FastdaimaResearch https://fastdaima.github.io/FastdaimaResearch/blog/ knowledge base about compilers, machine learning kernels, python and c++ quarto-1.6.39 Thu, 05 Dec 2024 00:00:00 GMT How to use azureopenai embeddings in lancedb fastdaima https://fastdaima.github.io/FastdaimaResearch/blog/posts/rag/how-to-use-azureopenai-embeddings-in-lancedb.html
  • I was unable to use azure openai embeddings inside lancedb
  • However, lancedb has fixed in their recent commit
  • Still I cannot able to install the dev version in my project, hence copied the changes from the commit, did an minor change on top of that, like added attribute api_version which worked for me.
  • let’s go.

References

  • https://github.com/lancedb/lancedb/commit/bd82e1f66d184bc13869030f03783cb4cb71da18
# !pip install lancedb openai -Uq

Implementation

from functools import cached_property
from typing import TYPE_CHECKING, List, Optional, Union
import logging
import os 
from lancedb.util import attempt_import_or_raise
from lancedb.embeddings.base import TextEmbeddingFunction
from lancedb.embeddings.registry import register, EmbeddingFunctionRegistry

if TYPE_CHECKING:
    import numpy as np


@register("azure-openai")
class AzureOpenAIEmbeddings(TextEmbeddingFunction):
    """
    An embedding function that uses the OpenAI API
    https://platform.openai.com/docs/guides/embeddings
    This can also be used for open source models that
    are compatible with the OpenAI API.
    Notes
    -----
    If you're running an Ollama server locally,
    you can just override the `base_url` parameter
    and provide the Ollama embedding model you want
    to use (https://ollama.com/library):
    ```python
    from lancedb.embeddings import get_registry
    openai = get_registry().get("openai")
    embedding_function = openai.create(
        name="<ollama-embedding-model-name>",
        base_url="http://localhost:11434",
        )
    ```
    """

    name: str = "text-embedding-ada-002"
    dim: Optional[int] = None
    azure_endpoint: Optional[str] = None
    default_headers: Optional[dict] = None
    azure_deployment: Optional[str] = None
    api_key: Optional[str] = None
    api_version: Optional[str] = None 
    use_azure: bool = False
    
    def ndims(self):
        return self._ndims

    @staticmethod
    def model_names():
        return [
            "text-embedding-ada-002",
            "text-embedding-3-large",
            "text-embedding-3-small",
        ]

    @cached_property
    def _ndims(self):
        if self.name == "text-embedding-ada-002":
            return 1536
        elif self.name == "text-embedding-3-large":
            return self.dim or 3072
        elif self.name == "text-embedding-3-small":
            return self.dim or 1536
        else:
            raise ValueError(f"Unknown model name {self.name}")

    def generate_embeddings(
        self, texts: Union[List[str], "np.ndarray"]
    ) -> List["np.array"]:
        """
        Get the embeddings for the given texts
        Parameters
        ----------
        texts: list[str] or np.ndarray (of str)
            The texts to embed
        """
        openai = attempt_import_or_raise("openai")

        valid_texts = []
        valid_indices = []
        for idx, text in enumerate(texts):
            if text:
                valid_texts.append(text)
                valid_indices.append(idx)

        # TODO retry, rate limit, token limit
        try:
            kwargs = {
                "input": valid_texts,
                "model": self.name,
            }
            if self.name != "text-embedding-ada-002":
                kwargs["dimensions"] = self.ndims()

            rs = self._openai_client.embeddings.create(**kwargs)
            valid_embeddings = {
                idx: v.embedding for v, idx in zip(rs.data, valid_indices)
            }
        except openai.BadRequestError:
            logging.exception("Bad request: %s", texts)
            return [None] * len(texts)
        except Exception:
            logging.exception("OpenAI embeddings error")
            raise
        return [valid_embeddings.get(idx, None) for idx in range(len(texts))]

    @cached_property
    def _openai_client(self):
        openai = attempt_import_or_raise("openai")
        kwargs = {}
        if self.azure_endpoint:
            kwargs["azure_endpoint"] = self.azure_endpoint
        if self.default_headers:
            kwargs["default_headers"] = self.default_headers
        if self.azure_deployment:
            kwargs["azure_deployment"] = self.azure_deployment
        if self.api_key:
            kwargs["api_key"] = self.api_key
        if self.api_version: 
            kwargs['api_version'] = self.api_version 
        if self.use_azure:
            return openai.AzureOpenAI(**kwargs)
        else:
            return openai.OpenAI(**kwargs)
# REQUIRED ENV VARIABLES, ALWAYS LOAD IT FROM .env

AZURE_EMBEDDING_MODEL= os.environ['AZURE_EMBEDDING_MODEL']
AZURE_EMBEDDING_ENDPOINT=os.environ['AZURE_EMBEDDING_ENDPOINT']
AZURE_EMBEDDING_DEPLOYMENT=os.environ['AZURE_EMBEDDING_DEPLOYMENT']
AZURE_EMBEDDING_API_KEY=os.environ['AZURE_EMBEDDING_API_KEY']
AZURE_API_VERSION=os.environ['AZURE_API_VERSION']
reg = EmbeddingFunctionRegistry.get_instance() 
azure_openai = reg.get('azure-openai')
  • Here by default, organization, base_url parameter is not working as expected, so I am passing azure_endpoint, azure_deployment and api_version parameters which are required for openai.AzureOpenAI API’s
embed_fn = azure_openai.create(
    api_version=AZURE_API_VERSION,
    name=AZURE_EMBEDDING_MODEL,
    azure_endpoint=AZURE_EMBEDDING_ENDPOINT,
    azure_deployment=AZURE_EMBEDDING_DEPLOYMENT,
    api_key=AZURE_EMBEDDING_API_KEY,
    use_azure=True
)
embed_fn.generate_embeddings('what is 2+3')
[[-0.011524682864546776,
  -0.012060285545885563,
  0.008892844431102276,
  0.023437215015292168,
  0.02672470547258854,
  0.010296491906046867,
  0.044251829385757446,
  0.05115925148129463,
  -0.0017534048529341817,
  0.04366081953048706,
  -0.002740344265475869,
  0.057808104902505875,
  -0.030270762741565704,
  0.007553838659077883,
  -0.023197118192911148,
  0.02506249211728573,
  -0.03551597148180008,
  -0.0005330050480552018,
  -0.028534671291708946,
  0.026780113577842712,
  0.017739515751600266,
  -0.028405388817191124,
  -0.012217272073030472,
  0.03211766481399536,
  0.013944127596914768,
  0.003534513060003519,
  0.03867417573928833,
  -0.009520791471004486,
  -0.030123010277748108,
  -0.021036241203546524,
  -2.9561291739810258e-05,
  0.016492854803800583,
  -0.014018003828823566,
  0.033114995807409286,
  -0.05585038661956787,
  0.03331815451383591,
  0.055222440510988235,
  0.002488703466951847,
  0.00489891367033124,
  -0.016178881749510765,
  0.01125688198953867,
  -0.03457405045628548,
  -0.02737112157046795,
  0.04779788479208946,
  0.016529792919754982,
  0.025727378204464912,
  0.01761946640908718,
  -0.015532465651631355,
  -0.01589261181652546,
  -0.002008508425205946,
  -0.008652746677398682,
  -0.0034560197964310646,
  0.030584735795855522,
  0.004522607196122408,
  -0.005060517694801092,
  0.008685067296028137,
  0.05939644202589989,
  0.00847267359495163,
  -0.041370656341314316,
  -0.030270762741565704,
  -0.020759204402565956,
  -0.015643279999494553,
  0.01741630770266056,
  0.010942908003926277,
  -0.03004913404583931,
  -0.006090166978538036,
  -0.006801225244998932,
  0.03527587279677391,
  -0.015550934709608555,
  0.059433382004499435,
  -0.01955871656537056,
  0.003534513060003519,
  0.0047211493365466595,
  0.0036499446723610163,
  -0.038858868181705475,
  -0.031175745651125908,
  0.022088976576924324,
  0.022255197167396545,
  0.011478510685265064,
  0.04702218249440193,
  0.0012235741596668959,
  -0.023825066164135933,
  -0.0035529821179807186,
  0.013981065712869167,
  0.012974502518773079,
  -0.027408059686422348,
  -0.012014112435281277,
  -0.0800078958272934,
  0.031175745651125908,
  0.021239399909973145,
  0.010628934018313885,
  -0.010739748366177082,
  -0.03815704211592674,
  -0.01571715623140335,
  0.03981925547122955,
  0.017425542697310448,
  -0.004931234754621983,
  0.02040829323232174,
  0.00021484692115336657,
  0.03250551596283913,
  0.007207544054836035,
  0.01857062242925167,
  -0.04081658646464348,
  -0.022790798917412758,
  0.02421291545033455,
  -0.016594436019659042,
  0.026669299229979515,
  0.014498199336230755,
  0.011598559096455574,
  0.0009546186774969101,
  -0.02057451382279396,
  0.003349822713062167,
  0.016382042318582535,
  0.006731966510415077,
  0.014950690791010857,
  0.00165528804063797,
  -0.012568184174597263,
  -0.020352885127067566,
  -0.004704988561570644,
  0.021276338025927544,
  -0.01083209365606308,
  -0.04668974131345749,
  0.02219979092478752,
  -0.00964084081351757,
  -0.006131722591817379,
  0.017425542697310448,
  -0.024046694859862328,
  0.018866127356886864,
  -0.015911079943180084,
  -0.06586060672998428,
  -0.026429200544953346,
  0.0021839644759893417,
  -0.024803925305604935,
  0.006755052600055933,
  0.007406086195260286,
  0.01877378299832344,
  0.013510105200111866,
  -0.008255662396550179,
  -0.0024586913641542196,
  -0.002260149223729968,
  -0.054779183119535446,
  -0.018219711259007454,
  0.007152136880904436,
  0.01915239728987217,
  -0.04912765696644783,
  -0.0019738790579140186,
  0.03163747116923332,
  0.0026202956214547157,
  -0.02081461250782013,
  0.017850330099463463,
  -0.018986176699399948,
  -0.04735462740063667,
  0.012854454107582569,
  0.0031466633081436157,
  -0.022402949631214142,
  -0.01293756440281868,
  -0.01546782348304987,
  0.034241605550050735,
  -0.001703769201412797,
  -0.021682657301425934,
  -0.012263445183634758,
  0.0052405912429094315,
  0.0325978621840477,
  0.030436983332037926,
  -0.02118399180471897,
  0.015532465651631355,
  0.027555812150239944,
  -0.015874141827225685,
  -0.026152165606617928,
  0.04524915665388107,
  8.1739948655013e-05,
  -0.012014112435281277,
  -0.03198838233947754,
  -0.032136134803295135,
  -0.00663038669154048,
  -0.0058500696904957294,
  -0.021571842953562737,
  -0.022809268906712532,
  -0.004095510579645634,
  0.01378714106976986,
  0.004963555373251438,
  0.03837867081165314,
  0.018755313009023666,
  -0.030954116955399513,
  -0.02631838619709015,
  0.016816064715385437,
  0.010638169012963772,
  0.012143395841121674,
  -0.004231719765812159,
  -0.05559182167053223,
  -0.01478446926921606,
  -0.04336531460285187,
  0.03433395177125931,
  -0.00033706004614941776,
  0.006879718508571386,
  -0.04957091063261032,
  -0.03065861202776432,
  -0.013593215495347977,
  -0.03453711047768593,
  0.006053228862583637,
  0.019207805395126343,
  0.0019253977807238698,
  0.011266116052865982,
  0.024674640968441963,
  -0.00035524051054380834,
  0.004931234754621983,
  -0.017425542697310448,
  -0.027426529675722122,
  0.02081461250782013,
  -0.021405620500445366,
  -0.0050743697211146355,
  0.02988291159272194,
  0.016529792919754982,
  -0.03195144608616829,
  -0.004300978500396013,
  0.0031535890884697437,
  -0.03684573993086815,
  -0.003400612622499466,
  -0.03154512494802475,
  -0.010157973505556583,
  0.010139504447579384,
  -0.018118131905794144,
  -0.025376465171575546,
  -0.002911183051764965,
  -0.026189103722572327,
  -0.007498431485146284,
  0.06002439185976982,
  0.03756603226065636,
  -0.013473167084157467,
  0.020888488739728928,
  -0.013759437017142773,
  -0.03520199656486511,
  0.002844232600182295,
  0.012743639759719372,
  -0.05847299098968506,
  -0.015486292541027069,
  -0.010721279308199883,
  -0.016659077256917953,
  -0.020870018750429153,
  0.02956893853843212,
  -0.017600998282432556,
  -0.0021539521403610706,
  -0.05621976777911186,
  -0.0346294566988945,
  -0.005374491680413485,
  -0.014876814559102058,
  -0.017259320244193077,
  -0.022458355873823166,
  0.010758217424154282,
  -0.0022959329653531313,
  -0.008823585696518421,
  0.029347309842705727,
  -0.016622139140963554,
  0.008075589314103127,
  0.0022001247853040695,
  -0.004416409879922867,
  0.051048435270786285,
  0.008740474469959736,
  -0.016308166086673737,
  -0.011293820105493069,
  -0.01332541462033987,
  -0.0075630731880664825,
  -0.013029909692704678,
  0.013602450489997864,
  0.002488703466951847,
  -0.0021216312889009714,
  -0.002479469170793891,
  -0.0022428345400840044,
  0.020796142518520355,
  0.001661059563048184,
  0.004584940150380135,
  0.013648622669279575,
  0.028091415762901306,
  0.03826785832643509,
  0.0418877899646759,
  -0.03226541727781296,
  -0.004361002705991268,
  -0.014581309631466866,
  0.013630153611302376,
  0.001094867940992117,
  -0.020722266286611557,
  -0.001871144981123507,
  0.028867114335298538,
  0.01799808256328106,
  -0.02040829323232174,
  -0.008629660122096539,
  -0.006671941839158535,
  0.01844133995473385,
  -0.02094389498233795,
  -0.03089870885014534,
  -0.03416772931814194,
  0.009031361900269985,
  0.02583819068968296,
  0.013491636142134666,
  0.006279474589973688,
  0.02020513266324997,
  0.04155534878373146,
  0.008888226933777332,
  -0.003869264619424939,
  -0.019780345261096954,
  -0.004698062781244516,
  -0.019226273521780968,
  -0.05163944512605667,
  -0.02295702137053013,
  0.034888021647930145,
  0.042294107377529144,
  0.00296197272837162,
  -0.008121761493384838,
  -0.014008769765496254,
  -0.034888021647930145,
  0.03459251672029495,
  -0.022162852808833122,
  0.007364531047642231,
  -0.04465814679861069,
  -0.0055037750862538815,
  -0.004337916616350412,
  -0.0026456904597580433,
  0.0033752175513654947,
  -0.012041816487908363,
  -0.0012570492690429091,
  -0.008426501415669918,
  0.03863723948597908,
  -0.04994029179215431,
  0.02988291159272194,
  0.025524217635393143,
  0.025727378204464912,
  0.005896241869777441,
  0.007470727898180485,
  0.022052038460969925,
  -0.04971866309642792,
  -0.03453711047768593,
  0.00634873379021883,
  -0.017028458416461945,
  -0.03466639295220375,
  0.03904355689883232,
  0.006667324807494879,
  -0.009188349358737469,
  -0.004531841725111008,
  -0.0012039507273584604,
  -0.00868968479335308,
  0.013953362591564655,
  0.00034110015258193016,
  0.024674640968441963,
  -0.030510859563946724,
  0.019060052931308746,
  0.017259320244193077,
  0.037233591079711914,
  -0.01663137413561344,
  7.647338497918099e-05,
  0.02255070209503174,
  0.00033821436227299273,
  -0.020315947011113167,
  0.026743175461888313,
  0.04174003750085831,
  0.028460795059800148,
  -0.023455685004591942,
  -0.008869757875800133,
  -0.000525790557730943,
  -0.007175223436206579,
  0.03649482876062393,
  0.005392960738390684,
  0.013510105200111866,
  0.0346294566988945,
  -0.02207050658762455,
  -0.007429172750562429,
  0.008107909932732582,
  0.02696480229496956,
  -0.005305232480168343,
  -0.03180369362235069,
  0.050198860466480255,
  -0.0325239859521389,
  0.030880240723490715,
  -0.033558253198862076,
  0.023197118192911148,
  0.004213250707834959,
  -0.020906956866383553,
  0.0007762769819237292,
  -0.012318851426243782,
  -0.01176478061825037,
  0.0014140361454337835,
  -0.0001131228927988559,
  0.011718607507646084,
  -0.001879225135780871,
  -0.027666626498103142,
  0.018626030534505844,
  -0.021516434848308563,
  0.002749578794464469,
  -0.04458427056670189,
  0.033059585839509964,
  0.018293587490916252,
  0.0024240619968622923,
  -0.023215588182210922,
  -0.00713828532025218,
  0.014488964341580868,
  -0.00939150806516409,
  -0.010028690099716187,
  0.021848877891898155,
  0.005346788093447685,
  0.027038678526878357,
  -0.038452547043561935,
  0.013417759910225868,
  -0.06065233796834946,
  -0.006990532856434584,
  -0.0324501097202301,
  0.0013644006103277206,
  -0.022790798917412758,
  -0.004986641462892294,
  -0.045138340443372726,
  -0.010989081114530563,
  0.009880938567221165,
  -0.040927398949861526,
  0.0022116678301244974,
  -0.024157509207725525,
  0.014027238823473454,
  0.015116912312805653,
  -0.00713828532025218,
  0.04281124100089073,
  0.01581873558461666,
  -0.003915437031537294,
  -0.03684573993086815,
  -0.004838889464735985,
  -0.008468056097626686,
  0.017397839576005936,
  0.010610464960336685,
  -0.03171134740114212,
  0.03235776349902153,
  0.0007451104465872049,
  0.023621905595064163,
  0.00937303900718689,
  -0.013371587730944157,
  -0.032339293509721756,
  -0.0011023710248991847,
  0.0031720581464469433,
  0.05754953995347023,
  -0.012946799397468567,
  -0.032745614647865295,
  0.032413169741630554,
  -0.01884765923023224,
  0.011774014681577682,
  0.005360639654099941,
  -0.0021839644759893417,
  0.03374294191598892,
  0.015495527535676956,
  -0.015837203711271286,
  -0.010361133143305779,
  0.030270762741565704,
  -0.04469508305191994,
  -0.021276338025927544,
  0.011533917859196663,
  0.028017539530992508,
  -0.010444244369864464,
  0.002571814227849245,
  0.001522541744634509,
  -0.02199663035571575,
  -0.005452984943985939,
  -0.018293587490916252,
  -0.016520559787750244,
  0.03765837848186493,
  -0.019614124670624733,
  0.0024309877771884203,
  0.010296491906046867,
  0.007253716699779034,
  0.0029504296835511923,
  -0.0025787402410060167,
  -0.013510105200111866,
  -0.03139737248420715,
  -0.026798581704497337,
  0.012263445183634758,
  0.002539493376389146,
  0.01544935442507267,
  -0.023344870656728745,
  -0.03896968066692352,
  0.02903333678841591,
  -0.010231849737465382,
  0.01026878785341978,
  0.03812010586261749,
  -0.03479567915201187,
  -0.03243163973093033,
  0.014664419926702976,
  -0.042294107377529144,
  -0.009834765456616879,
  -0.006574979517608881,
  0.0049912589602172375,
  0.02631838619709015,
  0.013583981432020664,
  0.003271329216659069,
  0.03682727366685867,
  0.014747531153261662,
  -0.001956564374268055,
  0.03326274827122688,
  -0.014581309631466866,
  0.010878266766667366,
  0.020038912072777748,
  6.497532012872398e-06,
  0.0001659328117966652,
  0.009331484325230122,
  0.0017834169557318091,
  -0.01727779023349285,
  -0.006057846359908581,
  0.05581345036625862,
  0.011164536699652672,
  0.0016829916276037693,
  0.03944987803697586,
  0.00853731483221054,
  0.00818178616464138,
  0.003312884597107768,
  0.029550468549132347,
  -0.017462480813264847,
  -0.025339527055621147,
  0.018422869965434074,
  0.04576629027724266,
  -0.01180171873420477,
  -0.01478446926921606,
  -0.03032616898417473,
  0.00633949926123023,
  0.009530026465654373,
  -0.03714124485850334,
  -0.004377163480967283,
  -0.005342170596122742,
  -0.01592954993247986,
  -0.02984597347676754,
  0.02094389498233795,
  -0.010656638070940971,
  0.025007084012031555,
  -0.0007099038339219987,
  0.02077767439186573,
  0.0028534671291708946,
  -0.006745818071067333,
  0.03558984771370888,
  -0.015726391226053238,
  0.01877378299832344,
  0.040964338928461075,
  -0.010638169012963772,
  -0.0063302647322416306,
  -0.011303054168820381,
  -0.016391275450587273,
  -0.04377163201570511,
  -0.0009303780389018357,
  -0.02875629998743534,
  -0.004386398009955883,
  -0.0006833546212874353,
  -0.007636949419975281,
  -0.009890172630548477,
  -0.024471482262015343,
  0.00011478221858851612,
  -0.016141943633556366,
  0.003924671560525894,
  -0.026540014892816544,
  -0.014636716805398464,
  0.006514954846352339,
  0.01018567755818367,
  0.013094551861286163,
  -0.021571842953562737,
  -0.005633058026432991,
  0.04709605872631073,
  0.00607631541788578,
  0.014165756292641163,
  -0.00964084081351757,
  -0.004088584333658218,
  -0.021534904837608337,
  0.010924438945949078,
  -0.0034190816804766655,
  0.008477291092276573,
  0.017120802775025368,
  0.007969392463564873,
  0.014036472886800766,
  0.0020558354444801807,
  0.0067873732186853886,
  -0.034924961626529694,
  -0.053597163408994675,
  -0.0035968462470918894,
  -0.02000197395682335,
  -0.01809042878448963,
  -0.010896735824644566,
  0.018626030534505844,
  -0.002387123880907893,
  0.021128585562109947,
  -0.036956556141376495,
  -0.008712771348655224,
  -0.028276104480028152,
  0.0012108766240999103,
  0.000872662290930748,
  0.021811939775943756,
  0.024323729798197746,
  0.0011843274114653468,
  0.01083209365606308,
  0.007369148079305887,
  -0.03619932383298874,
  -0.04004088416695595,
  0.014553606510162354,
  -0.021978162229061127,
  0.03475873917341232,
  0.01674218848347664,
  -0.019133929163217545,
  0.042183294892311096,
  -3.502625258988701e-05,
  -0.006408757995814085,
  0.0016899174079298973,
  -0.012752874754369259,
  -0.027426529675722122,
  -0.039708442986011505,
  -0.0007699282141402364,
  -0.002396358409896493,
  0.002304013120010495,
  0.0011623954633250833,
  0.01768410950899124,
  0.027389591559767723,
  0.011026019230484962,
  -0.007290654815733433,
  0.009096004068851471,
  0.03270867466926575,
  -0.012198803015053272,
  0.0017949601169675589,
  -0.010619699954986572,
  0.0018168921815231442,
  -0.010915204882621765,
  0.002839615335687995,
  -0.005785427987575531,
  0.0029388864059001207,
  -0.013039144687354565,
  0.008551167324185371,
  -0.01283598504960537,
  -0.01565251499414444,
  0.023584967479109764,
  -0.00979782734066248,
  0.010351899079978466,
  -0.01939249597489834,
  -0.01079515554010868,
  0.0018088119104504585,
  0.00042449941975064576,
  -0.011515448801219463,
  -0.04916459321975708,
  -0.00858810544013977,
  -0.03758450224995613,
  0.04750237986445427,
  -0.0025625796988606453,
  0.013667091727256775,
  -0.009142176248133183,
  -0.01492298673838377,
  0.004963555373251438,
  0.021128585562109947,
  -0.011838656850159168,
  0.020592983812093735,
  -0.05415123328566551,
  -0.018478278070688248,
  -0.007170605938881636,
  -0.01079515554010868,
  0.010379602201282978,
  0.01121994387358427,
  0.037233591079711914,
  0.0074430243112146854,
  0.01168167032301426,
  0.025561155751347542,
  0.004065498244017363,
  0.0021585694048553705,
  0.010361133143305779,
  0.016880705952644348,
  -0.024194447323679924,
  -0.012522011063992977,
  -0.012161864899098873,
  0.0018053490202873945,
  -0.03645789250731468,
  -0.024286791682243347,
  0.0012812898494303226,
  0.002353648655116558,
  -0.014452026225626469,
  -0.005711551755666733,
  -0.017564060166478157,
  0.03837867081165314,
  0.02410210110247135,
  0.021147053688764572,
  -0.02316018007695675,
  0.01427657064050436,
  0.0010267633479088545,
  0.02838691882789135,
  -0.03786153718829155,
  -0.003178984159603715,
  0.0022405257914215326,
  -0.0073599135503172874,
  0.02162724919617176,
  0.010804390534758568,
  -0.008939016610383987,
  0.02652154676616192,
  0.018856892362236977,
  0.013159193098545074,
  0.0027195666916668415,
  -0.020906956866383553,
  0.0027588133234530687,
  0.013066847808659077,
  0.009165262803435326,
  -0.00964084081351757,
  0.02883017621934414,
  0.031231151893734932,
  -0.0059839701279997826,
  -0.015864908695220947,
  0.020666860044002533,
  0.034038446843624115,
  -0.02149796672165394,
  -0.020722266286611557,
  -0.008514229208230972,
  0.007170605938881636,
  -0.008731240406632423,
  -0.0027680478524416685,
  0.006907422095537186,
  0.046911370009183884,
  0.014932221733033657,
  0.03025229275226593,
  -0.007369148079305887,
  -0.012069519609212875,
  -0.0019046200904995203,
  0.011413868516683578,
  0.022218259051442146,
  0.01176478061825037,
  0.04539690911769867,
  0.015384713187813759,
  0.01894923858344555,
  0.002484086435288191,
  -0.030547797679901123,
  -0.00304277497343719,
  0.010952142998576164,
  0.02631838619709015,
  -0.005332936067134142,
  -0.037196654826402664,
  0.027481935918331146,
  -0.014211928471922874,
  -0.007235247641801834,
  0.015837203711271286,
  -0.008417266421020031,
  -0.007821639999747276,
  0.009862469509243965,
  -0.01883842423558235,
  -0.015033801086246967,
  0.033022649586200714,
  0.02515483647584915,
  -0.000569077383261174,
  0.02478545531630516,
  -0.010037925094366074,
  -0.0025117897894233465,
  0.003259786171838641,
  -0.004270966164767742,
  -0.002929651876911521,
  0.002442531054839492,
  0.003218230791389942,
  0.03062167391180992,
  -0.009229904040694237,
  -0.010776686482131481,
  -0.008800499141216278,
  -0.018561389297246933,
  -0.0025002467446029186,
  0.009520791471004486,
  0.03377988189458847,
  -0.03435242176055908,
  0.011201474815607071,
  -0.03878499194979668,
  0.007821639999747276,
  0.019244743511080742,
  -0.011441572569310665,
  0.009345335885882378,
  -0.0050051105208694935,
  -0.0029896763153374195,
  -0.001390949822962284,
  0.026798581704497337,
  0.02461923472583294,
  0.001214339630678296,
  0.022366011515259743,
  -0.04524915665388107,
  -0.0016379732405766845,
  -0.023307932540774345,
  0.025967475026845932,
  0.013777906075119972,
  -0.0313604362308979,
  0.004746544174849987,
  0.020279008895158768,
  -0.0015952636022120714,
  0.028017539530992508,
  -0.0005399309447966516,
  -6.089012822485529e-05,
  -0.010019456036388874,
  -0.00937303900718689,
  -0.013971831649541855,
  -0.047280751168727875,
  -0.008915930986404419,
  -0.03215460479259491,
  -0.021479496732354164,
  -0.0060670808888971806,
  0.019410964101552963,
  0.03438935801386833,
  -0.00442102737724781,
  0.03871111571788788,
  0.00515286298468709,
  0.01840440183877945,
  0.01989115960896015,
  0.03139737248420715,
  0.0011046796571463346,
  0.022255197167396545,
  0.01293756440281868,
  0.021368682384490967,
  0.03588535264134407,
  0.011007550172507763,
  0.012891392223536968,
  0.019133929163217545,
  0.013454698026180267,
  -0.008675833232700825,
  0.017157740890979767,
  0.029107213020324707,
  0.008204872719943523,
  -0.020186664536595345,
  -0.013426994904875755,
  -0.0007630023756064475,
  0.012115692719817162,
  -0.0007191383629105985,
  0.006524189375340939,
  -0.00712905079126358,
  0.022735392674803734,
  -0.0025210243184119463,
  0.02631838619709015,
  0.020906956866383553,
  -0.04491671174764633,
  -0.014664419926702976,
  0.029070274904370308,
  0.010518120601773262,
  0.02264304645359516,
  0.024360667914152145,
  -0.025930536910891533,
  -0.01737936958670616,
  0.023418746888637543,
  0.07882587611675262,
  0.028663955628871918,
  -0.022088976576924324,
  -0.0038507955614477396,
  -0.02101777121424675,
  0.013897955417633057,
  0.012161864899098873,
  0.008546549826860428,
  -0.0052405912429094315,
  0.013694795779883862,
  -0.02441607601940632,
  -0.0008645820780657232,
  -0.003709969110786915,
  0.028700893744826317,
  -0.010425775311887264,
  -0.017240852117538452,
  0.014387384988367558,
  -0.021775001659989357,
  -0.014913752675056458,
  0.02737112157046795,
  -0.010656638070940971,
  -0.025856660678982735,
  -0.004012399818748236,
  0.004494903609156609,
  0.008380328305065632,
  -0.0075492216274142265,
  0.009409977123141289,
  0.011053722351789474,
  -0.004063189495354891,
  0.02757428213953972,
  -0.02016819454729557,
  0.006136339623481035,
  0.011718607507646084,
  -0.015698686242103577,
  0.004769630264490843,
  -0.0041255224496126175,
  -0.02960587665438652,
  -0.030566265806555748,
  0.02040829323232174,
  -0.009465384297072887,
  0.019669530913233757,
  -0.005674613639712334,
  0.006491868756711483,
  -0.015624810941517353,
  -0.002671085298061371,
  -0.02247682586312294,
  -0.025764314457774162,
  -0.00442102737724781,
  -0.0334659069776535,
  -0.009964048862457275,
  0.0023605746682733297,
  0.014082645066082478,
  -0.0007306815241463482,
  0.006399523466825485,
  0.01427657064050436,
  0.000962121703196317,
  -0.058251362293958664,
  -0.021848877891898155,
  0.010905969887971878,
  -0.024933207780122757,
  -0.024914739653468132,
  -0.010231849737465382,
  -0.0005038586095906794,
  0.0019784963224083185,
  -0.01069357618689537,
  0.008629660122096539,
  0.022329073399305344,
  -0.045212216675281525,
  -0.018007317557930946,
  0.042183294892311096,
  -0.013630153611302376,
  -0.016271227970719337,
  -0.03344743698835373,
  -0.004095510579645634,
  0.012891392223536968,
  -0.008758943527936935,
  0.01867220364511013,
  0.003613006556406617,
  -0.005272911861538887,
  -0.010121035389602184,
  0.01728702522814274,
  0.043550003319978714,
  0.01534777507185936,
  0.011718607507646084,
  -0.00782625749707222,
  -0.003527587279677391,
  -0.00833415612578392,
  -0.029956787824630737,
  0.0007220241823233664,
  -0.009714717045426369,
  -0.015283132903277874,
  -0.008343390189111233,
  0.02600441314280033,
  -0.03812010586261749,
  -0.03762144222855568,
  0.01519078854471445,
  -0.025376465171575546,
  0.03904355689883232,
  0.002650307724252343,
  -0.015532465651631355,
  -0.015320071019232273,
  -0.014766000211238861,
  0.009603902697563171,
  0.018256649374961853,
  -0.01602189429104328,
  0.008241810835897923,
  0.031692877411842346,
  -0.010342664085328579,
  0.026152165606617928,
  0.016548262909054756,
  0.007747763767838478,
  0.009354570880532265,
  -0.019263211637735367,
  -0.026022881269454956,
  -0.001999273896217346,
  0.015236960723996162,
  -0.009192965924739838,
  -0.05086374655365944,
  -0.02105470933020115,
  0.012438900768756866,
  -0.013445463962852955,
  0.019854221493005753,
  -0.015827970579266548,
  -0.014165756292641163,
  -0.009456150233745575,
  0.013639388605952263,
  -0.032099198549985886,
  -0.0058777728118002415,
  0.020500637590885162,
  -0.027851317077875137,
  -0.032782550901174545,
  0.035165056586265564,
  -0.014202694408595562,
  -0.0016252758214250207,
  0.007267568726092577,
  -0.0065934485755860806,
  -0.01425810158252716,
  -0.015615575946867466,
  -0.03004913404583931,
  -0.005919328425079584,
  0.01111836452037096,
  -0.007267568726092577,
  0.022015100345015526,
  0.009603902697563171,
  -5.948330363025889e-05,
  -0.006270240060985088,
  -0.004709606058895588,
  0.01656673103570938,
  0.024046694859862328,
  0.03618085756897926,
  -0.016751421615481377,
  0.02356649935245514,
  -0.009026745334267616,
  0.01759176328778267,
  0.011653966270387173,
  -0.001303221913985908,
  -0.034961897879838943,
  0.02624450996518135,
  0.04990335553884506,
  -0.021073179319500923,
  0.015163084492087364,
  -0.015707921236753464,
  0.0075261350721120834,
  0.008943634107708931,
  0.015153850428760052,
  0.013925658538937569,
  -0.00845420453697443,
  -0.009880938567221165,
  0.004665742162615061,
  0.018044255673885345,
  0.010370368137955666,
  -0.013861017301678658,
  0.02936577796936035,
  -0.03510965034365654,
  ...],
 [-0.03789591044187546,
  -0.014004245400428772,
  0.007991744205355644,
  0.011233304627239704,
  0.0224138293415308,
  0.00478097191080451,
  0.016898339614272118,
  0.04222385585308075,
  -0.003962884657084942,
  0.06154127046465874,
  -0.015139011666178703,
  -0.005665034055709839,
  -0.04880373924970627,
  0.017470119521021843,
  -0.0020540147088468075,
  0.012922259047627449,
  -0.024683361873030663,
  0.018138663843274117,
  -0.015297351405024529,
  -0.007191250566393137,
  0.041097886860370636,
  -0.02593248523771763,
  -0.03237162157893181,
  -0.0058013820089399815,
  0.03947930410504341,
  0.0036835914943367243,
  0.02195640467107296,
  -0.007116478867828846,
  0.0033977008424699306,
  -0.0031228060834109783,
  -0.018912767991423607,
  0.054574333131313324,
  -0.022167524322867393,
  0.0005297224852256477,
  0.004200393799692392,
  0.02213233709335327,
  0.03652363643050194,
  0.0016867551021277905,
  -0.032705895602703094,
  -0.014180177822709084,
  0.01574597880244255,
  -0.03698106110095978,
  -0.0581633597612381,
  0.0337790846824646,
  0.021710099652409554,
  -0.015350131317973137,
  0.010186505503952503,
  0.00944758765399456,
  0.00048518949188292027,
  -0.013423667289316654,
  -0.03515135869383812,
  0.024173157289624214,
  0.015314944088459015,
  0.010811066254973412,
  -0.028536289930343628,
  0.008321618661284447,
  0.06952861696481705,
  0.031192872673273087,
  -0.00997538585215807,
  0.011629153974354267,
  -0.013010225258767605,
  0.0031799841672182083,
  0.008449169807136059,
  0.010811066254973412,
  -0.04437023401260376,
  -0.004649022594094276,
  0.023381460458040237,
  0.014109805226325989,
  -0.006949342787265778,
  0.05140754580497742,
  0.0014514450449496508,
  -0.02026745118200779,
  0.012130562216043472,
  0.017505306750535965,
  -0.012033798731863499,
  -0.019247040152549744,
  0.019458159804344177,
  0.037368111312389374,
  -0.012552800588309765,
  0.03490505367517471,
  -0.009649910032749176,
  -0.03418372943997383,
  -0.02605563774704933,
  0.013960261829197407,
  0.02593248523771763,
  -0.013995449058711529,
  -0.002700567478314042,
  -0.0844477117061615,
  0.017953936010599136,
  0.02473614178597927,
  0.012816699221730232,
  0.012570393271744251,
  -0.003861723467707634,
  0.0016966513358056545,
  -0.008726263418793678,
  0.015904318541288376,
  0.0019253637874498963,
  0.020249856635928154,
  0.011074965819716454,
  0.027726998552680016,
  -0.012271308340132236,
  0.038916319608688354,
  -0.00045467616291716695,
  -0.005753000266849995,
  0.006447934545576572,
  0.021622132509946823,
  0.01744372956454754,
  0.025211161002516747,
  -0.0021045953035354614,
  0.03624214231967926,
  -0.015526063740253448,
  -0.0007306706393137574,
  0.005577067378908396,
  0.0015592038398608565,
  -0.021886032074689865,
  -0.03659400716423988,
  -0.025175973773002625,
  0.02948632650077343,
  0.0007960955845192075,
  0.010485590435564518,
  -0.03221328184008598,
  -0.03601343184709549,
  0.00016191309259738773,
  0.0037165789399296045,
  0.008765848353505135,
  -0.004939311649650335,
  -0.0519353412091732,
  -0.012878276407718658,
  -0.002693969989195466,
  -0.05415209382772446,
  -0.015482080169022083,
  0.008497551083564758,
  0.015420503914356232,
  -0.022994408383965492,
  0.0230471882969141,
  -0.018912767991423607,
  0.03852926939725876,
  0.03411335498094559,
  -0.022149931639432907,
  0.0016812571557238698,
  -0.0665377601981163,
  0.0059861112385988235,
  -0.00815008394420147,
  -0.038458894938230515,
  -0.02144620008766651,
  -0.03495783358812332,
  0.010424014180898666,
  -0.03374389931559563,
  -0.01757567934691906,
  -0.0005668332451023161,
  0.01657286286354065,
  -0.026829741895198822,
  0.017144644632935524,
  0.001324993441812694,
  -0.005616652313619852,
  -0.004053050186485052,
  9.511363168712705e-05,
  0.014144991524517536,
  -0.0034768707118928432,
  -0.028501102700829506,
  -0.028518695384263992,
  -0.0021903624292463064,
  -0.021393420174717903,
  -0.00316019169986248,
  -0.01770762912929058,
  -0.0016229794127866626,
  0.023135153576731682,
  0.015385317616164684,
  -0.030928974971175194,
  0.04912041872739792,
  -0.007138470653444529,
  0.005955323111265898,
  -0.015552453696727753,
  -0.0021221886854618788,
  -0.03705143183469772,
  -0.008084109053015709,
  0.0058145769871771336,
  0.010415217839181423,
  0.009157299064099789,
  0.022695321589708328,
  0.005607855971902609,
  0.05682627111673355,
  -0.005295575130730867,
  -0.005563872866332531,
  -0.012816699221730232,
  -0.004684208892285824,
  0.02793811820447445,
  0.029820598661899567,
  -0.0024300708901137114,
  -0.03432447463274002,
  0.0006602975190617144,
  -0.009605927392840385,
  0.038951508700847626,
  -0.012077782303094864,
  0.018156258389353752,
  -0.055489182472229004,
  -0.017135847359895706,
  -0.005277981981635094,
  -0.0094299940392375,
  -0.004121224395930767,
  -0.018191443756222725,
  -0.012368070892989635,
  -0.0014470467576757073,
  -0.013115785084664822,
  -0.0004472539876587689,
  -0.005009684711694717,
  0.03289942070841789,
  0.0059157381765544415,
  0.010661523789167404,
  -0.002326710382476449,
  0.03951449319720268,
  0.015508470125496387,
  0.012658360414206982,
  -0.024349089711904526,
  0.03719218075275421,
  -0.0013128980062901974,
  -0.0010347043862566352,
  0.0076574720442295074,
  -0.0028325170278549194,
  -0.02865944243967533,
  0.007982947863638401,
  -0.023117560893297195,
  0.02711123414337635,
  0.0360662117600441,
  -0.001710945856757462,
  0.0311752799898386,
  0.0383533351123333,
  0.0571429505944252,
  -0.006641460582613945,
  0.035309698432683945,
  -0.014188975095748901,
  0.01575477607548237,
  0.00021744186233263463,
  0.010212895460426807,
  -0.03235403075814247,
  0.005036074668169022,
  -0.01293985266238451,
  0.005229600239545107,
  0.021833252161741257,
  0.013159768655896187,
  -0.032019756734371185,
  0.03085860051214695,
  -0.017109457403421402,
  -0.008251245133578777,
  -0.0006064181216061115,
  0.0021155911963433027,
  -0.009078129194676876,
  -0.00021702951926272362,
  -0.012447240762412548,
  0.004132220055907965,
  0.03997191786766052,
  -0.0010984799591824412,
  0.00986102968454361,
  0.0004829903191421181,
  0.021833252161741257,
  0.017830781638622284,
  0.02524634636938572,
  0.03277626633644104,
  -0.04556657746434212,
  -0.044299863278865814,
  0.0057398052886128426,
  -0.0033647133968770504,
  -0.005291176959872246,
  0.009852233342826366,
  -0.012895869091153145,
  -0.002696169074624777,
  0.010591150261461735,
  -0.001248022774234414,
  0.022149931639432907,
  -0.015578843653202057,
  0.03694587200880051,
  0.01513021532446146,
  0.041097886860370636,
  -0.015605233609676361,
  0.044159114360809326,
  0.001531714340671897,
  -0.043842438608407974,
  -0.016097845509648323,
  0.023909257724881172,
  0.035344887524843216,
  -0.025809332728385925,
  -0.001118822256103158,
  0.036312516778707504,
  -0.0011578572448343039,
  -0.029187239706516266,
  -0.032846640795469284,
  -0.01460241712629795,
  -0.020601723343133926,
  -0.011066168546676636,
  -0.034764308482408524,
  -0.028905747458338737,
  0.016678422689437866,
  0.007947761565446854,
  0.014118601568043232,
  0.016361743211746216,
  0.038318149745464325,
  0.03867001459002495,
  0.025439873337745667,
  0.028501102700829506,
  0.006918554659932852,
  0.013784329406917095,
  -0.0005508893518708646,
  -0.023909257724881172,
  -0.02355739288032055,
  0.006944944616407156,
  0.041871991008520126,
  -0.03824777528643608,
  0.003954088315367699,
  -0.023909257724881172,
  -0.02287125587463379,
  0.031703077256679535,
  -0.02309996820986271,
  -0.025721365585923195,
  -0.008567923679947853,
  -0.019510939717292786,
  -0.026038045063614845,
  -0.0031931791454553604,
  0.023856477811932564,
  -0.01652888022363186,
  -0.02086562290787697,
  0.0004593493649736047,
  0.036171771585941315,
  -0.05608735606074333,
  0.01104857586324215,
  0.015095028094947338,
  0.034623563289642334,
  -0.032653115689754486,
  0.0010105136316269636,
  0.02359258010983467,
  -0.03761441633105278,
  -0.04120344668626785,
  -0.021481387317180634,
  -0.012077782303094864,
  -0.011136542074382305,
  0.03011968359351158,
  -0.028729815036058426,
  0.010784676298499107,
  0.012033798731863499,
  -0.023311087861657143,
  -0.0019616498611867428,
  -0.002685173414647579,
  0.006065281108021736,
  -0.0019077705219388008,
  0.008915390819311142,
  -0.012544004246592522,
  0.0010940816719084978,
  0.031016940250992775,
  0.005792585201561451,
  0.029750224202871323,
  0.016898339614272118,
  0.00016452459385618567,
  -0.02885296754539013,
  0.0627376139163971,
  0.037227366119623184,
  0.014092211611568928,
  -0.020566536113619804,
  -0.01465519703924656,
  0.00471499701961875,
  0.0028193220496177673,
  0.001749431132338941,
  0.024841701611876488,
  -0.011259694583714008,
  0.020091518759727478,
  -0.030788227915763855,
  -0.0020441184751689434,
  0.007213241886347532,
  0.022536983713507652,
  -0.006588680669665337,
  -0.005849763285368681,
  -0.00987862329930067,
  -0.04342019930481911,
  -0.01218334212899208,
  0.029415952041745186,
  0.0024234734009951353,
  0.002575215417891741,
  -0.004886531736701727,
  0.0011732514249160886,
  0.0006179637275636196,
  0.02118230052292347,
  0.009350825101137161,
  -0.008110499009490013,
  0.03406057879328728,
  -0.004820556845515966,
  -0.03538007289171219,
  0.012552800588309765,
  -0.009755469858646393,
  0.008954975754022598,
  -0.0571429505944252,
  0.01139164436608553,
  0.022836068645119667,
  0.013265327550470829,
  -0.04039415344595909,
  -0.00619723042473197,
  0.049050044268369675,
  0.004952506627887487,
  -0.015781166031956673,
  0.001991338562220335,
  0.006698638666421175,
  0.048592619597911835,
  -0.00828203372657299,
  0.0037451679818332195,
  -0.04451098293066025,
  0.002608202863484621,
  -0.01898314245045185,
  -0.0009967689402401447,
  -0.016036268323659897,
  0.0033779083751142025,
  -0.020302636548876762,
  0.00304583553224802,
  0.019141480326652527,
  -0.014672789722681046,
  0.003969482146203518,
  -0.0060300943441689014,
  -0.02355739288032055,
  0.017778001725673676,
  -0.006245612166821957,
  0.04665736109018326,
  0.006276400294154882,
  -0.024982448667287827,
  -0.049401912838220596,
  0.017329374328255653,
  0.005502296146005392,
  -0.004556657746434212,
  0.01579875871539116,
  -0.015517267398536205,
  0.021147115156054497,
  0.0028677035588771105,
  0.008172075264155865,
  -0.02067209593951702,
  -0.045988816767930984,
  -0.006861376576125622,
  0.03126324713230133,
  0.03196697682142258,
  0.023680545389652252,
  -0.018789615482091904,
  -0.02670658938586712,
  0.024542616680264473,
  -0.003556040348485112,
  0.023363865911960602,
  0.012728733010590076,
  0.0003012848028447479,
  -0.02281847596168518,
  -0.006988927721977234,
  -0.009887419641017914,
  -0.04222385585308075,
  -0.0016526681138202548,
  -0.017628459259867668,
  -0.04451098293066025,
  0.013432464562356472,
  -0.0021947608329355717,
  -0.028272390365600586,
  -0.012904666364192963,
  -0.03026043064892292,
  -0.008400787599384785,
  -0.01917666755616665,
  -0.024859294295310974,
  -0.0312984324991703,
  0.02144620008766651,
  -0.014092211611568928,
  -0.015693198889493942,
  -0.0033229293767362833,
  0.0259676706045866,
  0.0022145533002913,
  -0.02793811820447445,
  0.020144298672676086,
  -0.00437192851677537,
  -0.031879011541604996,
  0.020742468535900116,
  0.02396203763782978,
  0.008893399499356747,
  -0.007235233671963215,
  -0.026231570169329643,
  0.0006306088762357831,
  0.0008087407913990319,
  0.006856978405267,
  0.006984529551118612,
  -0.032529961317777634,
  -0.039408933371305466,
  0.004020062740892172,
  -0.04176643118262291,
  0.010424014180898666,
  -0.024313904345035553,
  0.01779559627175331,
  0.021710099652409554,
  0.029152054339647293,
  0.008290830068290234,
  0.04391280934214592,
  0.025211161002516747,
  -0.03409576416015625,
  0.01748771406710148,
  0.01424175500869751,
  0.0006504013435915112,
  0.02605563774704933,
  -0.030524328351020813,
  -0.013713956810534,
  -0.02756865881383419,
  -0.001957251690328121,
  -0.02542228065431118,
  0.02730475924909115,
  0.02661862224340439,
  0.01033604796975851,
  -0.005084455944597721,
  0.010503184050321579,
  -0.010133725591003895,
  -0.02450742945075035,
  -0.004444500431418419,
  0.00418719882145524,
  -0.007244030013680458,
  -0.054715078324079514,
  0.014725569635629654,
  0.0224138293415308,
  0.00987862329930067,
  0.011127745732665062,
  -0.013713956810534,
  0.01615062542259693,
  0.01903592236340046,
  -0.014092211611568928,
  0.004312551114708185,
  -0.01875443011522293,
  -0.002735754009336233,
  0.0012139359023422003,
  0.0051944139413535595,
  0.0009819245897233486,
  0.03666438162326813,
  -0.02118230052292347,
  0.008581118658185005,
  0.027287166565656662,
  0.019299820065498352,
  0.022290676832199097,
  -0.04394799470901489,
  0.034993018954992294,
  0.02765662595629692,
  -0.015490877442061901,
  -0.008092905394732952,
  0.006883368361741304,
  -0.013344497419893742,
  -0.008044524118304253,
  0.008642695844173431,
  -0.01607145555317402,
  -0.008757052011787891,
  -0.03666438162326813,
  0.01465519703924656,
  0.04422948881983757,
  -0.004561055917292833,
  -0.021164707839488983,
  0.0026697791181504726,
  0.01629137061536312,
  -0.023029593750834465,
  -0.010124928317964077,
  0.005942128133028746,
  0.004591844510287046,
  0.028518695384263992,
  0.004396119154989719,
  0.011576374061405659,
  0.020197076722979546,
  -0.002821521135047078,
  0.0005459412350319326,
  -0.04053490236401558,
  0.007050504442304373,
  -0.02086562290787697,
  0.021252673119306564,
  0.002269532298669219,
  0.0241907499730587,
  -0.0067646135576069355,
  -0.024630581960082054,
  0.019827619194984436,
  0.0018967747455462813,
  0.010758286342024803,
  -0.0007053802837617695,
  -0.024929668754339218,
  -0.004908523056656122,
  -0.029873378574848175,
  0.0005827771965414286,
  -0.03291701525449753,
  -0.0006498515140265226,
  -0.013775533065199852,
  -0.01229769829660654,
  -0.00661946926265955,
  -0.0267945546656847,
  -0.021287860348820686,
  -0.00926285795867443,
  0.01756688393652439,
  0.010793473571538925,
  -0.014303331263363361,
  0.02871222235262394,
  0.01216574851423502,
  0.010362437926232815,
  -0.04243497550487518,
  -0.028553882613778114,
  0.01857849583029747,
  -0.017382154241204262,
  0.023803697898983955,
  0.02888815477490425,
  -0.03673475608229637,
  0.04363131895661354,
  -0.0015372122870758176,
  -0.022660136222839355,
  0.009896215982735157,
  -0.00460064085200429,
  -0.006241213530302048,
  -0.046129561960697174,
  0.009060535579919815,
  0.014021839015185833,
  -0.007112080696970224,
  0.022149931639432907,
  0.050457507371902466,
  0.014488060027360916,
  -0.02067209593951702,
  0.01610664092004299,
  -0.0038023460656404495,
  -0.0032569547183811665,
  -0.011593966744840145,
  -0.011198118329048157,
  0.010591150261461735,
  0.024208344519138336,
  0.011057372204959393,
  -0.02350461296737194,
  0.0031997766345739365,
  0.00042031428893096745,
  -0.0047018020413815975,
  0.04563695192337036,
  0.00033894539228640497,
  0.017285391688346863,
  0.008941780775785446,
  -0.007195648737251759,
  0.0031513951253145933,
  -0.01898314245045185,
  -0.03288182616233826,
  0.006892164703458548,
  0.0344124436378479,
  0.010116131976246834,
  -0.03495783358812332,
  0.005484702996909618,
  -0.025140786543488503,
  0.04180161654949188,
  0.028237203136086464,
  0.0212702676653862,
  0.01097820233553648,
  0.00873945839703083,
  -0.0010946315014734864,
  -0.007446352858096361,
  0.015367724001407623,
  0.002100196899846196,
  -0.08705151826143265,
  -0.05056306719779968,
  0.005000887904316187,
  0.0005228500813245773,
  0.012482427060604095,
  0.004908523056656122,
  0.040464527904987335,
  0.0039980714209377766,
  -0.015490877442061901,
  0.018648870289325714,
  0.025035228580236435,
  0.009931402280926704,
  -0.0373329259455204,
  0.004242178052663803,
  -0.010573557578027248,
  -0.03008449636399746,
  -0.011646746657788754,
  -0.0212702676653862,
  -0.03122805990278721,
  0.01022169180214405,
  0.008123693987727165,
  -0.0002546901232562959,
  0.024806514382362366,
  -0.007156063802540302,
  -0.014021839015185833,
  0.039690423756837845,
  0.007666268851608038,
  -0.015226977877318859,
  -0.023451833054423332,
  0.0023311087861657143,
  0.009693893603980541,
  0.01812107115983963,
  -0.011083762161433697,
  -0.028923342004418373,
  0.019704466685652733,
  -0.008915390819311142,
  0.03648844733834267,
  0.010555963963270187,
  -0.00636876467615366,
  -0.015622826293110847,
  0.014963079243898392,
  0.00478097191080451,
  -0.01852571591734886,
  0.009157299064099789,
  0.009500367566943169,
  -0.014593619853258133,
  -0.029521511867642403,
  -0.0186664629727602,
  0.010538370348513126,
  0.004360932391136885,
  0.002764343051239848,
  0.012139358557760715,
  0.03803665563464165,
  0.02730475924909115,
  -0.032512370496988297,
  -0.010186505503952503,
  0.010142521932721138,
  -0.004325746092945337,
  0.011690730229020119,
  -0.015385317616164684,
  -0.015490877442061901,
  0.039831168949604034,
  0.003929897211492062,
  0.018384970724582672,
  -0.01666083000600338,
  -0.02223789691925049,
  -0.007336394861340523,
  0.019546126946806908,
  0.03054192289710045,
  0.012922259047627449,
  0.027005674317479134,
  0.024947261437773705,
  0.012500020675361156,
  -0.009095721878111362,
  0.00017510804173070937,
  -0.0037935494910925627,
  0.004917319864034653,
  0.015939505770802498,
  0.004154211841523647,
  -0.012693546712398529,
  0.017637256532907486,
  0.00849315244704485,
  0.0018824802245944738,
  0.047537025064229965,
  0.019686872139573097,
  0.015904318541288376,
  0.008699873462319374,
  -0.005884950049221516,
  -0.020707283169031143,
  -0.012420850805938244,
  0.01730298437178135,
  -0.011268491856753826,
  0.003756163874641061,
  -0.011022185906767845,
  0.01776040904223919,
  -0.006685443688184023,
  0.01829700358211994,
  0.005774992052465677,
  -0.008752653375267982,
  0.0030656277667731047,
  -0.010424014180898666,
  -0.0007273718947544694,
  0.009940199553966522,
  0.0003620365750975907,
  -0.0005154279060661793,
  -0.013546820729970932,
  0.021762879565358162,
  0.04764258489012718,
  -0.020619316026568413,
  0.02172769233584404,
  -0.07867711782455444,
  -0.0016394731355831027,
  0.03976079821586609,
  -0.010899032466113567,
  0.009377215057611465,
  -0.019194260239601135,
  -0.01610664092004299,
  0.017012694850564003,
  0.022800881415605545,
  0.01346765086054802,
  0.030665075406432152,
  0.025316720828413963,
  -0.012376868166029453,
  0.015165401622653008,
  -0.027093641459941864,
  0.053202059119939804,
  0.025035228580236435,
  -0.03048914298415184,
  0.014681586064398289,
  -0.019071107730269432,
  -0.016177015379071236,
  0.045390643179416656,
  -0.01620340533554554,
  0.019106294959783554,
  -0.028131643310189247,
  0.0041630081832408905,
  0.022431423887610435,
  0.004468691535294056,
  -0.004693005699664354,
  -0.014505653642117977,
  0.013001428917050362,
  -0.0028896951116621494,
  -0.00033509687636978924,
  0.015332537703216076,
  0.02132304757833481,
  0.023715732619166374,
  -0.005058065988123417,
  0.00872186478227377,
  -0.006769011728465557,
  0.03603102266788483,
  0.041696056723594666,
  0.013485243543982506,
  -0.027990898117423058,
  0.014092211611568928,
  0.001541610574349761,
  0.0010413018753752112,
  -0.006395154632627964,
  0.03515135869383812,
  0.013177361339330673,
  0.014628807082772255,
  0.0250176340341568,
  -0.009025349281728268,
  -0.0032899421639740467,
  -0.002795131178572774,
  -0.032019756734371185,
  -0.007411166559904814,
  0.0032305647619068623,
  -0.017461324110627174,
  -0.021534167230129242,
  0.0425405353307724,
  0.049050044268369675,
  0.024665769189596176,
  0.005946526303887367,
  0.01634415052831173,
  -0.02716401405632496,
  -0.0032349631655961275,
  0.01921185478568077,
  0.002718160627409816,
  -0.019598906859755516,
  0.004763378761708736,
  -0.011470814235508442,
  -0.011831476353108883,
  0.011585170403122902,
  0.04394799470901489,
  0.002412477508187294,
  -0.024806514382362366,
  0.010300861671566963,
  0.005110845901072025,
  -0.023856477811932564,
  -0.004286161158233881,
  0.010371234267950058,
  -0.016915932297706604,
  0.019880399107933044,
  -0.022449016571044922,
  0.013476447202265263,
  0.017619663849473,
  0.011541186831891537,
  0.022343456745147705,
  -0.016924729570746422,
  -0.00826004147529602,
  0.0142857376486063,
  -0.02464817650616169,
  -0.0007757534040138125,
  -0.02931039407849312,
  -0.022308269515633583,
  0.0023443037644028664,
  0.009940199553966522,
  0.010511980392038822,
  -0.001803310471586883,
  0.005770593415945768,
  -0.006184035446494818,
  -0.0034900654572993517,
  0.014549637213349342,
  -0.004508276470005512,
  -0.0018747831927612424,
  0.018736835569143295,
  0.00851514469832182,
  0.0012920059962198138,
  0.0029798606410622597,
  -0.023029593750834465,
  -0.033849459141492844,
  0.006856978405267,
  0.008897798135876656,
  -0.00619723042473197,
  0.008783441968262196,
  -0.014294534921646118,
  -0.0142857376486063,
  -0.022888848558068275,
  -0.006615070626139641,
  0.008998959325253963,
  -0.0010176609503105283,
  0.017373356968164444,
  -0.008532737381756306,
  -0.012148154899477959,
  0.004154211841523647,
  -0.022888848558068275,
  -0.017584476619958878,
  -0.005511092953383923,
  0.00697133457288146,
  -0.05978194251656532,
  -0.005753000266849995,
  0.02593248523771763,
  -0.00773664191365242,
  -0.03613658249378204,
  -0.011787492781877518,
  -0.015332537703216076,
  0.003578031901270151,
  -0.0016284773591905832,
  0.025316720828413963,
  -0.008479957468807697,
  -0.024437056854367256,
  0.010397624224424362,
  0.05112605169415474,
  -0.018006715923547745,
  0.006892164703458548,
  -0.014857519418001175,
  0.03272348642349243,
  0.024296309798955917,
  -0.013898685574531555,
  -0.0016163819236680865,
  -0.00043708286830224097,
  -0.025985265150666237,
  -0.015561250038444996,
  -0.023944444954395294,
  0.018384970724582672,
  0.024577802047133446,
  0.0036088202614337206,
  -0.013045412488281727,
  0.020091518759727478,
  -0.019018327817320824,
  -0.02980300411581993,
  0.006355570163577795,
  -0.011365254409611225,
  0.022255491465330124,
  -0.02670658938586712,
  0.010080945678055286,
  -0.036453261971473694,
  -0.0062192222103476524,
  0.00826444011181593,
  -0.018543310463428497,
  0.017681239172816277,
  -0.020848028361797333,
  -0.04894448444247246,
  -0.017426136881113052,
  0.020003551617264748,
  0.022079557180404663,
  0.003098615212365985,
  -0.005933331325650215,
  0.017725223675370216,
  0.03277626633644104,
  0.03972560912370682,
  -0.015992285683751106,
  0.004547860939055681,
  -0.020971180871129036,
  -0.014144991524517536,
  -0.015297351405024529,
  -0.025263940915465355,
  -0.03610139712691307,
  0.04729072004556656,
  -0.002467456506565213,
  -0.037086620926856995,
  -0.03870519995689392,
  0.010608743876218796,
  -0.030735448002815247,
  0.009579537436366081,
  -0.003622015006840229,
  0.006412748247385025,
  -0.006984529551118612,
  0.020249856635928154,
  -0.03017246350646019,
  0.008919789455831051,
  0.007411166559904814,
  -0.01262317318469286,
  -0.015534860081970692,
  0.004547860939055681,
  0.0073671829886734486,
  0.030137276276946068,
  -0.009016552940011024,
  -0.010441607795655727,
  -0.01981002651154995,
  -0.015526063740253448,
  -0.024753736332058907,
  5.4257379815680906e-05,
  0.01863127574324608,
  -0.01820903830230236,
  -0.004110228270292282,
  0.003197577316313982,
  -0.011813882738351822,
  0.017223814502358437,
  -0.003076623659580946,
  0.037227366119623184,
  -0.0028808985371142626,
  0.008884603157639503,
  -0.005379143171012402,
  0.02109433524310589,
  0.0048337518237531185,
  0.014989469200372696,
  0.009588333778083324,
  0.002460859017446637,
  -0.009773063473403454,
  -0.001707647112198174,
  0.023891665041446686,
  -0.016308965161442757,
  -0.010608743876218796,
  -0.0010495487367734313,
  -0.0022035574074834585,
  -0.0002794306492432952,
  -0.022677728906273842,
  -0.0062148235738277435,
  0.009966589510440826,
  0.00366379925981164,
  -0.007912574335932732,
  0.0408867672085762,
  -0.004983294755220413,
  -0.005168023984879255,
  0.009931402280926704,
  -0.028008490800857544,
  ...],
 [0.017731226980686188,
  -0.005076537374407053,
  0.0035393729340285063,
  0.028076734393835068,
  0.030422572046518326,
  -0.010528775863349438,
  0.017951149493455887,
  0.024814555421471596,
  0.02102089673280716,
  0.029799457639455795,
  0.0007588461157865822,
  0.018317686393857002,
  -0.020251169800758362,
  -0.002234730403870344,
  -0.007701858878135681,
  -0.006845078431069851,
  -0.012324806302785873,
  0.01962805725634098,
  0.006386907305568457,
  -0.001644834759645164,
  0.04984903335571289,
  -0.022743621841073036,
  -0.006199057213962078,
  0.0475398488342762,
  -0.0030628747772425413,
  -0.013653503730893135,
  0.028351636603474617,
  0.03342817351222038,
  0.000502556620631367,
  0.000946123618632555,
  0.0029391685966402292,
  0.006396070588380098,
  0.001806340180337429,
  -0.008283736184239388,
  -0.07675284892320633,
  0.039659302681684494,
  0.023678289726376534,
  0.002833789214491844,
  -0.0462203174829483,
  0.008705253712832928,
  -0.02017786167562008,
  -0.039952535182237625,
  -0.04046568647027016,
  0.03896288201212883,
  0.021057551726698875,
  0.009447491727769375,
  0.051205217838287354,
  0.014844749122858047,
  -0.025309380143880844,
  -0.0036470431368798018,
  0.007706440519541502,
  0.01567861996591091,
  -0.0027444458100944757,
  0.015724437311291695,
  -0.02186393178999424,
  0.03223692998290062,
  0.009062627330422401,
  -0.0016001630574464798,
  -0.03700191155076027,
  0.009630760177969933,
  -0.013616849668323994,
  -0.03907284513115883,
  -0.02265198715031147,
  0.025969145819544792,
  -0.04739323630928993,
  -0.03234688937664032,
  0.01862007938325405,
  -0.010858658701181412,
  0.011445118114352226,
  0.06282444298267365,
  0.00811421312391758,
  -0.023788250982761383,
  -0.0010011042468249798,
  -0.0005755776655860245,
  -0.012425604276359081,
  -0.007798074744641781,
  0.027710197493433952,
  0.026335684582591057,
  -0.013653503730893135,
  0.04009914770722389,
  0.005832519847899675,
  -0.007449864409863949,
  5.798730126116425e-05,
  0.012599709443747997,
  0.009694904088973999,
  0.0052735512144863605,
  0.0036035168450325727,
  -0.0657200813293457,
  0.04299478977918625,
  0.05219487100839615,
  -0.033226579427719116,
  -0.00017596640100236982,
  -0.006272364407777786,
  0.004343463573604822,
  0.048419538885354996,
  -0.011875798925757408,
  -0.0004504396056290716,
  0.010437141172587872,
  -0.018400156870484352,
  0.036085568368434906,
  -0.0382298082113266,
  0.006593084428459406,
  -0.027160391211509705,
  -0.008352462202310562,
  0.0026253212708979845,
  -0.005585107486695051,
  -0.025492649525403976,
  0.0027169554959982634,
  0.012608872726559639,
  0.044937435537576675,
  -0.008993901312351227,
  -0.0016746158944442868,
  -0.001075556967407465,
  -0.017923658713698387,
  -0.009369602426886559,
  -0.0187850221991539,
  -7.005008956184611e-05,
  -0.0187300406396389,
  0.033061638474464417,
  0.02673887461423874,
  -0.0015532005345448852,
  -0.024319730699062347,
  0.013580195605754852,
  -0.026610586792230606,
  -0.0018040493596345186,
  0.0012026994954794645,
  0.012654690071940422,
  0.01272799726575613,
  0.003198035294190049,
  -0.009749884717166424,
  -0.02831498347222805,
  0.04570716619491577,
  0.007614806294441223,
  0.014075021259486675,
  0.0072161974385380745,
  0.028113387525081635,
  0.023971520364284515,
  0.000456166744697839,
  0.022908562794327736,
  0.011729184538125992,
  -0.03133891522884369,
  0.009511635638773441,
  -0.012425604276359081,
  -0.014734787866473198,
  -0.023091832175850868,
  -0.004939086269587278,
  0.027600236237049103,
  0.02499782294034958,
  -0.005974553059786558,
  0.002114460337907076,
  0.019499769434332848,
  -0.038376424461603165,
  0.017383016645908356,
  0.013910079374909401,
  -0.0053239502012729645,
  0.023513348773121834,
  0.014881402254104614,
  -0.0003868683415930718,
  0.006492286454886198,
  -0.022523699328303337,
  -0.01482642162591219,
  -0.013094534166157246,
  0.003296542214229703,
  -0.017923658713698387,
  -0.004013580270111561,
  -0.030367590487003326,
  0.03280506283044815,
  -0.01539455447345972,
  -0.027398640289902687,
  0.06205471605062485,
  0.001533728325739503,
  0.004966576583683491,
  -0.003823439124971628,
  -0.012489748187363148,
  -0.008673181757330894,
  0.014863075688481331,
  -0.02809506095945835,
  -0.005960808135569096,
  -0.01497303694486618,
  -0.009520798921585083,
  -0.044204361736774445,
  0.003353813663125038,
  0.00875107105821371,
  -0.02158902958035469,
  -0.013497725129127502,
  0.024851208552718163,
  0.03199867904186249,
  0.00897099357098341,
  -0.05014226213097572,
  -0.016081811860203743,
  -0.012709670700132847,
  -0.03326323255896568,
  0.03348315507173538,
  0.01474395114928484,
  0.015431207604706287,
  -0.01042797788977623,
  -0.012434767559170723,
  0.012434767559170723,
  -0.044607553631067276,
  -0.01979299820959568,
  0.014643153175711632,
  -0.03797323256731033,
  -0.011481771245598793,
  0.012388951145112514,
  -0.015861889347434044,
  -0.021204166114330292,
  -0.014166655018925667,
  -0.0013504597591236234,
  0.012233172543346882,
  0.004096051212400198,
  0.023128485307097435,
  0.019811324775218964,
  -0.00902597326785326,
  -0.016375040635466576,
  0.014203309081494808,
  0.022615334019064903,
  -0.041125450283288956,
  -0.014313270337879658,
  -0.04130871966481209,
  -0.022303776815533638,
  0.014890565536916256,
  -0.02785681188106537,
  0.0012702797539532185,
  0.016897356137633324,
  -0.012773814611136913,
  -0.008237918838858604,
  0.029121365398168564,
  -0.006881732027977705,
  -0.004595457576215267,
  0.011133561842143536,
  -0.013635176233947277,
  0.0017181421862915158,
  0.0023149102926254272,
  -0.0009472690871916711,
  0.0028727338649332523,
  0.011857472360134125,
  -0.014762277714908123,
  0.018033619970083237,
  0.01249891147017479,
  0.02153404988348484,
  -0.046110354363918304,
  0.002051461720839143,
  -0.0647304356098175,
  -0.034674402326345444,
  0.003078910754993558,
  0.003674533450976014,
  -0.0018292487366124988,
  -0.006006625015288591,
  0.02124081924557686,
  -0.015073833987116814,
  0.004984903149306774,
  -0.010730370879173279,
  -0.02298187091946602,
  0.0371851809322834,
  -0.006116586271673441,
  0.0008418897050432861,
  0.015651129186153412,
  -0.01505550742149353,
  -0.046953391283750534,
  -0.008554057218134403,
  0.015339573845267296,
  -0.0007788911461830139,
  -0.004469460807740688,
  -0.00436408119276166,
  0.012508075684309006,
  -0.02607910707592964,
  -0.021882258355617523,
  -0.040392376482486725,
  0.015889378264546394,
  -0.057839538902044296,
  0.03269509971141815,
  -0.014221635647118092,
  0.00810046773403883,
  0.0071474714204669,
  0.014239962212741375,
  -0.021570703014731407,
  -0.03538914769887924,
  0.008201265707612038,
  -0.009401674382388592,
  0.01755712181329727,
  -0.040612298995256424,
  0.006588502787053585,
  0.020031247287988663,
  -0.041235413402318954,
  -0.004313682205975056,
  0.011655876412987709,
  -0.017877843230962753,
  -0.0010938838822767138,
  -0.03172377869486809,
  0.0035233369562774897,
  0.0042472477070987225,
  0.001700960798189044,
  0.008228755556046963,
  -0.027600236237049103,
  -0.004128123167902231,
  0.030184322968125343,
  0.04882273077964783,
  -0.004359499551355839,
  -0.006574757397174835,
  0.01889498345553875,
  0.054467398673295975,
  -0.028516579419374466,
  0.03033093735575676,
  -0.012911265715956688,
  0.04372786357998848,
  0.035920627415180206,
  0.0007972180028446019,
  -0.004634402226656675,
  -0.021662337705492973,
  -0.02618906833231449,
  0.014368250966072083,
  -0.049812380224466324,
  -0.026482298970222473,
  -0.07037510722875595,
  -0.02276194840669632,
  -0.0185376089066267,
  -0.01558698620647192,
  -0.00884728692471981,
  0.013195332139730453,
  -0.0007405193173326552,
  -0.0007771729724481702,
  0.0070512555539608,
  -0.03527918830513954,
  0.004054815508425236,
  -0.005021556746214628,
  0.0372951403260231,
  -0.030880741775035858,
  0.0033790129236876965,
  0.026427317410707474,
  -0.0038371842820197344,
  -0.016292570158839226,
  -0.02080097422003746,
  -0.00189224723726511,
  0.009795701131224632,
  -0.0002816321502905339,
  -0.013552705757319927,
  -0.0018475755350664258,
  0.02202887460589409,
  -0.01794198714196682,
  -0.003857801901176572,
  -0.006583921145647764,
  -0.0008647982613183558,
  -0.009749884717166424,
  -0.019866306334733963,
  -0.003779912833124399,
  0.016356714069843292,
  0.042811520397663116,
  -0.028553232550621033,
  0.004524440970271826,
  0.005860010161995888,
  0.00038400478661060333,
  -0.013094534166157246,
  0.010473795235157013,
  -0.008462423458695412,
  0.0285898856818676,
  -0.023531675338745117,
  -0.009177170693874359,
  -0.03841307759284973,
  -0.00789429061114788,
  0.005195661913603544,
  0.005337695125490427,
  0.00420601200312376,
  0.03163214400410652,
  -0.008398279547691345,
  0.04233502224087715,
  0.00769727723672986,
  0.012013250030577183,
  0.017300546169281006,
  -0.00834329891949892,
  0.03469272702932358,
  -0.002312619471922517,
  -0.0010921657085418701,
  -0.001508528832346201,
  -0.007349066901952028,
  -0.0063914889469742775,
  0.01682404801249504,
  0.0017490687314420938,
  -0.005731722339987755,
  -0.006290691439062357,
  0.010290526784956455,
  -0.009960643015801907,
  -0.00863194651901722,
  -0.018427647650241852,
  -0.06062522158026695,
  -0.03236521780490875,
  -0.027783505618572235,
  -0.0010681117419153452,
  -0.00811421312391758,
  0.005204825662076473,
  0.029139691963791847,
  -0.005026138853281736,
  -0.0279850997030735,
  -0.005411002319306135,
  -0.006593084428459406,
  0.004487787373363972,
  -0.007133726496249437,
  0.04237167909741402,
  0.01922486536204815,
  0.0017696864670142531,
  -0.019829651340842247,
  0.02595081925392151,
  -0.0288098081946373,
  0.014157491736114025,
  -0.012242335826158524,
  -0.007179543375968933,
  0.011234358884394169,
  -0.01649416610598564,
  -0.010812841355800629,
  0.027490274980664253,
  0.01573360152542591,
  -0.04035572335124016,
  -0.003928818739950657,
  -0.016512492671608925,
  -0.020196188241243362,
  -0.020049573853611946,
  -0.011481771245598793,
  0.06025868281722069,
  0.02433805726468563,
  -0.0022473300341516733,
  -0.04273821413516998,
  0.03579233959317207,
  0.015220449306070805,
  -0.031760431826114655,
  -0.0004962567472830415,
  -0.028516579419374466,
  -0.008164611645042896,
  0.011014437302947044,
  0.014560682699084282,
  -0.010776188224554062,
  -0.030587512999773026,
  -0.013314456678926945,
  -0.00779349310323596,
  0.05270802229642868,
  0.04519401118159294,
  0.03733179345726967,
  -0.030312610790133476,
  0.044204361736774445,
  -0.044094402343034744,
  0.00205260724760592,
  -0.005337695125490427,
  -0.013461071066558361,
  -0.011536751873791218,
  0.004540476948022842,
  0.014487375505268574,
  -0.0657200813293457,
  0.021094204857945442,
  -0.008013415150344372,
  -0.03309829160571098,
  -1.5508023352595046e-05,
  0.006835915148258209,
  -0.022285450249910355,
  -0.007010020315647125,
  -0.018390994518995285,
  -0.00831580813974142,
  -0.005795866250991821,
  -0.02635401114821434,
  -0.015348737128078938,
  0.06861572712659836,
  -0.005085701122879982,
  -0.0023893630132079124,
  -0.0017971767811104655,
  0.02213883586227894,
  0.01962805725634098,
  -0.029964400455355644,
  -0.014340760186314583,
  -0.017190584912896156,
  -0.03242019936442375,
  0.02943292073905468,
  0.0072940862737596035,
  -0.0014638571301475167,
  0.008888522163033485,
  -0.023898212239146233,
  -0.04013580083847046,
  -0.00044356705620884895,
  -0.0007972180028446019,
  0.034179575741291046,
  -0.05450405180454254,
  -0.0285898856818676,
  -0.011325993575155735,
  -0.034069616347551346,
  -0.03245685249567032,
  -0.0025886676739901304,
  -0.017923658713698387,
  0.04431432485580444,
  0.03874295949935913,
  -0.0021430959459394217,
  0.04336132854223251,
  0.01927984692156315,
  -0.011564242653548717,
  0.0239898469299078,
  0.022578679025173187,
  -0.00672595389187336,
  -0.00012062789755873382,
  -0.001292042899876833,
  -0.020544398576021194,
  -0.002357291057705879,
  -0.007028346881270409,
  -0.042921483516693115,
  0.01884000189602375,
  0.06744281202554703,
  0.019976267591118813,
  0.0007284922758117318,
  -0.006698463577777147,
  -0.006272364407777786,
  0.0036012260243296623,
  -0.03238354250788689,
  -0.010134748183190823,
  0.006941294297575951,
  -0.017190584912896156,
  0.006194475572556257,
  0.02164400927722454,
  -0.00447862409055233,
  -0.00012120061728637666,
  -0.004384698811918497,
  0.014808095060288906,
  0.015321246348321438,
  -0.00021147467487026006,
  -0.003099528606981039,
  -0.018821675330400467,
  0.006134913302958012,
  8.182652527466416e-05,
  -0.0012794432695955038,
  -0.025364359840750694,
  -0.007252851035445929,
  0.005543872248381376,
  0.0016654524952173233,
  0.03313494473695755,
  -0.012517238967120647,
  0.10314351320266724,
  -0.016265079379081726,
  -0.0233300793915987,
  0.03694692999124527,
  0.0004822252376470715,
  0.021790625527501106,
  -0.01802445761859417,
  -0.002218694193288684,
  -0.016952335834503174,
  0.004819961730390787,
  0.008792306296527386,
  0.023293426260352135,
  -0.003871547058224678,
  0.017080623656511307,
  0.025309380143880844,
  -0.03199867904186249,
  -0.02215716242790222,
  0.018281033262610435,
  0.007477354723960161,
  -0.021387433633208275,
  0.01827186904847622,
  -0.004746654070913792,
  0.039439380168914795,
  0.018986616283655167,
  0.0014924928545951843,
  0.005685905460268259,
  0.0038738378789275885,
  0.01370848435908556,
  0.004279319662600756,
  0.002623030450195074,
  -0.033611442893743515,
  -0.014359086751937866,
  0.004723745863884687,
  0.006396070588380098,
  0.03711187094449997,
  0.017877843230962753,
  -0.012746324762701988,
  0.0369102768599987,
  0.005511800292879343,
  0.008654855191707611,
  -0.019866306334733963,
  -0.014459884725511074,
  -0.004556512925773859,
  -0.001292042899876833,
  0.012205682694911957,
  -0.03139389306306839,
  -0.009768211282789707,
  0.02254202589392662,
  0.03196202591061592,
  0.0002946613822132349,
  0.0033858856186270714,
  0.011756674386560917,
  -0.004572548903524876,
  -0.009740720503032207,
  0.017245566472411156,
  -0.012874612584710121,
  0.005131518002599478,
  0.011939942836761475,
  -0.0185925904661417,
  -0.0385596938431263,
  -0.031320586800575256,
  0.016466675326228142,
  -0.002696337876841426,
  0.031027358025312424,
  0.03559074178338051,
  0.006354835350066423,
  0.027508601546287537,
  0.0020995698869228363,
  -0.03586564585566521,
  -0.005681323353201151,
  -0.010299690067768097,
  -0.018436811864376068,
  0.03676366060972214,
  0.021332453936338425,
  0.025877512991428375,
  -0.002645938890054822,
  -0.014670643955469131,
  0.06575673818588257,
  0.03936607390642166,
  -0.01774955354630947,
  0.009080953896045685,
  -0.029084710404276848,
  0.0380098856985569,
  -0.016411693766713142,
  -0.01399254985153675,
  0.0028131715953350067,
  0.0141025111079216,
  0.03637879714369774,
  0.012975409626960754,
  -0.01674157753586769,
  -0.043471287935972214,
  0.011225195601582527,
  0.0025084875524044037,
  -0.029414594173431396,
  0.006721372250467539,
  -0.0004816525324713439,
  -0.013735974207520485,
  0.023311752825975418,
  -0.001837266725488007,
  -0.016732413321733475,
  -0.00013294124801177531,
  0.04053899273276329,
  -0.0068267518654465675,
  -0.011903288774192333,
  -0.0031751268543303013,
  -0.02853490598499775,
  0.046733468770980835,
  -0.0005804456886835396,
  0.006107422988861799,
  0.021002570167183876,
  0.01587105169892311,
  0.028223348781466484,
  0.028351636603474617,
  -0.003278215415775776,
  0.028663193807005882,
  -0.07037510722875595,
  -0.051388487219810486,
  0.0072940862737596035,
  -0.01665910705924034,
  0.018858328461647034,
  -0.005589689128100872,
  0.026940470561385155,
  0.013937569223344326,
  0.01827186904847622,
  0.011976596899330616,
  0.03298833221197128,
  0.03126560524106026,
  -0.021552376449108124,
  0.007830146700143814,
  -0.02494284324347973,
  -0.03373973071575165,
  0.004432806745171547,
  -0.011325993575155735,
  -0.030514204874634743,
  -0.007404047530144453,
  -0.034289535135030746,
  0.007559825666248798,
  0.028076734393835068,
  -2.0313452750997385e-06,
  -0.012214845977723598,
  0.0022461845073848963,
  0.04871276766061783,
  0.006611410994082689,
  -0.02175397053360939,
  -0.03493097797036171,
  -0.00013959905481897295,
  0.007500263396650553,
  0.0053606037981808186,
  0.0043686628341674805,
  0.004897850565612316,
  -0.006758025847375393,
  0.013433581218123436,
  0.01658579893410206,
  0.002572631696239114,
  -0.004581712651997805,
  0.001886520185507834,
  0.014047530479729176,
  0.006570175755769014,
  -0.023128485307097435,
  0.015229612588882446,
  -0.009649086743593216,
  -0.0004412761772982776,
  -0.024979496374726295,
  -0.008540311828255653,
  0.035462453961372375,
  0.014963873662054539,
  -0.006845078431069851,
  0.002920841798186302,
  0.03324490785598755,
  -0.009768211282789707,
  -0.015412881039083004,
  0.01676906831562519,
  0.001429494353942573,
  -0.006441887933760881,
  -0.0006036406266503036,
  -0.012526402249932289,
  0.04999564588069916,
  0.004102923441678286,
  0.019298173487186432,
  -0.007541498634964228,
  -0.014734787866473198,
  -0.045890431851148605,
  -0.011765837669372559,
  0.03918280452489853,
  0.015651129186153412,
  0.037368446588516235,
  0.01084949541836977,
  -0.005612597800791264,
  0.0033309049904346466,
  -0.024008173495531082,
  0.00421975739300251,
  0.015412881039083004,
  -0.004907014314085245,
  0.017163095995783806,
  -0.018913310021162033,
  0.036800313740968704,
  -0.011701693758368492,
  -0.016411693766713142,
  0.03672700747847557,
  0.03207198902964592,
  0.0187025498598814,
  -0.0023366734385490417,
  -0.01379095483571291,
  -0.002900223946198821,
  0.02703210338950157,
  -0.02719704620540142,
  -0.017053134739398956,
  0.04424101486802101,
  -0.014368250966072083,
  0.02785681188106537,
  -0.008324971422553062,
  0.023733271285891533,
  0.0012473711976781487,
  -0.010592919774353504,
  -0.017657920718193054,
  0.010986946523189545,
  -0.02965284325182438,
  0.005814193282276392,
  -0.011976596899330616,
  -0.007060418836772442,
  0.010409651324152946,
  0.016833212226629257,
  0.027068758383393288,
  -0.010134748183190823,
  0.012792141176760197,
  -0.060588568449020386,
  -0.008260827511548996,
  0.05226817727088928,
  0.00997896958142519,
  0.03353813663125038,
  -0.025712570175528526,
  0.04405774921178818,
  0.01082200463861227,
  0.02164400927722454,
  0.0073032500222325325,
  0.005365185439586639,
  0.011518425308167934,
  -0.013231986202299595,
  0.012828795239329338,
  -0.01984797790646553,
  0.024356383830308914,
  0.006803843192756176,
  0.0013470235280692577,
  -0.014258289709687233,
  0.01995793916285038,
  -0.007843892090022564,
  0.009103862568736076,
  0.014927219599485397,
  0.0039356909692287445,
  -0.02921300008893013,
  -0.03603058680891991,
  0.007032928988337517,
  -0.00978653784841299,
  0.014221635647118092,
  -0.03980591893196106,
  -0.00494824955239892,
  0.01609097421169281,
  -0.03463774546980858,
  -0.008705253712832928,
  -0.033061638474464417,
  0.0078072380274534225,
  0.01558698620647192,
  0.006331926677376032,
  -0.010831168852746487,
  0.0663798525929451,
  0.025309380143880844,
  0.013800118118524551,
  0.008865613490343094,
  0.0377899669110775,
  -0.016759904101490974,
  -0.012718833982944489,
  -0.029799457639455795,
  0.029506228864192963,
  0.00038429113919846714,
  -0.0003582326462492347,
  0.02118583954870701,
  -0.00894350279122591,
  0.02298187091946602,
  -0.026665566489100456,
  -0.009539125487208366,
  -0.03936607390642166,
  0.006359416991472244,
  -0.014597335830330849,
  -0.003420248394832015,
  0.0235500019043684,
  0.015403717756271362,
  -0.0020308441016823053,
  0.00599746173247695,
  -0.001799467601813376,
  -0.05109525844454765,
  -0.016842374578118324,
  0.02703210338950157,
  0.010189728811383247,
  -0.0037318048998713493,
  0.007344485260546207,
  -0.0015428917249664664,
  -0.010840332135558128,
  0.006272364407777786,
  0.0563000850379467,
  0.03483934327960014,
  -0.009186333976686,
  -0.014111674390733242,
  -0.03342817351222038,
  0.03269509971141815,
  -0.025034477934241295,
  0.0469900444149971,
  0.007426956202834845,
  0.019371479749679565,
  -0.036195527762174606,
  0.011096907779574394,
  0.01567861996591091,
  0.030862415209412575,
  0.014047530479729176,
  -0.010702880099415779,
  -0.01219651848077774,
  0.012214845977723598,
  -0.0017467779107391834,
  -0.03441782295703888,
  -0.025547629222273827,
  -0.007523172069340944,
  0.00811421312391758,
  -0.004070851486176252,
  0.0032736335415393114,
  0.010070604272186756,
  0.041968487203121185,
  0.0187300406396389,
  -0.008824378252029419,
  0.021552376449108124,
  -0.022853583097457886,
  0.017767881974577904,
  0.0029277142602950335,
  0.011417627334594727,
  -0.0024764155969023705,
  -0.01040048711001873,
  -0.010858658701181412,
  -0.028461597859859467,
  0.011500098742544651,
  -0.0005377784837037325,
  -0.0017891587922349572,
  -0.027710197493433952,
  0.007940107956528664,
  -0.016943173483014107,
  -0.03852304071187973,
  -0.022505372762680054,
  -0.013323619961738586,
  -0.00441677076742053,
  0.027893466874957085,
  0.010024786926805973,
  0.013882588595151901,
  0.00515900831669569,
  -0.015312083065509796,
  0.0014856202760711312,
  0.003931109327822924,
  0.017172258347272873,
  -0.03218194842338562,
  -0.02752692811191082,
  0.03342817351222038,
  -0.0048382882960140705,
  -0.04827292263507843,
  0.02499782294034958,
  -0.0038623837754130363,
  0.018555935472249985,
  -0.0013584777479991317,
  0.01995793916285038,
  0.009905662387609482,
  -0.02265198715031147,
  0.0026024128310382366,
  0.024851208552718163,
  0.01040048711001873,
  -0.0012336261570453644,
  -0.0017880132654681802,
  0.009878172539174557,
  0.008718999102711678,
  -0.012902102433145046,
  -0.0007221924606710672,
  0.0003035384579561651,
  -0.01598101295530796,
  -0.008297481574118137,
  -0.0007330740336328745,
  0.05567697063088417,
  0.005947062745690346,
  0.05410086363554001,
  0.03245685249567032,
  -0.019481441006064415,
  0.0021052968222647905,
  -0.04035572335124016,
  0.03095404990017414,
  -0.033501483500003815,
  -0.02629902958869934,
  0.008485331200063229,
  -0.0032003261148929596,
  -0.01535790041089058,
  -0.012847121804952621,
  -0.005039883777499199,
  -0.009145097807049751,
  0.00758273433893919,
  0.004517568740993738,
  -0.0009003065060824156,
  -0.005346858408302069,
  0.027343660593032837,
  0.010327179916203022,
  0.036360472440719604,
  -0.012104884721338749,
  0.04204179346561432,
  0.04574381932616234,
  0.0007393738487735391,
  0.0012645527021959424,
  -0.013131188228726387,
  -0.00955745205283165,
  -0.0009730412275530398,
  0.00426786532625556,
  -0.007798074744641781,
  0.006900059059262276,
  0.05281798169016838,
  0.006602247711271048,
  -0.012856285087764263,
  -0.011399300768971443,
  0.005136099644005299,
  -0.011619223281741142,
  -0.0024214349687099457,
  -0.0236049834638834,
  0.007523172069340944,
  -0.033061638474464417,
  0.018418485298752785,
  -0.02314681187272072,
  -0.005992880091071129,
  0.049299225211143494,
  -0.0034362843725830317,
  -0.02063603326678276,
  0.017437998205423355,
  0.024246422573924065,
  0.006835915148258209,
  0.007463609799742699,
  -0.016375040635466576,
  -0.022963542491197586,
  -0.010446304455399513,
  -0.017621265724301338,
  0.04160194844007492,
  0.010776188224554062,
  0.01253556553274393,
  0.009158843196928501,
  0.026830509305000305,
  -0.02028782293200493,
  0.011701693758368492,
  0.006441887933760881,
  0.04834623262286186,
  0.029982727020978928,
  0.003933400381356478,
  -0.023696618154644966,
  0.013864262029528618,
  -0.012288153171539307,
  0.03038591705262661,
  0.025254400447010994,
  0.011344320140779018,
  0.021295800805091858,
  0.0119949234649539,
  0.02439303696155548,
  -0.017245566472411156,
  -0.002838370855897665,
  -0.03095404990017414,
  0.015156305395066738,
  -0.018088601529598236,
  -0.009484144859015942,
  -0.007820983417332172,
  -0.006625156383961439,
  -0.04966576397418976,
  0.009676576592028141,
  0.006872568745166063,
  0.02461295947432518,
  -0.0009243605309166014,
  0.006281527690589428,
  0.0009976678993552923,
  ...],
 [-0.027602434158325195,
  -0.02202529087662697,
  0.002326022833585739,
  0.01597895845770836,
  0.025548627600073814,
  -0.003600798547267914,
  0.0017926532309502363,
  0.030488383024930954,
  0.0029789027757942677,
  0.06812967360019684,
  -0.03183398023247719,
  0.022928256541490555,
  -0.027903422713279724,
  0.030205100774765015,
  -0.022343983873724937,
  0.028169000521302223,
  -0.0029833291191607714,
  0.011313633993268013,
  -0.010844445787370205,
  -0.002992181805893779,
  0.025283049792051315,
  0.009428028017282486,
  -0.01551862247288227,
  0.02418532595038414,
  -0.02152954414486885,
  0.018218668177723885,
  0.029585417360067368,
  0.0011785035021603107,
  0.020449526607990265,
  -0.03927017003297806,
  -0.03909311816096306,
  0.03565830737352371,
  0.035127151757478714,
  -0.015970105305314064,
  -0.030346741899847984,
  0.029656238853931427,
  0.01812128908932209,
  0.01903310790657997,
  0.012358241714537144,
  -0.003501206636428833,
  -0.0034148937556892633,
  -0.014739593490958214,
  -0.06770475208759308,
  0.023176129907369614,
  -0.006090594921261072,
  -0.009065071120858192,
  0.00655978312715888,
  -0.006351746618747711,
  0.03183398023247719,
  0.009976889938116074,
  0.017253734171390533,
  0.008923429064452648,
  0.009277533739805222,
  0.04985789209604263,
  -0.006493388209491968,
  0.01905081421136856,
  0.024999765679240227,
  0.022822024300694466,
  0.016217978671193123,
  -0.01104805525392294,
  -0.020237063989043236,
  0.0025893880520015955,
  0.009047365747392178,
  -7.663039286853746e-05,
  0.002715537790209055,
  -0.035445842891931534,
  0.04369647428393364,
  0.017917679622769356,
  -0.009720164351165295,
  0.06550930440425873,
  -0.0035698143765330315,
  0.014996319077908993,
  0.01583731733262539,
  0.02161807008087635,
  0.012482178397476673,
  -0.011703148484230042,
  0.016793398186564445,
  0.027177508920431137,
  -0.014836971648037434,
  0.03340974450111389,
  0.012597261928021908,
  -0.0013234650250524282,
  -0.004309007432311773,
  0.02671717293560505,
  0.020591167733073235,
  0.020237063989043236,
  -0.0060286265797913074,
  -0.0717061311006546,
  0.028452284634113312,
  0.04383811727166176,
  -0.005125660449266434,
  -0.003711456200107932,
  -0.029780175536870956,
  -0.010932971723377705,
  0.025176819413900375,
  0.018537363037467003,
  0.011304780840873718,
  -0.00399916572496295,
  0.008728671818971634,
  0.03873901441693306,
  0.01908622495830059,
  -0.006086168345063925,
  -0.018767530098557472,
  -0.030630026012659073,
  0.013031039386987686,
  -0.018625888973474503,
  0.03124970756471157,
  0.04150103032588959,
  0.027956537902355194,
  0.019068518653512,
  -0.01892687752842903,
  -0.01717406138777733,
  -0.018572771921753883,
  -0.027832601219415665,
  -0.032400548458099365,
  -0.028487693518400192,
  -0.013633017428219318,
  -0.016200274229049683,
  0.000939483055844903,
  -0.022981371730566025,
  -0.06547389179468155,
  -0.020697398111224174,
  0.016722576692700386,
  -0.017147501930594444,
  -0.024627957493066788,
  0.004198349546641111,
  -0.03895147889852524,
  -0.0237249918282032,
  0.004083265550434589,
  -0.030435267835855484,
  -0.04656472057104111,
  0.04596274346113205,
  0.02438008412718773,
  -0.0026425037067383528,
  0.022255457937717438,
  0.007985052652657032,
  0.02662864699959755,
  0.016315357759594917,
  -0.010021152906119823,
  0.03100183606147766,
  -0.05594848468899727,
  0.004643192980438471,
  0.020679693669080734,
  0.010260173119604588,
  -0.047343749552965164,
  -0.04390893876552582,
  0.037499651312828064,
  0.01945803314447403,
  0.0131372706964612,
  0.030895603820681572,
  0.0008957733516581357,
  -0.032471369951963425,
  0.015633706003427505,
  -0.011269371025264263,
  0.00028273017960600555,
  -0.007285696920007467,
  0.010127384215593338,
  0.03149758279323578,
  0.0004910431453026831,
  -0.047202110290527344,
  -0.019794432446360588,
  0.02901885099709034,
  -0.02368958108127117,
  0.009259828366339207,
  -0.007068807724863291,
  0.014084500260651112,
  0.05477994307875633,
  0.03325039893388748,
  -0.0666070282459259,
  0.053965501487255096,
  -0.0007685170858167112,
  0.032772358506917953,
  0.001623347052372992,
  -0.014898939989507198,
  -0.014836971648037434,
  -0.03144446387887001,
  0.014535983093082905,
  -0.007808000780642033,
  0.025070587173104286,
  0.0006285352283157408,
  -0.051699232310056686,
  0.030045753344893456,
  -0.029886405915021896,
  0.0031980047933757305,
  -0.03415336459875107,
  0.01456254068762064,
  0.03211726248264313,
  0.029355250298976898,
  -0.012508735992014408,
  -0.07021889090538025,
  -0.01078247744590044,
  -0.02400827407836914,
  0.04291744530200958,
  -0.02131708152592182,
  0.022237753495573997,
  -0.0160851888358593,
  -0.03387007862329483,
  -0.012296273373067379,
  0.01095067709684372,
  -0.008839329704642296,
  -0.007697343360632658,
  -0.02894802950322628,
  0.007838984951376915,
  0.008764082565903664,
  0.007139628753066063,
  0.017545869573950768,
  -0.015226487070322037,
  0.011198549531400204,
  0.014819266274571419,
  0.020768219605088234,
  0.03388778492808342,
  0.0007231474737636745,
  0.035233382135629654,
  -0.006511093582957983,
  0.0159966628998518,
  0.014828119426965714,
  0.0016565443947911263,
  -0.009428028017282486,
  -0.010667392984032631,
  -0.010844445787370205,
  0.006887329276651144,
  0.0015381406992673874,
  0.01808588020503521,
  -0.009516553953289986,
  -0.001600109040737152,
  0.0055638644844293594,
  0.054355017840862274,
  0.02629224769771099,
  0.011614622548222542,
  0.011499538086354733,
  -0.005001723766326904,
  0.0342773012816906,
  0.017758332192897797,
  -0.032418251037597656,
  -0.049893300980329514,
  0.020750515162944794,
  0.01719176582992077,
  0.009250976145267487,
  0.020272472873330116,
  0.004780408460646868,
  -0.004550240933895111,
  0.039624277502298355,
  -0.03647274896502495,
  -0.010809035040438175,
  -0.004014657810330391,
  -0.002441106829792261,
  -0.008029315620660782,
  -0.04295285791158676,
  0.015863874927163124,
  -0.020060010254383087,
  -0.013234649784862995,
  -0.037251777946949005,
  0.016757987439632416,
  -0.030205100774765015,
  -0.032843176275491714,
  0.00422269431874156,
  0.02650471031665802,
  0.0060330526903271675,
  -0.03615405410528183,
  -0.00027664401568472385,
  0.03194021061062813,
  -0.0070112659595906734,
  -0.016598641872406006,
  0.014730740338563919,
  0.007024544756859541,
  -0.017200618982315063,
  0.00948114413768053,
  -0.022662678733468056,
  0.0003875782713294029,
  0.0007259139092639089,
  0.009286385960876942,
  -0.0049574607983231544,
  0.01083559263497591,
  -0.027814896777272224,
  0.039588864892721176,
  -0.008648998104035854,
  -0.029408365488052368,
  -0.023211538791656494,
  0.0311965923756361,
  0.018023911863565445,
  0.001816997886635363,
  0.01883835159242153,
  -0.01965279132127762,
  -0.012535293586552143,
  0.03654356673359871,
  -0.005692227277904749,
  -0.00224524294026196,
  -0.003926131874322891,
  -0.008662277832627296,
  -0.03668520972132683,
  -0.009100481867790222,
  -0.007267991546541452,
  0.026062078773975372,
  -0.02126396633684635,
  0.037464238703250885,
  0.01226086262613535,
  0.020272472873330116,
  -0.03788916394114494,
  0.016899630427360535,
  0.00777259049937129,
  0.009109334088861942,
  0.006798803340643644,
  -0.020024601370096207,
  -0.030222805216908455,
  0.014270405285060406,
  -0.006825361400842667,
  0.00014634468243457377,
  0.007608816958963871,
  -0.037959985435009,
  -0.0172271765768528,
  0.021051503717899323,
  -0.04171349108219147,
  -0.0073653701692819595,
  -0.027673253789544106,
  -0.01792653277516365,
  -0.018820645287632942,
  -0.003941623959690332,
  0.008405552245676517,
  0.013500227592885494,
  -0.025229934602975845,
  -0.016775693744421005,
  -0.0060551841743290424,
  -0.05329270288348198,
  -0.016563231125473976,
  0.03682685270905495,
  0.0029811158310621977,
  -0.015368128195405006,
  0.005218612961471081,
  0.016262242570519447,
  -0.009321796707808971,
  -0.06402206420898438,
  0.00656420923769474,
  -0.025460101664066315,
  -0.029532302170991898,
  0.01967049576342106,
  0.007037824019789696,
  -0.0074804541654884815,
  -0.018732119351625443,
  -0.038668192923069,
  -0.013579901307821274,
  0.030488383024930954,
  0.024787303060293198,
  0.01959967613220215,
  0.022839730605483055,
  -0.017377670854330063,
  -0.0004003038920927793,
  0.015288455411791801,
  0.00023708392109256238,
  0.020520346239209175,
  0.01464221440255642,
  0.0003189705603290349,
  -0.0239020437002182,
  0.049255914986133575,
  0.012101516127586365,
  0.0237249918282032,
  -0.010189352557063103,
  -0.020573461428284645,
  -0.022680383175611496,
  0.01321694441139698,
  0.005794032476842403,
  0.011065760627388954,
  0.009401470422744751,
  -0.0021257326006889343,
  0.010224763303995132,
  -0.025300756096839905,
  -0.012482178397476673,
  0.023123012855648994,
  0.028275230899453163,
  -0.019794432446360588,
  0.032471369951963425,
  -0.029992638155817986,
  0.005749769043177366,
  0.032754652202129364,
  0.007808000780642033,
  0.009295239113271236,
  -0.027230624109506607,
  0.028275230899453163,
  -0.010136236436665058,
  0.01903310790657997,
  0.0239020437002182,
  -0.011986431665718555,
  0.0005502449348568916,
  0.020272472873330116,
  -0.024875830858945847,
  0.018590478226542473,
  0.01095952931791544,
  -0.005639111623167992,
  -0.02629224769771099,
  -0.022521035745739937,
  0.012402504682540894,
  0.00047665764577686787,
  -0.008706540800631046,
  0.025637153536081314,
  0.01192446332424879,
  0.015376981347799301,
  -0.007971773855388165,
  -0.029656238853931427,
  0.001699700835160911,
  0.015058287419378757,
  -0.03703931346535683,
  0.0007430658442899585,
  -0.06657161563634872,
  -0.003558748634532094,
  -0.015173370949923992,
  -0.012482178397476673,
  -0.006617324892431498,
  0.00163994578178972,
  -0.03912853077054024,
  0.017988501116633415,
  -0.021016092970967293,
  -0.018413426354527473,
  -0.0017074468778446317,
  -0.006683719344437122,
  -0.032329726964235306,
  0.032701537013053894,
  -0.012190042063593864,
  0.04783949628472328,
  -0.01079132966697216,
  -0.019068518653512,
  -0.03387007862329483,
  0.003419320099055767,
  -0.0025672565679997206,
  0.008011611178517342,
  0.008516209200024605,
  0.0022817598655819893,
  0.06218072399497032,
  0.04628143832087517,
  0.022326279431581497,
  -0.012703493237495422,
  -0.03909311816096306,
  -0.007639801129698753,
  -0.0030143132898956537,
  -0.0042182682082057,
  0.025991259142756462,
  -0.013057597912847996,
  -0.015421244315803051,
  0.016094041988253593,
  0.0009167982498183846,
  -0.006130431313067675,
  0.015899285674095154,
  0.002270694123581052,
  0.017377670854330063,
  0.008728671818971634,
  -0.0012083810288459063,
  -0.03335662931203842,
  -0.011827085167169571,
  0.009764427319169044,
  -0.05134512856602669,
  -0.0013024400686845183,
  0.021193144842982292,
  -0.008826050907373428,
  -0.015244191512465477,
  0.001915483153425157,
  -0.014491720125079155,
  -0.007396354340016842,
  -0.01099494006484747,
  -0.015137960202991962,
  0.03424188867211342,
  -0.02177741751074791,
  0.0026934060733765364,
  0.005054839421063662,
  0.04769785329699516,
  0.006586340721696615,
  -0.04281121492385864,
  0.006882903166115284,
  -0.023300066590309143,
  -0.011765116825699806,
  0.03399401530623436,
  -0.0006257687928155065,
  -0.004185070749372244,
  0.0006910567753948271,
  -0.034702226519584656,
  -0.04642307758331299,
  0.02167118526995182,
  0.02873556688427925,
  0.017439639195799828,
  -0.02662864699959755,
  -0.04086364060640335,
  -0.02414991706609726,
  -0.01209266297519207,
  0.0049574607983231544,
  -0.034967802464962006,
  0.016988156363368034,
  0.05060151219367981,
  0.014217289164662361,
  0.0029988212045282125,
  0.021228555589914322,
  -0.008772934786975384,
  0.0072193024680018425,
  0.029656238853931427,
  0.011658885516226292,
  -0.004390893969684839,
  -0.03530420362949371,
  -0.0017085534054785967,
  0.017563575878739357,
  -0.02197217382490635,
  0.02165348082780838,
  -0.03930558264255524,
  0.03838491067290306,
  0.04111151397228241,
  -0.005156644620001316,
  4.526587872533128e-05,
  0.025690270587801933,
  0.03316187113523483,
  -0.02894802950322628,
  -0.01762554422020912,
  0.006922740023583174,
  0.008564898744225502,
  -0.04330696165561676,
  0.0262745413929224,
  -0.0018988845404237509,
  -0.008392272517085075,
  -0.005864853039383888,
  -0.005515174940228462,
  0.018696708604693413,
  0.028009653091430664,
  0.0020626578480005264,
  0.006205678451806307,
  -0.026062078773975372,
  -0.004873360972851515,
  -0.016430441290140152,
  0.005276154726743698,
  -0.012880545109510422,
  0.04415681213140488,
  -0.022857435047626495,
  0.01919245533645153,
  0.007059955503791571,
  0.0002588004863355309,
  0.012703493237495422,
  -0.03640192747116089,
  0.016182567924261093,
  0.008892444893717766,
  0.002095855074003339,
  0.003341859672218561,
  -0.011729706078767776,
  -0.005457633174955845,
  -0.014208436943590641,
  -0.014580246061086655,
  -0.02928442880511284,
  -0.02622142620384693,
  0.01662519946694374,
  -0.016757987439632416,
  0.017041271552443504,
  -0.02147642895579338,
  -0.01644814759492874,
  0.020768219605088234,
  0.02884179912507534,
  -0.029939522966742516,
  -0.006041905377060175,
  -0.0026093064807355404,
  0.007768163923174143,
  0.03194021061062813,
  -0.006086168345063925,
  0.016341915354132652,
  0.02893032506108284,
  -0.008870313875377178,
  -0.01554518099874258,
  -0.01963508501648903,
  -0.02654012106359005,
  0.0038221138529479504,
  0.01446516253054142,
  0.008741950616240501,
  0.030789373442530632,
  -0.006249941419810057,
  -0.0012360455002635717,
  0.006453551817685366,
  -0.013845480047166348,
  0.010800182819366455,
  0.008817197754979134,
  -0.030116574838757515,
  0.025229934602975845,
  -0.017501607537269592,
  -0.006772245746105909,
  -0.02138790301978588,
  0.008179809898138046,
  0.013314323499798775,
  0.01550091803073883,
  -0.02404368482530117,
  -0.02446861006319523,
  0.001427483162842691,
  0.015022876672446728,
  0.0030939867720007896,
  0.0028616057243198156,
  -0.013872037641704082,
  -0.005262875929474831,
  0.002757587470114231,
  0.015952400863170624,
  -0.019989190623164177,
  -0.014819266274571419,
  0.006958150304853916,
  0.022662678733468056,
  0.0244331993162632,
  0.037393417209386826,
  -0.015120255760848522,
  0.02687652036547661,
  -0.008122268132865429,
  -0.00777701660990715,
  -0.004459501709789038,
  -0.02880638837814331,
  -0.028204411268234253,
  -0.04674177244305611,
  0.032843176275491714,
  0.016315357759594917,
  -0.006396009586751461,
  -0.009808690287172794,
  0.0020637642592191696,
  0.014943202957510948,
  0.0077371797524392605,
  0.005749769043177366,
  0.030594615265727043,
  -0.014872382394969463,
  0.008560472168028355,
  -0.003618503687903285,
  0.0292667243629694,
  0.029638532549142838,
  0.017501607537269592,
  -0.005028281826525927,
  0.009428028017282486,
  -0.006214531138539314,
  0.003600798547267914,
  0.02629224769771099,
  -0.039695095270872116,
  -0.012358241714537144,
  -0.012429062277078629,
  -0.0030630026012659073,
  0.0039194924756884575,
  -0.015704527497291565,
  -0.0291781984269619,
  0.014704182744026184,
  -0.025070587173104286,
  0.03927017003297806,
  0.006227809935808182,
  0.005794032476842403,
  -0.016687167808413506,
  0.01192446332424879,
  0.012721198610961437,
  -0.00038481183582916856,
  -0.020360998809337616,
  0.01929868757724762,
  -0.004767129663378,
  0.01209266297519207,
  0.00664388295263052,
  0.03367532417178154,
  -0.05035363882780075,
  -0.03824326768517494,
  -0.017306849360466003,
  -0.011880200356245041,
  0.01334088109433651,
  0.002465451369062066,
  0.037712112069129944,
  -0.0009754468337632716,
  0.006303057074546814,
  0.042138416320085526,
  0.008246204815804958,
  0.014819266274571419,
  -0.04854770377278328,
  0.030028048902750015,
  -0.014775003306567669,
  0.011198549531400204,
  -0.012508735992014408,
  -0.01580190658569336,
  -0.018625888973474503,
  -0.005302712321281433,
  -0.000531986472196877,
  -0.011543801985681057,
  -0.01191561110317707,
  0.004634340759366751,
  -0.014146468602120876,
  -0.005590422078967094,
  0.006356172729283571,
  0.017776038497686386,
  -0.044617146253585815,
  -0.007090939674526453,
  0.01428811065852642,
  0.03592388704419136,
  0.011216254904866219,
  -0.026132900267839432,
  0.005262875929474831,
  -0.007467175368219614,
  0.030700845643877983,
  -0.0009151384001597762,
  -0.00656420923769474,
  -0.011703148484230042,
  0.04652931168675423,
  0.025973552837967873,
  0.006280925590544939,
  0.02377810701727867,
  0.01792653277516365,
  0.0073875016532838345,
  -0.006196825765073299,
  -0.007529143709689379,
  0.01697930321097374,
  0.01087100338190794,
  0.013393997214734554,
  -0.004357696510851383,
  0.023335475474596024,
  -0.008113415911793709,
  -0.008843756280839443,
  -0.014252699911594391,
  -0.008016036823391914,
  0.00665273517370224,
  -0.0010390749666839838,
  0.00472286669537425,
  0.0077371797524392605,
  0.004656472243368626,
  -0.00023417915508616716,
  0.009560816921293736,
  -0.032205790281295776,
  0.02379581145942211,
  -0.03330351412296295,
  0.010277878493070602,
  0.01771407015621662,
  0.02682340331375599,
  0.013234649784862995,
  0.013411701656877995,
  0.008870313875377178,
  0.000889133894816041,
  0.004357696510851383,
  -0.04344860464334488,
  0.02147642895579338,
  0.0018833924550563097,
  0.02397286333143711,
  -0.0029789027757942677,
  0.032382842153310776,
  0.005696653388440609,
  -0.002405696315690875,
  0.00481139263138175,
  0.021140029653906822,
  0.02386663295328617,
  -0.004368762485682964,
  -0.021175440400838852,
  -0.018891466781497,
  0.0239905696362257,
  0.009171302430331707,
  0.015624853782355785,
  0.03156840056180954,
  -0.007267991546541452,
  0.016386179253458977,
  -0.0085250623524189,
  -0.0066128987818956375,
  0.03842031955718994,
  -0.023583348840475082,
  -0.000694376474712044,
  0.039695095270872116,
  0.006307483650743961,
  0.005063692107796669,
  0.03195791691541672,
  -0.017696363851428032,
  -0.019422622397542,
  -0.00543550169095397,
  0.02213152125477791,
  -0.025460101664066315,
  -0.012411356903612614,
  -0.03574683517217636,
  -0.00785669032484293,
  0.06175579875707626,
  -0.011968727223575115,
  -0.0015137960435822606,
  -0.0016089616110548377,
  0.0014407620765268803,
  0.0065686353482306,
  -0.013562195934355259,
  0.03360450267791748,
  0.03188709542155266,
  -0.0005737596657127142,
  -0.030470678582787514,
  0.000615809578448534,
  -0.02933754399418831,
  0.01064968854188919,
  0.0027000457048416138,
  -0.02152954414486885,
  0.02871786244213581,
  0.006351746618747711,
  -0.005068118218332529,
  0.020821334794163704,
  0.0050769709050655365,
  0.02179512195289135,
  -0.011490685865283012,
  -0.04072199761867523,
  -0.01334973331540823,
  0.008440962061285973,
  -0.0011851430172100663,
  -0.008591456338763237,
  -0.016005516052246094,
  0.0025473381392657757,
  -0.015102550387382507,
  0.002857179380953312,
  0.022999076172709465,
  0.008640145882964134,
  0.017430786043405533,
  -0.019865253940224648,
  -0.010605424642562866,
  0.04015543311834335,
  -0.01787341758608818,
  0.006086168345063925,
  0.017643248662352562,
  0.022467920556664467,
  0.0067810979671776295,
  0.009410322643816471,
  -0.002345941262319684,
  0.022768909111618996,
  0.01936950720846653,
  -0.017722921445965767,
  0.009959184564650059,
  0.011632327921688557,
  -0.009410322643816471,
  -0.02154724858701229,
  -0.01434122584760189,
  -0.025566333904862404,
  0.009073923341929913,
  -0.03174545615911484,
  0.005338123068213463,
  0.010419520549476147,
  0.03179857134819031,
  0.013544491492211819,
  0.030948719009757042,
  0.0243446733802557,
  -0.047060467302799225,
  -0.020060010254383087,
  -0.010676246136426926,
  0.002974476432427764,
  -0.020272472873330116,
  -0.00020236510317772627,
  -0.0026557825040072203,
  -0.016669461503624916,
  0.015527475625276566,
  0.03546354919672012,
  -0.002635864308103919,
  -0.016386179253458977,
  -0.025973552837967873,
  -0.011207402683794498,
  -0.01603207364678383,
  -0.01209266297519207,
  0.02414991706609726,
  -0.016421588137745857,
  0.018608182668685913,
  -0.059312477707862854,
  -0.002936852863058448,
  0.027690960094332695,
  0.03162151947617531,
  0.0035676013212651014,
  -0.022290868684649467,
  0.0006235556211322546,
  0.009985742159187794,
  0.005678948480635881,
  0.005780753213912249,
  -0.016722576692700386,
  0.0028461136389523745,
  0.01070280373096466,
  -0.0077150482684373856,
  -0.006851918995380402,
  0.012499882839620113,
  0.030045753344893456,
  -0.011526096612215042,
  -0.012110368348658085,
  0.0070156920701265335,
  -0.006626177579164505,
  0.013739248737692833,
  0.014987465925514698,
  -0.002328236121684313,
  0.00048495695227757096,
  0.013022187165915966,
  -0.005041560623794794,
  -0.008502930402755737,
  -0.0012780954129993916,
  0.008879166096448898,
  -0.015970105305314064,
  -0.02161807008087635,
  -0.0019796646665781736,
  -0.023158423602581024,
  -0.02455713599920273,
  -0.008724246174097061,
  0.020538052543997765,
  -0.0011685442877933383,
  0.008976545184850693,
  0.019776728004217148,
  -0.008733098395168781,
  -0.03792457655072212,
  -0.025070587173104286,
  -0.00300988694652915,
  0.005864853039383888,
  0.011773969046771526,
  -0.029850995168089867,
  -0.027938833460211754,
  -0.006431420333683491,
  -0.0017849071882665157,
  -0.030523793771862984,
  0.005705506075173616,
  0.01334088109433651,
  -0.011720853857696056,
  0.008175384253263474,
  0.002207619370892644,
  0.014695330522954464,
  -0.009861805476248264,
  0.044687967747449875,
  0.05616094917058945,
  0.009897216223180294,
  0.0019874107092618942,
  -0.02430926263332367,
  -0.01082674041390419,
  0.007484880741685629,
  0.0007004626677371562,
  -0.0073609440587460995,
  -0.01951115019619465,
  -5.273526403470896e-05,
  -0.011278223246335983,
  -0.013190386816859245,
  0.027814896777272224,
  -0.0009239910286851227,
  0.00943688116967678,
  0.0055417330004274845,
  0.025885026901960373,
  0.003886295249685645,
  -0.02413221076130867,
  0.005386812146753073,
  -0.017439639195799828,
  0.027018161490559578,
  -0.03611864149570465,
  0.02218463644385338,
  -0.013792363926768303,
  -0.02422073669731617,
  0.012933661229908466,
  -0.006506667472422123,
  -0.003720308654010296,
  0.022255457937717438,
  -0.01959967613220215,
  -0.013270060531795025,
  0.02880638837814331,
  0.02655782550573349,
  0.03337433561682701,
  -0.025973552837967873,
  0.026132900267839432,
  0.02432696893811226,
  0.04327154904603958,
  0.01312841847538948,
  -0.01805046945810318,
  0.011729706078767776,
  0.00656420923769474,
  -0.009605079889297485,
  -0.04075741022825241,
  -0.015288455411791801,
  0.04284662380814552,
  0.004393106792122126,
  -0.04334237053990364,
  -0.05803770199418068,
  -0.023671874776482582,
  -0.023370886221528053,
  -0.004165152553468943,
  -0.02684110961854458,
  0.01450057327747345,
  0.020431820303201675,
  0.008029315620660782,
  -0.016687167808413506,
  -0.027478497475385666,
  0.046954236924648285,
  0.034525174647569656,
  -0.042032185941934586,
  0.024716483429074287,
  -0.0004097097844351083,
  -0.012243157252669334,
  0.005028281826525927,
  -0.001136453589424491,
  -0.01794423721730709,
  -0.0011408799327909946,
  -0.02411450631916523,
  0.019900664687156677,
  -0.014137615449726582,
  -0.0011259411694481969,
  -0.011012645438313484,
  -0.0146333621814847,
  -0.02161807008087635,
  0.0010766985360533,
  -0.03337433561682701,
  0.025513218715786934,
  0.003572027664631605,
  0.017607837915420532,
  -0.012703493237495422,
  0.020325589925050735,
  0.027832601219415665,
  0.020431820303201675,
  0.022928256541490555,
  -0.0023171703796833754,
  0.0005798458587378263,
  -0.024663368239998817,
  0.02152954414486885,
  -0.04043871536850929,
  -0.00948114413768053,
  -0.0032754652202129364,
  0.011835937388241291,
  -0.012225452810525894,
  -0.017271438613533974,
  0.013721543364226818,
  0.014164173975586891,
  -0.0004478866758290678,
  -0.012756609357893467,
  0.015040582045912743,
  0.022892845794558525,
  0.0055417330004274845,
  0.0004135828057769686,
  -0.020573461428284645,
  ...],
 [0.007534371688961983,
  0.005438145250082016,
  -0.019238615408539772,
  0.009694576263427734,
  -0.005520940758287907,
  0.013721438124775887,
  -0.02991168014705181,
  0.011342954821884632,
  0.014835410751402378,
  0.031010599806904793,
  -0.020984843373298645,
  0.03110092133283615,
  -0.018771950155496597,
  0.009408555924892426,
  0.008836515247821808,
  -0.013608534820377827,
  -0.03817615658044815,
  0.011726823635399342,
  0.020984843373298645,
  -0.01193004846572876,
  0.06449000537395477,
  -0.015610676258802414,
  -0.004768256098031998,
  -0.005054275970906019,
  0.03901916369795799,
  0.01336014922708273,
  0.017793461680412292,
  0.03362994268536568,
  -0.03197403624653816,
  0.024627836421132088,
  -0.021617097780108452,
  0.02482353337109089,
  0.0012814452638849616,
  0.027066532522439957,
  0.0001524186081951484,
  0.03781486675143242,
  0.008693506009876728,
  -0.0028244485147297382,
  0.003283585887402296,
  -0.017703138291835785,
  -0.02116548642516136,
  -0.030303075909614563,
  -0.037905190140008926,
  0.019961191341280937,
  0.01180209219455719,
  0.04007292166352272,
  0.006597279570996761,
  -0.0071053411811590195,
  -0.008896730840206146,
  0.01910313218832016,
  -0.024552566930651665,
  0.002798104425892234,
  0.023002037778496742,
  0.028963297605514526,
  -0.01607733964920044,
  -0.008603183552622795,
  0.04762987419962883,
  -0.033720262348651886,
  -0.006811794359236956,
  -0.009754790924489498,
  0.03284715116024017,
  -0.034141767770051956,
  0.011117149144411087,
  -0.017597762867808342,
  -0.0033005212899297476,
  0.0014827883569523692,
  0.0029448780696839094,
  0.01120747160166502,
  0.037483684718608856,
  0.03718261420726776,
  0.0023860097862780094,
  -0.002058591926470399,
  0.01643862947821617,
  0.012885957956314087,
  -0.0007004669751040637,
  0.005867175292223692,
  -0.023483755066990852,
  0.029610607773065567,
  0.026614923030138016,
  -0.00781286507844925,
  0.015414977446198463,
  0.01001823041588068,
  0.00026155784144066274,
  0.008211787790060043,
  -0.0030954149551689625,
  -0.001138435211032629,
  0.006747816223651171,
  -0.060184650123119354,
  0.0003873972746077925,
  0.030378345400094986,
  -0.00048265891382470727,
  -0.016860131174325943,
  0.0025516003370285034,
  -0.014233263209462166,
  -0.014353692531585693,
  0.020759036764502525,
  0.012968753464519978,
  -0.00401933491230011,
  -0.04094603657722473,
  0.026524601504206657,
  0.0029467595741152763,
  0.010627904906868935,
  0.0020830542780458927,
  -0.027548251673579216,
  0.01129026710987091,
  0.005415564868599176,
  0.006680074613541365,
  0.0383266918361187,
  0.0027585886418819427,
  0.026810619980096817,
  -0.03630949929356575,
  -0.019313883036375046,
  -0.028029968962073326,
  -0.008783827535808086,
  -0.018546145409345627,
  0.04425784572958946,
  0.006555881816893816,
  0.027518143877387047,
  -0.01010102592408657,
  -0.02965576760470867,
  0.0070639438927173615,
  -0.026057936251163483,
  0.009619307704269886,
  -0.033750370144844055,
  0.001979560125619173,
  -0.017928944900631905,
  -0.010334357619285583,
  -0.002495148917660117,
  -0.012953699566423893,
  -0.0005593386595137417,
  -0.01868162862956524,
  -0.014639712870121002,
  -0.015640784054994583,
  0.03299768641591072,
  -0.011297794058918953,
  -0.011297794058918953,
  0.010755861178040504,
  0.015083796344697475,
  -0.0034115423914045095,
  0.005927389953285456,
  -0.028632117435336113,
  0.0320342518389225,
  0.0068757724948227406,
  -0.0070074922405183315,
  -0.028120292350649834,
  0.0013181386748328805,
  0.04230086877942085,
  -0.013646169565618038,
  0.0031010599341243505,
  0.028707385063171387,
  0.004282774403691292,
  0.01238165982067585,
  0.013864448294043541,
  -0.03444284200668335,
  0.028872976079583168,
  0.012336498126387596,
  -0.011666608974337578,
  0.02187301032245159,
  0.0036222939379513264,
  0.0035394986625760794,
  -0.036038532853126526,
  0.011779512278735638,
  -0.03835679963231087,
  -0.011403169482946396,
  0.0007338673458434641,
  -0.044077202677726746,
  -0.02167731337249279,
  -0.03263640031218529,
  -0.007538135163486004,
  0.06659752130508423,
  -0.02258053421974182,
  0.006548354867845774,
  0.0036524012684822083,
  0.023664399981498718,
  -0.016935400664806366,
  0.013495632447302341,
  0.013089182786643505,
  -0.007541898172348738,
  0.013924662955105305,
  -0.0014150467468425632,
  -0.004820943810045719,
  -0.007353727240115404,
  -0.011350481770932674,
  -0.033268652856349945,
  -0.00729351257905364,
  0.04762987419962883,
  -0.022460104897618294,
  0.01598701812326908,
  -0.008708558976650238,
  -0.02318268083035946,
  -0.00182337814476341,
  -0.022023547440767288,
  -0.008226840756833553,
  0.016905292868614197,
  0.06846418231725693,
  0.004918793216347694,
  -0.005727928597480059,
  -0.029640713706612587,
  0.00882898923009634,
  -0.0117192966863513,
  -0.008919310756027699,
  -0.0016323844902217388,
  -0.018290232867002487,
  -0.012005317024886608,
  0.003951593302190304,
  -0.02057839371263981,
  0.01053005550056696,
  0.008520388044416904,
  0.00791071355342865,
  -0.02212892286479473,
  -0.011275213211774826,
  -0.027232123538851738,
  -0.007026309613138437,
  -0.003992991056293249,
  0.02730739302933216,
  0.004470945801585913,
  -0.00976231787353754,
  -0.001887356280349195,
  0.02032248117029667,
  0.016860131174325943,
  -0.0003081301983911544,
  -0.022204192355275154,
  -0.003016382921487093,
  0.023920312523841858,
  -0.02858695574104786,
  -0.00033423895365558565,
  0.01144833117723465,
  0.02980630472302437,
  -0.025862237438559532,
  -0.01621282286942005,
  -0.02759341336786747,
  -0.013601007871329784,
  -0.020217103883624077,
  0.011373062618076801,
  -0.03965141624212265,
  0.008234367705881596,
  -0.003947829827666283,
  -0.01453433744609356,
  0.017477333545684814,
  0.014918206259608269,
  -0.0020849360153079033,
  0.00803866982460022,
  -0.024386975914239883,
  -0.022821392863988876,
  0.012020370922982693,
  0.008136519230902195,
  -0.051754582673311234,
  -0.012622518464922905,
  -0.004312882199883461,
  0.019449366256594658,
  -0.0026964920107275248,
  0.009318233467638493,
  0.010138659738004208,
  0.016965508460998535,
  -0.0024631598498672247,
  0.029866520315408707,
  0.03959120437502861,
  -0.01412036083638668,
  -0.02881276048719883,
  -0.0014733797870576382,
  0.035677243024110794,
  0.007346200291067362,
  -0.023363325744867325,
  0.0032421883661299944,
  0.04416752606630325,
  -0.0010481131030246615,
  0.019539689645171165,
  0.0016700186533853412,
  -0.008558022789657116,
  0.014255844056606293,
  0.04296322911977768,
  -0.025681594386696815,
  0.046425577253103256,
  0.00786555279046297,
  0.013841867446899414,
  -0.0008566490141674876,
  -0.012682733125984669,
  -0.031401995569467545,
  0.012878431007266045,
  -0.002474450273439288,
  0.02270096354186535,
  0.03022780828177929,
  -0.044137418270111084,
  0.00802361685782671,
  -0.0031048234086483717,
  0.007696198765188456,
  -0.033660050481557846,
  -0.008098884485661983,
  -0.017552601173520088,
  -0.005423091351985931,
  -0.09923391789197922,
  0.0018516037380322814,
  0.011568760499358177,
  -0.0044972896575927734,
  0.030754687264561653,
  -0.016634326428174973,
  0.04398687928915024,
  0.024100957438349724,
  -0.028707385063171387,
  0.018967648968100548,
  -0.020623553544282913,
  -0.01204295177012682,
  -0.019343990832567215,
  -0.020819252356886864,
  -0.017522495239973068,
  0.022550426423549652,
  0.02810523845255375,
  -0.00905479397624731,
  -0.013081655837595463,
  -0.009739737026393414,
  0.03926002234220505,
  -0.01058274321258068,
  -0.004342989530414343,
  -0.01522680651396513,
  -0.04094603657722473,
  0.015023581683635712,
  0.013299934566020966,
  0.005189759191125631,
  0.03775465115904808,
  0.01884721964597702,
  0.00816662609577179,
  0.010048337280750275,
  -0.01836550049483776,
  0.048894383013248444,
  -0.009664468467235565,
  -0.0001768808433553204,
  -0.008889203891158104,
  -0.00980747863650322,
  0.015083796344697475,
  0.0015430031344294548,
  -0.02331816405057907,
  -0.03865787386894226,
  0.018666574731469154,
  -0.0074628666043281555,
  0.037574008107185364,
  -0.015475193038582802,
  -0.04871373996138573,
  0.0026080517563968897,
  0.020894519984722137,
  -0.008106411434710026,
  -0.02000635303556919,
  -0.03405144438147545,
  -0.02434181608259678,
  0.00718813668936491,
  -0.0016276801470667124,
  -0.03293747082352638,
  0.0327267199754715,
  0.005671477410942316,
  0.033720262348651886,
  0.011154783889651299,
  0.0005758036277256906,
  -0.030890170484781265,
  0.03393101692199707,
  0.021180540323257446,
  0.023077305406332016,
  0.010650484822690487,
  -0.03161274641752243,
  -0.028993405401706696,
  -0.004583848174661398,
  0.020653661340475082,
  -0.014135414734482765,
  -0.04091592878103256,
  -0.0202923733741045,
  -0.010191347450017929,
  0.002909125527366996,
  0.010078445076942444,
  0.024236440658569336,
  -0.011809619143605232,
  0.021105272695422173,
  -0.011659082025289536,
  -0.03591810166835785,
  0.009611780755221844,
  -0.04868363216519356,
  0.015535407699644566,
  -0.03820626437664032,
  -0.001242870232090354,
  -0.0027454164810478687,
  0.010191347450017929,
  -0.0454019270837307,
  0.003876324975863099,
  0.007730069570243359,
  0.004361806437373161,
  0.01120747160166502,
  -0.010319303721189499,
  0.024236440658569336,
  -0.023137520998716354,
  0.00308788800612092,
  0.01669454202055931,
  -0.012554776854813099,
  -0.004753202199935913,
  -0.016709595918655396,
  0.014105306938290596,
  -0.021797742694616318,
  -0.03919980674982071,
  0.02103000320494175,
  -0.0031876186840236187,
  0.03399122878909111,
  -0.0014225736958906054,
  0.008821462281048298,
  -0.009190277196466923,
  -0.006759106647223234,
  -0.03769443929195404,
  0.0005828600260429084,
  0.011147256940603256,
  -0.00802361685782671,
  -0.026072990149259567,
  -0.04064496234059334,
  -0.024402029812335968,
  0.002839502179995179,
  0.004534923937171698,
  0.02224935218691826,
  -0.006890826392918825,
  0.03868798166513443,
  -0.026569761335849762,
  0.017838621512055397,
  -0.020337535068392754,
  -0.0073048025369644165,
  0.011425750330090523,
  0.024778373539447784,
  -0.020789144560694695,
  0.002169613027945161,
  0.014090253040194511,
  0.006484376732259989,
  0.024672996252775192,
  -0.01881711184978485,
  0.054976072162389755,
  -0.006367710418999195,
  0.008941891603171825,
  0.003283585887402296,
  -0.0025120843201875687,
  0.006153195630759001,
  -0.03432241082191467,
  -0.004708041436970234,
  0.03387080132961273,
  0.0015646428801119328,
  -0.007873079739511013,
  0.016378413885831833,
  0.003268532222136855,
  -0.01141822338104248,
  -0.008964472450315952,
  0.01646873541176319,
  -0.0012014725944027305,
  0.022294513881206512,
  0.015460139140486717,
  -0.0069999657571315765,
  0.006604806054383516,
  0.016062285751104355,
  -0.011041881516575813,
  -0.012547249905765057,
  -0.019900977611541748,
  -0.030724579468369484,
  0.01604723371565342,
  -0.06009432673454285,
  -0.020081620663404465,
  0.00420374283567071,
  0.0024198805913329124,
  0.021586989983916283,
  0.046034183353185654,
  -0.01698056235909462,
  -0.006879535969346762,
  0.03131167218089104,
  0.03579767420887947,
  0.023694507777690887,
  -0.0040343888103961945,
  0.0007715015672147274,
  -0.0413675382733345,
  0.0032704139593988657,
  0.03236543387174606,
  -0.01294617261737585,
  0.008083831518888474,
  -0.03206435963511467,
  -0.0017960933037102222,
  -0.0551266111433506,
  0.02611815184354782,
  -0.014428961090743542,
  -0.020608501508831978,
  -0.06954804807901382,
  -0.007632220629602671,
  -0.04669654369354248,
  -0.03528584912419319,
  0.01497089397162199,
  0.002879018196836114,
  -0.0061117978766560555,
  -0.0165590588003397,
  0.017507441341876984,
  0.007222007494419813,
  0.0008373614982701838,
  0.02730739302933216,
  -0.0004958309000357985,
  -0.0076623279601335526,
  0.01022898219525814,
  -0.018967648968100548,
  0.005656423978507519,
  -0.009927907958626747,
  0.004640299826860428,
  -0.01205800473690033,
  -0.022746125236153603,
  -0.033780477941036224,
  0.03302779421210289,
  0.019012808799743652,
  0.030468666926026344,
  -0.01598701812326908,
  0.015317128971219063,
  0.023077305406332016,
  0.002991920802742243,
  -0.004572558216750622,
  -0.004884922411292791,
  0.010319303721189499,
  -0.012908538803458214,
  0.025485895574092865,
  0.048412665724754333,
  -0.00020981079433113337,
  0.003005092730745673,
  -0.01975044049322605,
  -0.01633325219154358,
  0.013563374057412148,
  0.00816662609577179,
  0.012833270244300365,
  0.012374132871627808,
  -0.04196968674659729,
  0.027472982183098793,
  0.004297828301787376,
  -0.0025403101462870836,
  0.007124158553779125,
  -0.0013661222765222192,
  0.015490246005356312,
  -0.013691330328583717,
  0.029761143028736115,
  0.005295135080814362,
  -0.01607733964920044,
  -0.02405579574406147,
  0.030950386077165604,
  0.033298760652542114,
  0.025802023708820343,
  0.010123605839908123,
  -0.001065048505552113,
  -0.0225052647292614,
  0.003962883725762367,
  -0.012968753464519978,
  0.032275110483169556,
  -0.021135380491614342,
  0.029309533536434174,
  0.012306391261518002,
  -0.030077271163463593,
  0.0016775454860180616,
  0.01014618668705225,
  0.009732210077345371,
  -0.03627939149737358,
  -0.0010669302428141236,
  -0.010025757364928722,
  0.013044022023677826,
  -0.003405897179618478,
  0.0003151866258122027,
  0.01556551456451416,
  0.04160839691758156,
  0.00934081431478262,
  -0.01123757939785719,
  -0.0008646462811157107,
  -0.021857956424355507,
  -0.026705244556069374,
  0.02955039218068123,
  0.029354693368077278,
  0.03495466709136963,
  0.04495031759142876,
  0.00892683770507574,
  0.012185961939394474,
  -0.021602043882012367,
  -0.002209128811955452,
  -0.005569864995777607,
  0.014127887785434723,
  0.006811794359236956,
  0.012614991515874863,
  0.01649884320795536,
  -0.02955039218068123,
  0.0016060405177995563,
  0.057836275547742844,
  0.03486434370279312,
  -0.03597831726074219,
  0.0028338569682091475,
  -0.006864482536911964,
  -0.01565583609044552,
  0.011019300669431686,
  0.03721271827816963,
  -0.01341283693909645,
  -0.005983841605484486,
  -0.0193590447306633,
  0.023242896422743797,
  -0.011681662872433662,
  -0.014444014988839626,
  -0.016227876767516136,
  0.02447729930281639,
  0.03251596912741661,
  0.019735386595129967,
  0.0039704106748104095,
  0.013984877616167068,
  0.026283740997314453,
  -0.041217003017663956,
  0.016197768971323967,
  -0.011809619143605232,
  -0.001350127742625773,
  0.01320208515971899,
  -0.008761247619986534,
  0.01604723371565342,
  -0.001405638293363154,
  -0.0006858837441541255,
  0.009197804145514965,
  0.017552601173520088,
  0.007673617918044329,
  -0.015700997784733772,
  -0.016137555241584778,
  0.0153547627851367,
  0.0018356092041358352,
  -0.036580465734004974,
  0.010936505161225796,
  -0.023498808965086937,
  -0.0268256738781929,
  0.016875185072422028,
  0.055487897247076035,
  -0.00040480311145074666,
  0.006059110164642334,
  -0.011508545838296413,
  -0.025877291336655617,
  0.021526776254177094,
  -0.005140834953635931,
  0.011327900923788548,
  0.019946137443184853,
  0.01193004846572876,
  -0.006668784189969301,
  -0.005441908724606037,
  0.019057970494031906,
  -0.025576218962669373,
  -0.013548320159316063,
  0.0009258018690161407,
  -0.0014169284841045737,
  -0.025034286081790924,
  -0.0051596518605947495,
  0.0036862720735371113,
  0.019313883036375046,
  0.009980595670640469,
  0.009152643382549286,
  0.031401995569467545,
  0.009175223298370838,
  0.061087872833013535,
  0.03317832946777344,
  -0.02190311811864376,
  -0.0027585886418819427,
  -0.010236509144306183,
  -0.01604723371565342,
  -0.00669889198616147,
  0.04636536166071892,
  -0.040855713188648224,
  0.018952595070004463,
  0.011199944652616978,
  -0.005404274445027113,
  0.05723412707448006,
  -0.008475227281451225,
  0.01678486354649067,
  -0.010131132788956165,
  -0.037965405732393265,
  0.003268532222136855,
  -0.0022938058245927095,
  0.004523633513599634,
  -0.018380554392933846,
  -0.030378345400094986,
  -0.01678486354649067,
  0.004275247920304537,
  -0.011087042279541492,
  -0.007752649951726198,
  -0.004211269784718752,
  0.010793494991958141,
  -0.00908490177243948,
  -0.03600842505693436,
  0.03045361302793026,
  -0.01621282286942005,
  0.026569761335849762,
  -0.027668680995702744,
  -0.03886862471699715,
  -0.003336273832246661,
  -0.03718261420726776,
  0.021722473204135895,
  0.003146221162751317,
  -0.027834272012114525,
  -0.005720402114093304,
  -0.011832199990749359,
  0.005983841605484486,
  -0.030769741162657738,
  -0.04040410369634628,
  -0.006672547664493322,
  -0.016152609139680862,
  0.007880606688559055,
  0.010138659738004208,
  0.00602147588506341,
  -0.0008557081455364823,
  -0.004617718979716301,
  -0.0003565842634998262,
  0.003131167497485876,
  0.029128888621926308,
  -0.023017091676592827,
  -0.02051817812025547,
  -0.004519870039075613,
  0.016995614394545555,
  0.057896487414836884,
  -0.017793461680412292,
  0.00908490177243948,
  -0.008821462281048298,
  -0.0055247037671506405,
  -0.0026833200827240944,
  0.003744605230167508,
  -0.028210613876581192,
  0.013292407616972923,
  0.0284966342151165,
  -0.004169872030615807,
  -0.02949017658829689,
  0.018531091511249542,
  0.026554707437753677,
  0.026765460148453712,
  -0.030498774722218513,
  -0.007191900163888931,
  -0.013563374057412148,
  0.00215455936267972,
  -0.009702103212475777,
  0.007094051223248243,
  -0.002948641311377287,
  0.02659986913204193,
  0.03697185963392258,
  0.01785367541015148,
  0.018290232867002487,
  0.017251528799533844,
  -0.023498808965086937,
  0.011041881516575813,
  -0.0037144978996366262,
  -0.010891344398260117,
  -0.011071988381445408,
  -0.0019682699348777533,
  -0.004561267793178558,
  -0.012471981346607208,
  0.010033284313976765,
  0.017868729308247566,
  0.0004589023010339588,
  -0.002427407307550311,
  0.010552636347711086,
  0.01823001727461815,
  0.028029968962073326,
  -0.004422021098434925,
  -0.01159134041517973,
  0.01949452795088291,
  0.00905479397624731,
  -0.023273004218935966,
  0.010605324059724808,
  0.008964472450315952,
  -0.018335392698645592,
  -0.026765460148453712,
  0.02968587540090084,
  -0.024176225066184998,
  -0.009942961856722832,
  0.011041881516575813,
  -0.023559024557471275,
  0.015347235836088657,
  -0.03877830505371094,
  -0.002783050760626793,
  -0.007933294400572777,
  0.014654766768217087,
  0.008670925162732601,
  0.005855885334312916,
  -0.025350412353873253,
  0.0009032213711179793,
  0.0005188818322494626,
  -0.03862776607275009,
  0.043625593185424805,
  0.0046967510133981705,
  -0.008580602705478668,
  -0.019148292019963264,
  -0.03314822539687157,
  0.009370921179652214,
  0.010853709653019905,
  -0.024898802861571312,
  0.0019626247230917215,
  -0.022264406085014343,
  -0.0027661153580993414,
  -0.013924662955105305,
  -0.007398888468742371,
  -0.03161274641752243,
  -0.04022345691919327,
  0.00045796146150678396,
  -0.023754721507430077,
  -0.02312246710062027,
  0.012374132871627808,
  -0.006074163597077131,
  0.008512861095368862,
  0.015911750495433807,
  -0.008211787790060043,
  -0.00674405274912715,
  0.02563643269240856,
  0.02643427811563015,
  0.020111728459596634,
  -0.012155854143202305,
  -0.002101871417835355,
  -0.016845079138875008,
  -0.0072370609268546104,
  -0.04826212674379349,
  -0.02106011100113392,
  -0.03676110878586769,
  -0.02575686201453209,
  0.007639747112989426,
  -0.019885923713445663,
  0.030829954892396927,
  0.007730069570243359,
  0.0017575182719156146,
  -0.05578897148370743,
  0.01618271693587303,
  0.01205800473690033,
  -0.0022542900405824184,
  0.013616061769425869,
  -0.006401581224054098,
  -0.01323219295591116,
  0.003117995336651802,
  -0.0023991817142814398,
  0.017251528799533844,
  0.0024048269260674715,
  -0.01061285100877285,
  -0.0004864223301410675,
  0.02611815184354782,
  0.041728828102350235,
  -0.011952629312872887,
  -0.008791354484856129,
  0.03483423590660095,
  0.001208999427035451,
  -0.02498912438750267,
  0.011470911093056202,
  -0.027217069640755653,
  -0.005106964148581028,
  0.0017688084626570344,
  0.024688050150871277,
  0.03384069353342056,
  -0.03525574132800102,
  0.012667679227888584,
  -0.011132203042507172,
  -0.01836550049483776,
  0.01556551456451416,
  0.04955674707889557,
  -0.003225252963602543,
  0.010070918127894402,
  -0.012088112533092499,
  -0.004403204191476107,
  -0.03384069353342056,
  0.005637606605887413,
  0.00446341885253787,
  0.009137589484453201,
  0.01627303846180439,
  -0.033298760652542114,
  0.015866588801145554,
  -0.009641887620091438,
  -0.03694175183773041,
  0.002092462731525302,
  -0.0279095396399498,
  -0.0010612851474434137,
  -0.022625695914030075,
  0.021045057103037834,
  0.015294548124074936,
  0.012833270244300365,
  0.003575251204892993,
  -0.0012842678697779775,
  -0.01762787066400051,
  -0.027262231335043907,
  0.01737195812165737,
  -0.001853485475294292,
  0.019855815917253494,
  0.007440285757184029,
  -0.034232091158628464,
  0.020111728459596634,
  0.01107951533049345,
  -0.03236543387174606,
  0.02337837964296341,
  -0.02988157421350479,
  -0.013593481853604317,
  0.0073311468586325645,
  0.015460139140486717,
  0.012201014906167984,
  0.022264406085014343,
  0.024386975914239883,
  -0.019283775240182877,
  0.02515471540391445,
  -0.01352574024349451,
  -0.01658916473388672,
  0.01669454202055931,
  -0.013939715921878815,
  0.017191313207149506,
  -0.023197734728455544,
  -0.020247211679816246,
  0.010710700415074825,
  -0.0069999657571315765,
  0.008685979060828686,
  -0.012359078973531723,
  -0.013352622278034687,
  0.017311742529273033,
  0.0029467595741152763,
  -0.008896730840206146,
  0.006570935249328613,
  -0.015369816683232784,
  -0.022806338965892792,
  0.009694576263427734,
  -0.025064392015337944,
  -0.010710700415074825,
  -0.004851051606237888,
  -0.019690224900841713,
  -0.014985947869718075,
  0.013578427955508232,
  0.00638276431709528,
  0.009815005585551262,
  0.007319856435060501,
  0.014639712870121002,
  -0.006341366562992334,
  -0.0024085904005914927,
  0.004083313047885895,
  0.019238615408539772,
  -0.0033400373067706823,
  -0.026389118283987045,
  0.0052838451229035854,
  0.03549659997224808,
  0.010793494991958141,
  -0.029640713706612587,
  -0.0019127593841403723,
  -0.010304250754415989,
  0.03248586133122444,
  0.024266546592116356,
  -0.0032948763109743595,
  -0.0320342518389225,
  -0.0012456928379833698,
  -0.0027360080275684595,
  0.029113834723830223,
  -0.020563339814543724,
  -0.002756706904619932,
  0.007820392027497292,
  0.0049940613098442554,
  -0.01746227964758873,
  -0.001295558176934719,
  0.03998259827494621,
  -0.007508027367293835,
  0.0011901822872459888,
  -0.008527914993464947,
  -0.006160722579807043,
  -0.02357407659292221,
  0.029866520315408707,
  -0.01318703219294548,
  0.02733749896287918,
  -0.022264406085014343,
  -0.010838656686246395,
  0.010733280330896378,
  -0.005201049614697695,
  0.005223629996180534,
  -0.012163381092250347,
  0.012118220329284668,
  0.0503094308078289,
  -0.0021827849559485912,
  -0.007918240502476692,
  -0.004060732666403055,
  -0.03387080132961273,
  -0.006815557833760977,
  -0.018952595070004463,
  0.00887414999306202,
  0.00921285804361105,
  -0.009039740078151226,
  0.007647274062037468,
  -0.0020115491934120655,
  -0.016890238970518112,
  0.0117192966863513,
  0.024462245404720306,
  -0.015971964225172997,
  0.004038152284920216,
  0.0026287504006177187,
  0.021300969645380974,
  -0.02187301032245159,
  0.03739336505532265,
  0.0051859961822628975,
  0.011613921262323856,
  0.03320843726396561,
  -0.0057166386395692825,
  -0.008241894654929638,
  -0.025877291336655617,
  0.015144011937081814,
  0.038025617599487305,
  0.010213928297162056,
  0.016830025240778923,
  -0.008053723722696304,
  0.014963367022573948,
  -0.023363325744867325,
  -0.0004229145706631243,
  -0.019961191341280937,
  0.010793494991958141,
  0.01762787066400051,
  -0.004109657369554043,
  0.009920381009578705,
  -0.02742782235145569,
  -0.016875185072422028,
  0.010454786941409111,
  -0.002252408303320408,
  0.002997565781697631,
  -0.008768773637712002,
  0.011621448211371899,
  -0.02578696981072426,
  ...],
 [-0.005651965271681547,
  -0.022757913917303085,
  -0.01466343179345131,
  -0.011637380346655846,
  0.01561376266181469,
  -0.01715596579015255,
  0.00256964354775846,
  0.03454534709453583,
  -0.007285866886377335,
  0.035645727068185806,
  0.01723932847380638,
  0.008207020349800587,
  -0.030760696157813072,
  -0.033928465098142624,
  -0.0061813159845769405,
  0.034378621727228165,
  -0.01295450422912836,
  -0.00957833044230938,
  -0.0142966378480196,
  -0.014963536523282528,
  0.01569712534546852,
  -0.02572561241686344,
  -0.02187427319586277,
  0.022791258990764618,
  -0.0017714493442326784,
  -0.0015119840390980244,
  -0.005781176965683699,
  -0.02182425558567047,
  -0.041914574801921844,
  -0.02540883608162403,
  -0.005322684068232775,
  0.036512695252895355,
  -0.019856905564665794,
  -0.0017495666397735476,
  -0.006214661058038473,
  0.018072951585054398,
  0.032661356031894684,
  -0.015438701957464218,
  0.004964226391166449,
  -0.017506087198853493,
  0.01715596579015255,
  -0.0027196956798434258,
  -0.016672464087605476,
  0.06548944115638733,
  -0.00679819704964757,
  -0.0236915722489357,
  0.039547085762023926,
  -0.026075733825564384,
  0.013446342200040817,
  -0.02160751260817051,
  -0.014580070041120052,
  0.005635292734950781,
  0.028193136677145958,
  0.05168464034795761,
  0.000825286959297955,
  0.009394933469593525,
  0.04888366535305977,
  0.019273368641734123,
  0.00040691232425160706,
  -0.00171309569850564,
  0.033645033836364746,
  -0.06068776920437813,
  -0.013588057830929756,
  -0.011353948153555393,
  -0.0024383477866649628,
  -0.012195907533168793,
  0.0031823564786463976,
  0.05278502032160759,
  -0.008031959645450115,
  0.05078432708978653,
  -0.017939571291208267,
  -0.00043061847100034356,
  0.00779437692835927,
  0.006147970911115408,
  0.00583119411021471,
  -0.026892684400081635,
  0.0010743319289758801,
  0.049850668758153915,
  0.02060716599225998,
  0.026759304106235504,
  0.007527617737650871,
  -0.013663084246218204,
  -0.0005595695693045855,
  0.022291084751486778,
  -0.003344913013279438,
  0.014721785672008991,
  -0.02250782586634159,
  -0.0766933336853981,
  0.012829461134970188,
  0.0522848479449749,
  0.002150747925043106,
  -0.008869750425219536,
  -0.039913877844810486,
  0.0003581974597182125,
  0.04464885964989662,
  0.02079056203365326,
  0.005860371049493551,
  0.02319139800965786,
  -0.03491213917732239,
  -0.004207713063806295,
  -0.019089970737695694,
  -0.026575908064842224,
  -0.04114764183759689,
  -0.02324141561985016,
  0.009453287348151207,
  0.02582564763724804,
  0.0022195216733962297,
  0.026759304106235504,
  0.02704273723065853,
  -0.005326852202415466,
  -0.04184788465499878,
  -0.012304278090596199,
  0.008611327968537807,
  0.025125402957201004,
  -0.015096915885806084,
  0.012087536044418812,
  -0.010220220312476158,
  0.016138944774866104,
  -0.008836406283080578,
  -0.03484544903039932,
  -0.017772845923900604,
  0.009278226643800735,
  0.005485240835696459,
  0.01536367554217577,
  0.020757216960191727,
  0.01454672496765852,
  -0.006985762622207403,
  0.015622098930180073,
  0.0014661347959190607,
  -0.02464190125465393,
  -0.021474134176969528,
  0.02232442982494831,
  0.023308103904128075,
  0.026575908064842224,
  0.020407095551490784,
  0.033561669290065765,
  0.021040650084614754,
  -0.01379646360874176,
  0.006906568072736263,
  0.001839181175455451,
  -0.052118122577667236,
  -0.010153530165553093,
  -0.009661693125963211,
  0.006243837997317314,
  -0.04384858161211014,
  -0.0060771130956709385,
  0.024308452382683754,
  0.026709286496043205,
  0.015080243349075317,
  0.04068081080913544,
  0.0006137550808489323,
  -0.050084080547094345,
  0.010195211507380009,
  1.7869168686956982e-06,
  0.0015307406429201365,
  0.025858990848064423,
  -0.007765199989080429,
  -0.006302191410213709,
  -0.006785692647099495,
  -0.009620011784136295,
  0.011487327516078949,
  0.02260786108672619,
  -0.015997229143977165,
  -0.009161518886685371,
  -0.008944776840507984,
  -0.03781314939260483,
  -0.006777356844395399,
  -0.01037027221173048,
  -0.016280660405755043,
  0.04825011268258095,
  0.025942353531718254,
  0.014304974116384983,
  0.022157704457640648,
  -0.005097605753690004,
  0.0219576358795166,
  0.027059409767389297,
  0.054018784314394,
  -0.002371658105403185,
  0.02722613327205181,
  -0.02740953117609024,
  -0.01370476558804512,
  0.021357426419854164,
  0.003730463795363903,
  -0.026475872844457626,
  -0.04221468046307564,
  0.012829461134970188,
  0.02219104953110218,
  -0.012537692673504353,
  -0.01578882336616516,
  -0.0473831444978714,
  -0.01715596579015255,
  -0.013471351005136967,
  0.0424814373254776,
  0.010245229117572308,
  0.016547420993447304,
  -0.032744720578193665,
  -0.00025894420105032623,
  -0.006994098890572786,
  -0.0376797690987587,
  -0.036779455840587616,
  -0.005347692873328924,
  0.011920811608433723,
  0.010828765109181404,
  -0.031344231218099594,
  -0.02210768684744835,
  -0.022657878696918488,
  -0.015405356884002686,
  -0.0021465797908604145,
  0.02424176223576069,
  -0.015530399978160858,
  -0.013021194376051426,
  0.010270237922668457,
  0.02255784347653389,
  -0.0018902405863627791,
  0.06265512108802795,
  0.002536298707127571,
  -0.010887118987739086,
  0.01188746653497219,
  -0.0032886434346437454,
  -0.030927421525120735,
  0.01788955368101597,
  -0.0018120884196832776,
  0.006235501728951931,
  0.049217112362384796,
  -0.019723525270819664,
  0.001973602920770645,
  0.007369229104369879,
  0.02074054442346096,
  0.008402922190725803,
  0.027059409767389297,
  -0.004301495850086212,
  -0.006398058030754328,
  0.0027134434785693884,
  -0.0056686378084123135,
  -0.04941718280315399,
  -0.005889547988772392,
  -0.014363327994942665,
  0.014980209060013294,
  0.0344119668006897,
  -0.004568255040794611,
  -0.025575559586286545,
  -0.004505733493715525,
  -0.05585275590419769,
  -0.009344915859401226,
  0.006939913146197796,
  0.014530052430927753,
  -0.0393136702477932,
  0.022707896307110786,
  0.008219525218009949,
  0.034245241433382034,
  0.006927408743649721,
  -0.0135046960785985,
  -0.03117750771343708,
  -0.013179582543671131,
  -0.044915616512298584,
  0.018122969195246696,
  -0.007365060970187187,
  0.05331854149699211,
  -0.01788955368101597,
  0.018623141571879387,
  0.015538736246526241,
  -0.011762423440814018,
  0.0114623187109828,
  -0.0077151828445494175,
  -0.0013452594866976142,
  -0.011870793998241425,
  0.013996533118188381,
  -0.03821328654885292,
  -0.022891292348504066,
  -0.02672595903277397,
  0.015680452808737755,
  -0.0035241420846432447,
  0.030927421525120735,
  0.04621607065200806,
  0.041547778993844986,
  -0.04968394339084625,
  -0.027676289901137352,
  0.02645920030772686,
  0.024591883644461632,
  0.00749844079837203,
  0.006327200215309858,
  0.005531089846044779,
  0.018756521865725517,
  0.013387988321483135,
  -0.009961796924471855,
  0.04184788465499878,
  -0.04038070887327194,
  -0.028943397104740143,
  -0.03644600510597229,
  -0.011654051952064037,
  -0.03821328654885292,
  -0.01191247534006834,
  -0.008719698525965214,
  0.01250434760004282,
  -0.00662730447947979,
  0.04121433198451996,
  0.044915616512298584,
  0.005818690173327923,
  0.025108730420470238,
  -0.0003430880606174469,
  0.03454534709453583,
  -0.02759292721748352,
  0.0015526232309639454,
  -0.021040650084614754,
  -0.029193485155701637,
  0.016238979995250702,
  0.035012174397706985,
  0.010928800329566002,
  -0.013838144950568676,
  -0.018172984942793846,
  0.030477264896035194,
  -0.036379314959049225,
  -0.030510609969496727,
  -0.059287283569574356,
  0.004405698738992214,
  -0.0065439422614872456,
  -0.009169855155050755,
  0.0022862115874886513,
  0.014338319189846516,
  -0.0211907010525465,
  -0.009036475792527199,
  0.006214661058038473,
  -0.03434527665376663,
  0.015321994200348854,
  -0.030010435730218887,
  0.035012174397706985,
  -0.03794652968645096,
  0.01238764077425003,
  0.006443907506763935,
  -0.014871837571263313,
  -0.016455721110105515,
  0.011770759709179401,
  -0.001244182582013309,
  -0.023358121514320374,
  0.018172984942793846,
  0.01755610480904579,
  -0.017572777345776558,
  0.029610296711325645,
  -0.021974308416247368,
  -0.015422029420733452,
  0.05795348435640335,
  -0.01905662566423416,
  -0.014580070041120052,
  -0.04658286273479462,
  -0.014713449403643608,
  0.0015484550967812538,
  -0.008423762395977974,
  -0.024258434772491455,
  0.025975698605179787,
  0.008436267264187336,
  0.00025464582722634077,
  -0.010912127792835236,
  0.028943397104740143,
  -0.02350817434489727,
  0.040013913065195084,
  -0.021574167534708977,
  -0.0344119668006897,
  -0.03149428591132164,
  -0.020040301606059074,
  0.018539780750870705,
  0.0024800291284918785,
  0.007881907746195793,
  0.011862457729876041,
  0.007936093024909496,
  -0.024525195360183716,
  -0.00794442929327488,
  0.026359165087342262,
  -0.00044103877735324204,
  -0.02682599425315857,
  0.046949658542871475,
  -0.005852034781128168,
  0.010045159608125687,
  -0.027126099914312363,
  0.027492893859744072,
  -0.01288781501352787,
  -0.03567907214164734,
  0.05912055820226669,
  0.00134734355378896,
  -0.014413345605134964,
  0.01037027221173048,
  0.01193748414516449,
  -0.022591188549995422,
  0.006589791271835566,
  -0.009795072488486767,
  -0.005397710017859936,
  0.01823967508971691,
  -0.006577286869287491,
  -0.03831332176923752,
  0.01565544307231903,
  0.010837101377546787,
  0.02650921791791916,
  -0.0253754910081625,
  -0.021090667694807053,
  0.026442527770996094,
  0.024658573791384697,
  0.011037170886993408,
  0.022757913917303085,
  -0.016130609437823296,
  0.02132408134639263,
  0.010095177218317986,
  0.016939222812652588,
  -0.050951048731803894,
  -0.02242446504533291,
  -0.014246620237827301,
  -0.010703722015023232,
  0.002090310212224722,
  -0.03534562513232231,
  -0.028976742178201675,
  -0.011195559985935688,
  0.00864467304199934,
  -0.06385553628206253,
  -0.0005840572412125766,
  0.032061148434877396,
  0.00932824332267046,
  -0.0006814869702793658,
  0.02224106714129448,
  0.03352832421660423,
  0.016639119014143944,
  -0.0442153736948967,
  -0.06325532495975494,
  0.023041345179080963,
  -0.01454672496765852,
  -0.036912836134433746,
  0.0063355364836752415,
  -0.031110817566514015,
  0.05948735028505325,
  -0.003324072575196624,
  0.033961810171604156,
  -0.029660314321517944,
  -0.03771311417222023,
  -0.04528241232037544,
  3.459861943611031e-07,
  0.0409475713968277,
  0.05241822823882103,
  -0.012345959432423115,
  0.005789513234049082,
  0.054018784314394,
  0.010912127792835236,
  -0.01474679447710514,
  -0.00047646777238696814,
  0.01892324723303318,
  0.012646064162254333,
  -0.0028176463674753904,
  0.010286910459399223,
  -0.030477264896035194,
  -0.014680104330182076,
  0.035612381994724274,
  -0.01417993102222681,
  -0.002257034881040454,
  0.005097605753690004,
  -0.007806881330907345,
  -0.015105252154171467,
  0.002561307279393077,
  -0.006468915846198797,
  -0.0035762435290962458,
  -0.00833206344395876,
  0.0012202159268781543,
  0.03974715247750282,
  -0.026025716215372086,
  0.004935049451887608,
  -0.003836750751361251,
  0.03294478729367256,
  0.00595623767003417,
  0.0005460231914184988,
  -0.011670724488794804,
  -0.006910736206918955,
  -0.028626620769500732,
  0.02137409895658493,
  -0.013613066636025906,
  0.013663084246218204,
  0.030193831771612167,
  -0.016589101403951645,
  0.03938036039471626,
  0.03404517099261284,
  -0.04334840551018715,
  0.02255784347653389,
  -0.004343176726251841,
  -0.018306365236639977,
  0.007452591322362423,
  -0.04541579261422157,
  0.0032969797030091286,
  -0.003759640734642744,
  -0.0064105624333024025,
  0.06172146275639534,
  0.031794387847185135,
  -0.007256689947098494,
  0.013513032346963882,
  0.020890597254037857,
  -0.015105252154171467,
  0.02764294482767582,
  0.007219176739454269,
  0.012237588874995708,
  -0.022407792508602142,
  -0.005997919011861086,
  0.018273020163178444,
  -0.06725671887397766,
  -0.012037519365549088,
  -0.03998056799173355,
  0.008398753590881824,
  0.009836753830313683,
  -0.01715596579015255,
  0.0038388348184525967,
  0.046182725578546524,
  0.033828429877758026,
  -0.028609948232769966,
  -0.011495663784444332,
  -0.012646064162254333,
  -0.014988545328378677,
  -0.0069440812803804874,
  0.007244185544550419,
  0.002709275344386697,
  -0.028893379494547844,
  0.0025988202542066574,
  -0.012295941822230816,
  0.03577910736203194,
  2.7613768907031044e-05,
  -0.008978121913969517,
  -0.03771311417222023,
  -0.02274124138057232,
  -0.02392498590052128,
  -0.010270237922668457,
  0.003826330415904522,
  -0.02991040050983429,
  0.0034637043718248606,
  0.01770615763962269,
  0.026892684400081635,
  0.0081444988027215,
  -0.005005907267332077,
  0.014646759256720543,
  -0.035745762288570404,
  0.010345264337956905,
  0.03471206873655319,
  0.017356034368276596,
  -0.032611340284347534,
  0.014988545328378677,
  0.011245576664805412,
  -0.011362284421920776,
  -0.008928104303777218,
  -0.03041057474911213,
  0.002006947761401534,
  0.009378260932862759,
  -0.008373744785785675,
  0.012812788598239422,
  -0.0374796986579895,
  -0.007702678442001343,
  0.05305178090929985,
  0.013196255080401897,
  -0.033378273248672485,
  0.008515461347997189,
  0.012304278090596199,
  0.03711290657520294,
  0.026425855234265327,
  -0.011353948153555393,
  0.032477959990501404,
  0.026109078899025917,
  0.011262249201536179,
  0.01168739702552557,
  0.0015411608619615436,
  -0.018806539475917816,
  0.015280312858521938,
  0.03974715247750282,
  0.020323732867836952,
  0.020990632474422455,
  0.0032636348623782396,
  -0.022157704457640648,
  -0.02069052867591381,
  -0.003553318791091442,
  0.032111164182424545,
  -0.009086492471396923,
  -0.006172979716211557,
  0.003434527665376663,
  0.0029593624640256166,
  0.00171309569850564,
  -0.026942702010273933,
  0.0033949303906410933,
  0.023758260533213615,
  0.00394720584154129,
  -0.007931924425065517,
  -0.009061484597623348,
  0.004136855248361826,
  0.003772144904360175,
  0.0032636348623782396,
  0.016630783677101135,
  -0.0024591884575784206,
  -0.01052866131067276,
  -0.004689130466431379,
  0.002917681122198701,
  -0.056486308574676514,
  -0.020573820918798447,
  -0.013196255080401897,
  0.011078852228820324,
  0.03994722291827202,
  0.005868707317858934,
  -0.01025356538593769,
  0.011478991247713566,
  0.012679409235715866,
  0.001019625342451036,
  0.02899341471493244,
  0.002457104390487075,
  -0.047716591507196426,
  -0.028259826824069023,
  0.03731297329068184,
  -0.0034199391957372427,
  -0.007435918785631657,
  -0.001953804399818182,
  0.0344119668006897,
  0.026209114119410515,
  0.006960753817111254,
  0.008290383033454418,
  -0.0006830500205978751,
  0.0229246374219656,
  0.006481420248746872,
  -0.0159805566072464,
  -0.006860719062387943,
  0.003240710124373436,
  -0.017589449882507324,
  0.011170551180839539,
  0.0004428623360581696,
  -0.008006950840353966,
  0.02277458645403385,
  0.03811325132846832,
  -0.011762423440814018,
  -0.0212407186627388,
  0.014204938896000385,
  -0.0017881217645481229,
  -0.01404655072838068,
  -0.013212927617132664,
  -0.009044812060892582,
  -0.009319907054305077,
  0.011895802803337574,
  -0.00019394764967728406,
  -0.004784997086971998,
  -0.007948596961796284,
  -0.011812441051006317,
  0.007069124840199947,
  0.007723519112914801,
  0.00149531161878258,
  0.020173681899905205,
  -0.016864197328686714,
  0.0041868723928928375,
  0.0023216407280415297,
  0.005114278290420771,
  0.01007850468158722,
  -0.057386621832847595,
  -0.02174089290201664,
  0.010311919264495373,
  0.008561310358345509,
  0.007811049465090036,
  0.020940614864230156,
  0.03481210395693779,
  -0.006531437858939171,
  0.029927073046565056,
  0.030744023621082306,
  0.027559582144021988,
  0.014288301579654217,
  -0.032878100872039795,
  -0.007865235209465027,
  0.004605768248438835,
  0.0003084405907429755,
  -0.03661273047327995,
  -0.009820081293582916,
  -0.0442153736948967,
  0.010628695599734783,
  -0.011412302032113075,
  0.007956933230161667,
  0.006239669863134623,
  0.0040826695039868355,
  0.015672115609049797,
  -0.031160835176706314,
  0.024075038731098175,
  0.007469263859093189,
  -0.005372701678425074,
  -0.007652660831809044,
  -0.0005653007538057864,
  0.03149428591132164,
  -0.027309495955705643,
  -0.022024324163794518,
  0.0015578333986923099,
  0.005447727628052235,
  0.03654604032635689,
  0.026609253138303757,
  -0.003332408843562007,
  0.004597431980073452,
  0.01823967508971691,
  0.0018318869406357408,
  -0.0109788179397583,
  -0.006548110395669937,
  0.002955194329842925,
  0.0018120884196832776,
  0.030077125877141953,
  -0.0036200087051838636,
  -0.016689136624336243,
  0.000700243515893817,
  0.01499688159674406,
  0.01502189040184021,
  0.004901704378426075,
  0.030744023621082306,
  -0.003928449470549822,
  0.0014546725433319807,
  0.02232442982494831,
  0.006873223464936018,
  -0.03286142647266388,
  -0.01810629665851593,
  0.01502189040184021,
  0.014555061236023903,
  0.041781194508075714,
  0.02409171126782894,
  -0.00019017029262613505,
  -0.0023278926964849234,
  -0.011328939348459244,
  0.010662040673196316,
  0.03481210395693779,
  -0.0014984377194195986,
  0.017456069588661194,
  0.014580070041120052,
  0.00980340875685215,
  0.004172284156084061,
  0.007081629242748022,
  -0.014955200254917145,
  0.012370968237519264,
  0.02305801771581173,
  0.0005788471316918731,
  -0.0065439422614872456,
  0.010186875239014626,
  -0.004389026202261448,
  0.010495316237211227,
  -0.004272318910807371,
  0.032427944242954254,
  0.04061412066221237,
  0.00675234803929925,
  -0.0063063595443964005,
  -0.023308103904128075,
  0.02672595903277397,
  0.009078157134354115,
  0.00846544373780489,
  0.014988545328378677,
  0.0024320955853909254,
  0.01166238822042942,
  0.01087044645100832,
  -0.024891989305615425,
  0.05265164002776146,
  -0.015146933495998383,
  -0.0020861420780420303,
  0.006769020576030016,
  -0.04528241232037544,
  -0.005897884257137775,
  0.008019454777240753,
  0.029743675142526627,
  0.0066398088820278645,
  0.018006261438131332,
  0.023758260533213615,
  -0.007902747951447964,
  0.020307060331106186,
  -0.06492257863283157,
  -0.004880863707512617,
  0.058020174503326416,
  0.008894759230315685,
  -0.017906226217746735,
  -0.0018068782519549131,
  0.011979165486991405,
  0.0023466493003070354,
  -0.002016326179727912,
  0.005214313045144081,
  0.009178191423416138,
  0.03117750771343708,
  0.02545885182917118,
  0.011720742098987103,
  -0.06638975441455841,
  0.02949358895421028,
  0.011037170886993408,
  -0.015738805755972862,
  0.00548940896987915,
  0.008423762395977974,
  -0.014054886996746063,
  0.003180272411555052,
  0.023941658437252045,
  -0.013187918812036514,
  0.022024324163794518,
  -0.013588057830929756,
  -0.027926377952098846,
  -0.0019600566010922194,
  0.009311570785939693,
  -0.014505043625831604,
  -0.012837797403335571,
  -0.004559918772429228,
  -0.04434875398874283,
  0.01947343721985817,
  0.0014869753504171968,
  0.03252797573804855,
  0.007348388433456421,
  -0.021224046126008034,
  0.0033261566422879696,
  0.03654604032635689,
  0.010153530165553093,
  0.0407475009560585,
  -0.014013205654919147,
  0.013171246275305748,
  0.0196068175137043,
  -0.002673846436664462,
  -0.013988197781145573,
  0.025858990848064423,
  -0.008223692886531353,
  -0.027209460735321045,
  0.02622578665614128,
  -0.010562006384134293,
  -0.0011462318943813443,
  0.014129913412034512,
  -0.03521224483847618,
  -0.05015077069401741,
  -0.027059409767389297,
  -0.0025821479503065348,
  0.0010284825693815947,
  0.02800973877310753,
  0.026492545381188393,
  0.009219872765243053,
  0.018539780750870705,
  0.02342481166124344,
  -0.009928452782332897,
  0.005047588609158993,
  0.02060716599225998,
  -0.00038502970710396767,
  0.013096220791339874,
  0.04144774377346039,
  0.007815217599272728,
  -0.0008419594378210604,
  0.020090319216251373,
  0.04818342253565788,
  0.020590493455529213,
  0.026525890454649925,
  -0.016355687752366066,
  -0.03754638880491257,
  0.01397986151278019,
  -0.0011639463482424617,
  0.014113240875303745,
  0.00022807410277891904,
  0.01168739702552557,
  -0.05805351957678795,
  0.00033162572071887553,
  0.021124012768268585,
  0.006064608693122864,
  0.01395485270768404,
  0.005289338994771242,
  -0.025658922269940376,
  -0.004643281456083059,
  -0.007719350978732109,
  0.016855861991643906,
  0.004434875212609768,
  0.009253217838704586,
  0.03494548425078392,
  -0.0036491856444627047,
  0.03137757629156113,
  0.032011132687330246,
  0.0004160300886724144,
  0.03817994147539139,
  -0.02387496829032898,
  0.014204938896000385,
  -0.0011420637601986527,
  0.0016630783211439848,
  -0.004068081267178059,
  -0.008769716136157513,
  0.006493924651294947,
  -0.0036095883697271347,
  -0.06105456501245499,
  -0.017506087198853493,
  0.011745750904083252,
  -0.012354295700788498,
  -0.004747484344989061,
  -0.02867663837969303,
  -0.006764852441847324,
  -0.04301495850086212,
  -0.015397020615637302,
  -0.019089970737695694,
  0.012079199776053429,
  -0.010486979968845844,
  -0.008719698525965214,
  -0.016255652531981468,
  0.023174725472927094,
  -0.0037971537094563246,
  -0.015071907080709934,
  -0.012554365210235119,
  -0.013079548254609108,
  0.01575547829270363,
  -0.026709286496043205,
  -0.011412302032113075,
  0.009128173813223839,
  -0.018723176792263985,
  -0.03734631836414337,
  0.0011556101962924004,
  0.013396324589848518,
  -0.0020507131703197956,
  0.003988887183368206,
  0.022407792508602142,
  -0.006764852441847324,
  -0.020823907107114792,
  0.0027634608559310436,
  0.039780497550964355,
  -0.0011253913398832083,
  0.01742272451519966,
  0.01216256245970726,
  0.01002848707139492,
  -0.006218829192221165,
  -0.01670580916106701,
  0.029793692752718925,
  0.014646759256720543,
  0.013012858107686043,
  -0.00921153649687767,
  0.02932686358690262,
  0.00513928709551692,
  0.019640162587165833,
  0.007635988760739565,
  0.005268498789519072,
  -0.006781524978578091,
  0.013779791072010994,
  -0.031794387847185135,
  0.03081071376800537,
  -0.007548457942903042,
  0.00409517390653491,
  -0.014905182644724846,
  0.02160751260817051,
  -0.006248005665838718,
  -0.03227788954973221,
  0.000384248181944713,
  -0.00839458592236042,
  0.02319139800965786,
  -0.0014140333514660597,
  -0.008911431767046452,
  -0.024391815066337585,
  0.002709275344386697,
  -0.011612371541559696,
  0.0440153032541275,
  0.0006283434922806919,
  0.018956592306494713,
  0.019190005958080292,
  0.008628000505268574,
  -0.01075373962521553,
  0.0009899275610223413,
  -0.017639467492699623,
  0.0021257391199469566,
  -0.021407444030046463,
  -0.012120881117880344,
  -0.008577982895076275,
  0.023308103904128075,
  0.0003720043459907174,
  -0.039647120982408524,
  0.014029878191649914,
  0.007777704391628504,
  -0.033194877207279205,
  -0.009878435172140598,
  -0.009720046073198318,
  0.0016266072634607553,
  -0.033378273248672485,
  0.007665165234357119,
  -0.0043931943364441395,
  -0.0013410913525149226,
  0.034011825919151306,
  -0.004305663984268904,
  -0.03898021951317787,
  0.01810629665851593,
  -0.019173333421349525,
  -0.0009763811831362545,
  0.005080933216959238,
  -0.0269593745470047,
  -0.02137409895658493,
  0.000657520315144211,
  -0.04501565173268318,
  0.02972700260579586,
  0.004222301300615072,
  0.002131991321220994,
  -0.01059535052627325,
  0.0009649188723415136,
  -0.00675234803929925,
  -0.01929004117846489,
  0.016355687752366066,
  0.044682204723358154,
  0.01900660991668701,
  0.012521020136773586,
  -0.01770615763962269,
  0.04641614109277725,
  -0.024608556181192398,
  0.0018558535957708955,
  0.04811673238873482,
  -0.014680104330182076,
  0.008257037959992886,
  0.006910736206918955,
  0.05105108395218849,
  -0.006327200215309858,
  0.024158399552106857,
  0.009311570785939693,
  0.04514903202652931,
  0.038913529366254807,
  0.006197988521307707,
  -0.006769020576030016,
  -0.007740191649645567,
  0.002152831992134452,
  0.006056272424757481,
  0.0006934702978469431,
  -0.0012889899080619216,
  0.014513379894196987,
  0.013529704883694649,
  0.02832651697099209,
  ...],
 [0.0010050542186945677,
  -0.013880135491490364,
  -0.001135751255787909,
  0.03053629770874977,
  0.014446299523115158,
  0.01115889847278595,
  0.04244399443268776,
  0.03272789716720581,
  -0.006451523397117853,
  0.061730075627565384,
  -0.020728886127471924,
  0.02229953370988369,
  -0.03130335733294487,
  0.014327587559819221,
  -0.011971617117524147,
  -0.0008988985791802406,
  -0.026828840374946594,
  0.006392167881131172,
  -0.00864768959581852,
  -0.026463573798537254,
  0.029951872304081917,
  -0.029988398775458336,
  -0.007647772319614887,
  0.043393686413764954,
  0.0065108793787658215,
  0.004597794730216265,
  0.0459870807826519,
  0.0012864237651228905,
  -0.025586934760212898,
  0.0048900083638727665,
  -0.013660975731909275,
  0.03126683086156845,
  0.013076548464596272,
  -0.015514704398810863,
  -0.0409829281270504,
  0.025440827012062073,
  -0.011606350541114807,
  -0.03409764915704727,
  -0.0028034220449626446,
  0.014144954271614552,
  -0.012364278547465801,
  -0.044818226248025894,
  -0.02586088515818119,
  0.046206239610910416,
  0.01729537919163704,
  -0.0054150791838765144,
  0.01910345070064068,
  0.024217184633016586,
  -0.046169713139534,
  -0.019596559926867485,
  0.029038704931735992,
  0.002874192548915744,
  -0.0012030972866341472,
  0.01256517507135868,
  -0.0194321908056736,
  0.016016945242881775,
  0.041165560483932495,
  -0.03601529821753502,
  -0.031175514683127403,
  0.02980576455593109,
  -0.0003087645454797894,
  -0.044781699776649475,
  0.03908354043960571,
  0.04679066687822342,
  -0.013176997192203999,
  -0.028691701591014862,
  -0.017003165557980537,
  0.0015386861050501466,
  0.019249556586146355,
  0.060926489531993866,
  -0.019961826503276825,
  -0.0475212000310421,
  -0.008373740129172802,
  -0.01261083409190178,
  0.009350827895104885,
  -0.00018791260663419962,
  -0.00530549930408597,
  0.043320633471012115,
  -0.012181645259261131,
  0.02551388181746006,
  0.00826415978372097,
  -0.005360289476811886,
  0.010839289985597134,
  0.058807943016290665,
  0.02306659333407879,
  -0.007561021484434605,
  0.01135979499667883,
  -0.10687704384326935,
  0.03342190757393837,
  0.025751303881406784,
  0.016774874180555344,
  0.01545078307390213,
  -0.013478342443704605,
  -0.005798609461635351,
  0.025057297199964523,
  -0.016254369169473648,
  -0.014628932811319828,
  -0.008035868406295776,
  -0.011752457357943058,
  0.042955368757247925,
  -0.020071405917406082,
  0.004351240117102861,
  0.006090822629630566,
  -0.039010487496852875,
  -0.03196083754301071,
  -0.001073541701771319,
  -0.0018331824103370309,
  0.032545264810323715,
  -0.03634404018521309,
  -0.005113734398037195,
  -0.007122701033949852,
  -0.013176997192203999,
  0.001637993031181395,
  -0.01259257085621357,
  0.0006283729453571141,
  -0.04123861342668533,
  -0.004985890816897154,
  -0.017541933804750443,
  0.0067939613945782185,
  -4.055745012010448e-05,
  -0.008866850286722183,
  -0.001185975386761129,
  0.03208868205547333,
  -0.002981489757075906,
  0.009149931371212006,
  0.0026618812698870897,
  -0.02270132675766945,
  -0.026353994384407997,
  0.023304017260670662,
  -0.010373574681580067,
  -0.010729710571467876,
  -0.009044917300343513,
  5.136563413543627e-05,
  -0.006223232019692659,
  0.006478918716311455,
  0.02951355092227459,
  0.02618962526321411,
  0.03787815943360329,
  0.001463349792174995,
  0.039704494178295135,
  -0.022116901353001595,
  -0.024600714445114136,
  0.019651349633932114,
  0.0279976949095726,
  -0.05102776363492012,
  0.01150590181350708,
  0.04076376557350159,
  0.007250544615089893,
  -0.013889267109334469,
  0.025002507492899895,
  -0.0035339556634426117,
  -0.032855741679668427,
  0.0001852016430348158,
  -0.00430558156222105,
  -0.01654658280313015,
  -0.014446299523115158,
  -0.0008600890287198126,
  0.004848915617913008,
  0.024965981021523476,
  0.004999588243663311,
  0.009241248480975628,
  0.001885689445771277,
  -0.030463244765996933,
  -0.007451441138982773,
  -0.016300028190016747,
  -0.01866512931883335,
  0.04865352809429169,
  -0.025769567117094994,
  -0.04058113321661949,
  0.06856056302785873,
  -0.010665788315236568,
  -0.011953353881835938,
  0.024673767387866974,
  -0.03280095383524895,
  -0.013496605679392815,
  0.0008161428268067539,
  -0.005976676940917969,
  -0.0409829281270504,
  -0.0032896834891289473,
  -0.028162064030766487,
  -0.021002836525440216,
  0.020345356315374374,
  -0.013944057747721672,
  -0.006045164540410042,
  -0.015222490765154362,
  0.0005359147908166051,
  -0.003999670501798391,
  0.0019313478842377663,
  0.0006266607088036835,
  -0.029038704931735992,
  -0.0007071335567161441,
  -0.028527332469820976,
  0.027431530877947807,
  0.014848092570900917,
  0.011012791655957699,
  0.0022931902203708887,
  -0.01310394424945116,
  -0.0020009768195450306,
  -0.04131166636943817,
  -0.003983690403401852,
  -0.022445641458034515,
  0.006930936127901077,
  0.007204886060208082,
  0.01976093091070652,
  0.0029198508709669113,
  -0.007414914667606354,
  -0.012629097327589989,
  0.002109415363520384,
  0.003077372210100293,
  -0.015103779733181,
  0.005584015045315027,
  0.009834806434810162,
  -0.010154414921998978,
  0.007716259453445673,
  0.005456171929836273,
  0.013944057747721672,
  -0.05241577699780464,
  0.006214100401848555,
  -0.003075089305639267,
  0.003906071186065674,
  -7.237560203066096e-05,
  -0.019596559926867485,
  -0.013724897056818008,
  0.030371928587555885,
  -0.035796139389276505,
  0.005730121862143278,
  0.05511875078082085,
  0.048032574355602264,
  -0.016382213681936264,
  -0.011049318127334118,
  -0.01561515312641859,
  0.03225305303931236,
  -0.001187116838991642,
  -0.009300604462623596,
  -0.08401134610176086,
  -0.030791984871029854,
  0.00233656563796103,
  0.010894080623984337,
  0.0345359705388546,
  0.02692015841603279,
  -0.04573139548301697,
  0.0006643288652412593,
  -0.05058944225311279,
  -0.02737674117088318,
  0.01653745025396347,
  -0.009227550588548183,
  -0.004031631629914045,
  0.005310065113008022,
  0.013186128810048103,
  -0.013030890375375748,
  0.03340364247560501,
  -0.020582780241966248,
  -0.0315590463578701,
  -0.006159310229122639,
  0.018162887543439865,
  0.02045493759214878,
  0.004794125910848379,
  0.014254533685743809,
  -0.0330931656062603,
  -0.022445641458034515,
  0.002714388305321336,
  0.008556373417377472,
  0.02160552702844143,
  -0.003488297341391444,
  0.019833983853459358,
  0.028490804135799408,
  -0.0073555586859583855,
  -0.005186787340790033,
  -0.010729710571467876,
  -0.02268306352198124,
  0.01694837585091591,
  -0.0004360371676739305,
  0.011314136907458305,
  0.007775615435093641,
  0.057310350239276886,
  -0.010857553221285343,
  -0.001623154035769403,
  -0.004995022434741259,
  0.030335402116179466,
  -0.012072065845131874,
  -0.03239915892481804,
  0.01870165579020977,
  -0.013551395386457443,
  -0.03999670594930649,
  0.0025774133391678333,
  0.015962155535817146,
  0.0113050052896142,
  0.009423881769180298,
  -0.018829500302672386,
  -0.03459075838327408,
  -0.036526672542095184,
  0.012948705814778805,
  -0.005675331689417362,
  -0.022555220872163773,
  0.01081189513206482,
  -0.0011962485732510686,
  0.04072723910212517,
  0.000418629904743284,
  -0.022372586652636528,
  0.029933609068393707,
  -0.0017247438663616776,
  -0.044818226248025894,
  -0.00045287367538549006,
  0.011880300007760525,
  -0.00593558419495821,
  0.0036503844894468784,
  -0.01908518746495247,
  0.006949199363589287,
  -0.04164040833711624,
  -0.042663153260946274,
  0.012702150270342827,
  -0.029714448377490044,
  -0.003207498462870717,
  -0.019925300031900406,
  -5.9284502640366554e-05,
  -0.053365468978881836,
  0.014921145513653755,
  0.026500102132558823,
  -0.018866026774048805,
  -0.04383200779557228,
  -0.0013275162782520056,
  0.007022252772003412,
  -0.05431516095995903,
  -0.007410348858684301,
  -0.01834552176296711,
  0.01764238253235817,
  -0.032892268151044846,
  0.041129034012556076,
  0.023157911375164986,
  -0.025988727807998657,
  0.0017761094495654106,
  -0.014172349125146866,
  0.049749329686164856,
  -0.015441651456058025,
  0.022536957636475563,
  -0.012811730615794659,
  -0.009460408240556717,
  -0.0038626957684755325,
  -0.02408934012055397,
  -0.023797128349542618,
  0.0266462080180645,
  -0.021349839866161346,
  -0.0002865060814656317,
  -0.020199250429868698,
  -0.023267490789294243,
  0.02914828434586525,
  0.03239915892481804,
  -0.013734028674662113,
  0.016994034871459007,
  0.004328410606831312,
  0.0003307376173324883,
  -0.05756603553891182,
  0.031778205186128616,
  0.025696514174342155,
  0.023285754024982452,
  0.0058260043151676655,
  -0.0425170473754406,
  -0.00807696022093296,
  0.00025996717158704996,
  0.01649179309606552,
  0.003974558785557747,
  -0.012455595657229424,
  0.043357159942388535,
  0.0147659070789814,
  0.03152251988649368,
  0.028947388753294945,
  0.009067746810615063,
  -0.008223067037761211,
  -0.018455101177096367,
  0.011515033431351185,
  -0.04788646847009659,
  0.02405281364917755,
  0.009661304764449596,
  0.027778534218668938,
  -0.017742831259965897,
  -0.030408455058932304,
  0.01039183884859085,
  0.00027880124980583787,
  0.008373740129172802,
  0.005013285670429468,
  -0.005501830019056797,
  0.009515197947621346,
  0.000745943165384233,
  -0.023486651480197906,
  0.00848331954330206,
  0.014090163633227348,
  -0.02012619748711586,
  -0.03260005638003349,
  -0.005798609461635351,
  0.02195253036916256,
  -0.01403537392616272,
  -0.009496934711933136,
  0.007985644042491913,
  0.01348747406154871,
  -0.0025774133391678333,
  0.002362819155678153,
  0.0076751671731472015,
  0.01623610593378544,
  0.04302842170000076,
  -0.030207559466362,
  -0.003960860893130302,
  -0.06140133738517761,
  -0.00718205701559782,
  -0.015889102593064308,
  -0.030262349173426628,
  -0.0016802269965410233,
  0.015496441163122654,
  0.006688946858048439,
  -0.011496770195662975,
  -0.001992986537516117,
  0.00026624518795870245,
  -0.013377893716096878,
  -0.016793137416243553,
  -0.013478342443704605,
  0.010190941393375397,
  0.0018320409581065178,
  0.03126683086156845,
  -0.01686619035899639,
  -0.024965981021523476,
  -0.037238944321870804,
  -0.007908024825155735,
  -0.0006894409307278693,
  0.02193426713347435,
  -0.01059273537248373,
  -0.002928982488811016,
  0.026353994384407997,
  0.010519681498408318,
  0.01405363716185093,
  -0.0032896834891289473,
  -0.04142124578356743,
  -0.015971288084983826,
  0.020692359656095505,
  0.02156900055706501,
  0.04200567305088043,
  0.0065108793787658215,
  -0.010601866990327835,
  0.025294721126556396,
  -0.019870510324835777,
  0.008145447820425034,
  -0.024345027282834053,
  -0.010638393461704254,
  -0.00601776922121644,
  0.046754140406847,
  -0.00319380103610456,
  -0.051210395991802216,
  0.025331247597932816,
  -0.025093823671340942,
  3.0427003366639838e-05,
  0.001981572015210986,
  0.01238254178315401,
  -0.03095635585486889,
  0.02191600389778614,
  0.018939079716801643,
  -0.007990209385752678,
  -0.022189954295754433,
  -0.0005952705978415906,
  -0.015669941902160645,
  0.06574801355600357,
  -0.0036115748807787895,
  0.008341779001057148,
  0.04967627674341202,
  0.021824687719345093,
  0.009853069670498371,
  -0.016281764954328537,
  -0.003269137116149068,
  -0.0063282460905611515,
  -0.039704494178295135,
  0.02986055426299572,
  0.01348747406154871,
  0.005095470696687698,
  0.010528813116252422,
  -0.01979745738208294,
  -0.0024176090955734253,
  0.0020009768195450306,
  0.007487968076020479,
  0.02728542499244213,
  0.003125313436612487,
  -0.02591567486524582,
  0.007337295450270176,
  -0.027522848919034004,
  0.0014804716920480132,
  0.029604868963360786,
  0.007821273989975452,
  0.04196914657950401,
  0.02334054373204708,
  -0.00031390110962092876,
  0.04635234922170639,
  -0.012254699133336544,
  -0.04412421956658363,
  -0.004250791389495134,
  -0.005967545323073864,
  -0.0024792479816824198,
  -0.025787830352783203,
  -0.024162394925951958,
  0.02807074785232544,
  -0.017185799777507782,
  -0.01150590181350708,
  -0.02690189518034458,
  -0.003237176453694701,
  0.04430685564875603,
  0.01840031147003174,
  -0.015386860817670822,
  0.018108097836375237,
  0.021769898012280464,
  -0.007739088963717222,
  -0.03267310932278633,
  0.03711109980940819,
  -0.0015866273315623403,
  0.003349039237946272,
  0.0233770702034235,
  0.02881954424083233,
  -0.015414255671203136,
  -0.0024381554685533047,
  -0.006218666210770607,
  -0.01261083409190178,
  0.0067620002664625645,
  0.013341367244720459,
  -0.007725391536951065,
  -0.004093270283192396,
  -0.039047013968229294,
  -0.03409764915704727,
  -0.01583431288599968,
  -0.001624295487999916,
  0.017742831259965897,
  0.0015934761613607407,
  0.0013640429824590683,
  0.026353994384407997,
  -0.0014827545965090394,
  0.07111743092536926,
  0.002746349200606346,
  -0.015669941902160645,
  0.030481508001685143,
  0.008035868406295776,
  -0.008364608511328697,
  -0.010382707230746746,
  -0.0019895622972398996,
  -0.018792973831295967,
  -0.005657068453729153,
  -0.013350498862564564,
  0.007314466405659914,
  0.011944222263991833,
  0.0019062357023358345,
  0.039047013968229294,
  -0.03128509595990181,
  -0.006606761831790209,
  -0.012291225604712963,
  0.0034152439329773188,
  -0.045548759400844574,
  0.008191106840968132,
  -0.006483484525233507,
  0.024363290518522263,
  0.03051803447306156,
  -0.027194108814001083,
  -0.004159474745392799,
  0.009615646675229073,
  0.00021987344371154904,
  -0.02480161003768444,
  -0.0041891527362167835,
  -0.00746970484033227,
  0.014619801193475723,
  0.023998023942112923,
  0.00936909206211567,
  0.013478342443704605,
  0.007474270649254322,
  -0.001320667564868927,
  0.016619635745882988,
  -0.024253711104393005,
  0.011021923273801804,
  -0.010647525079548359,
  -0.03570482134819031,
  -0.0017772509017959237,
  -0.013524000532925129,
  0.006593064405024052,
  -0.028253382071852684,
  0.024527661502361298,
  0.01761498861014843,
  0.015962155535817146,
  -0.029696185141801834,
  0.005561186000704765,
  0.00968869961798191,
  -0.015313807874917984,
  0.0014188329223543406,
  0.015624284744262695,
  -0.019596559926867485,
  -0.01864686608314514,
  0.008378305472433567,
  0.0025774133391678333,
  -0.044745173305273056,
  -0.003152708522975445,
  0.01838204823434353,
  -0.019231293350458145,
  0.01827246882021427,
  0.011560692451894283,
  -0.0020626154728233814,
  0.025678250938653946,
  0.024600714445114136,
  -0.014291061088442802,
  -0.023888444527983665,
  0.007834970951080322,
  -0.03596051037311554,
  -0.015432518906891346,
  0.04123861342668533,
  0.018546417355537415,
  0.026883631944656372,
  0.006173007655888796,
  0.01981572061777115,
  0.04711940884590149,
  -0.01081189513206482,
  0.020765414461493492,
  -0.022792644798755646,
  0.035832665860652924,
  0.0027417833916842937,
  -0.015907365828752518,
  -0.017103614285588264,
  0.018847763538360596,
  0.013569658622145653,
  -0.016656162217259407,
  0.016482660546898842,
  -0.02224474400281906,
  0.025678250938653946,
  0.02405281364917755,
  -0.022007320076227188,
  0.005647936835885048,
  -0.013533132150769234,
  -0.007063345517963171,
  0.003479165490716696,
  -0.014674590900540352,
  -0.015642547979950905,
  -0.0007271090871654451,
  0.020071405917406082,
  0.003100201254710555,
  -0.016637898981571198,
  0.008866850286722183,
  -0.02449113503098488,
  -0.0034609022550284863,
  0.01827246882021427,
  0.003878676099702716,
  0.014738512225449085,
  -0.006465221289545298,
  -0.006150178611278534,
  -0.0018046459881588817,
  -0.03559524193406105,
  -0.008159145712852478,
  -0.08707959204912186,
  -0.03482818230986595,
  -0.00630541704595089,
  -0.019961826503276825,
  0.030664142221212387,
  0.005629673600196838,
  -0.01834552176296711,
  0.03426202014088631,
  0.003759964369237423,
  0.04211525246500969,
  0.008309817872941494,
  0.03526650369167328,
  0.0019564598333090544,
  0.017450617626309395,
  -0.019340872764587402,
  -0.026025254279375076,
  -0.003075089305639267,
  -0.019322609528899193,
  -0.02118547074496746,
  -0.0024632674176245928,
  -0.007771049626171589,
  -0.012455595657229424,
  0.015085515566170216,
  -0.00828698929399252,
  -0.010839289985597134,
  0.0005567463813349605,
  0.01756932958960533,
  -0.0032531567849218845,
  -0.00575295090675354,
  -0.0226100105792284,
  0.002310312120243907,
  -0.015642547979950905,
  -0.036526672542095184,
  -0.005355723202228546,
  0.016756610944867134,
  -0.008414831943809986,
  0.008716177195310593,
  0.006529142614454031,
  0.010702314786612988,
  -0.0016277198446914554,
  0.027851589024066925,
  0.011432848870754242,
  0.00746513856574893,
  -0.012711281888186932,
  -0.007095306180417538,
  -0.019980089738965034,
  -0.02303006686270237,
  -0.02198905684053898,
  0.004166323691606522,
  0.01650092378258705,
  0.017779357731342316,
  -0.013231786899268627,
  0.010236600413918495,
  0.03513865917921066,
  -0.010894080623984337,
  -0.033586274832487106,
  0.004995022434741259,
  -0.018957342952489853,
  0.01613565720617771,
  0.011332400143146515,
  0.0008606597548350692,
  0.01766977831721306,
  0.0035955943167209625,
  0.014373245649039745,
  -0.01368837058544159,
  0.0025705646257847548,
  -0.01651005633175373,
  -0.025751303881406784,
  0.03479165583848953,
  0.04134819284081459,
  0.05131997540593147,
  0.030426718294620514,
  0.003139010863378644,
  0.006775697693228722,
  -0.038316480815410614,
  -0.0012704433174803853,
  -0.004255357198417187,
  0.031504254788160324,
  0.0073829535394907,
  -0.030773721635341644,
  0.022847434505820274,
  -0.0194321908056736,
  0.011432848870754242,
  0.01840031147003174,
  0.006926370318979025,
  0.01945045404136181,
  0.006647854577749968,
  0.021459421142935753,
  -0.01438237726688385,
  0.008821191266179085,
  0.005953847896307707,
  0.0024472870863974094,
  0.014537615701556206,
  -0.007784747052937746,
  0.021477684378623962,
  -0.002837665844708681,
  0.006483484525233507,
  0.02591567486524582,
  -0.01039183884859085,
  -0.004449405241757631,
  -0.009999176487326622,
  0.0035818968899548054,
  0.012108592316508293,
  0.006478918716311455,
  0.01022746879607439,
  0.007848668843507767,
  0.006218666210770607,
  0.03912006691098213,
  -0.03922964632511139,
  0.01259257085621357,
  -0.04160388186573982,
  -0.008743572048842907,
  0.04931100830435753,
  0.012556043453514576,
  0.013907530345022678,
  -0.01242820080369711,
  0.007214017678052187,
  0.012848257087171078,
  0.008072394877672195,
  0.02407107688486576,
  0.027395004406571388,
  0.0034266584552824497,
  -0.03561350703239441,
  0.005442474503070116,
  -0.03743983805179596,
  0.028691701591014862,
  0.03440812602639198,
  -0.00032360348268412054,
  0.0025933936703950167,
  0.017815884202718735,
  -0.009022087790071964,
  0.014071900397539139,
  0.01614478975534439,
  0.0076066795736551285,
  -0.03192431107163429,
  -0.02228127047419548,
  -0.012912179343402386,
  -0.00068601657403633,
  -0.005921886768192053,
  -0.004150343127548695,
  0.00504068098962307,
  0.0037165889516472816,
  -0.05172177031636238,
  -0.010547076351940632,
  -0.008328081108629704,
  -0.0005550342029891908,
  0.026116570457816124,
  -0.005597712472081184,
  0.0233770702034235,
  0.040398500859737396,
  0.009606515057384968,
  -0.007341861259192228,
  -0.00292669958434999,
  0.030755458399653435,
  0.005237011704593897,
  0.005565751809626818,
  0.009277774952352047,
  0.011451112106442451,
  0.003600160125643015,
  0.007140964735299349,
  0.03199736401438713,
  0.007793878670781851,
  0.0019016698934137821,
  -0.036471884697675705,
  -0.003086503827944398,
  -0.04054460674524307,
  -0.0009342838311567903,
  -0.00107696617487818,
  0.009081443771719933,
  0.021769898012280464,
  0.040069758892059326,
  0.025440827012062073,
  0.006757434457540512,
  0.00773452315479517,
  -0.04134819284081459,
  -0.022756116464734077,
  0.010364443063735962,
  0.0194321908056736,
  0.02018098719418049,
  0.014994199387729168,
  -0.020948046818375587,
  0.011350663378834724,
  0.032198261469602585,
  0.03711109980940819,
  -0.009067746810615063,
  -0.03552218899130821,
  -0.028600385412573814,
  -0.0266462080180645,
  0.03860869258642197,
  -0.004467668477445841,
  0.007109003607183695,
  0.011478506959974766,
  0.017167536541819572,
  -0.009095141664147377,
  0.007077042944729328,
  0.024271974340081215,
  0.0026961250696331263,
  0.008414831943809986,
  -0.0002211575920227915,
  0.020034879446029663,
  0.007688864599913359,
  -0.03345843404531479,
  0.02297527715563774,
  -0.023924970999360085,
  -0.003828451968729496,
  -0.010556208901107311,
  -0.016455266624689102,
  0.0012373410863801837,
  0.006232363637536764,
  0.030846774578094482,
  0.020710622891783714,
  0.022756116464734077,
  0.026700997725129128,
  -0.01584344357252121,
  0.03422549366950989,
  0.025659987702965736,
  -0.02518514171242714,
  0.014875487424433231,
  -0.010455760173499584,
  -0.03838953375816345,
  -0.011259347200393677,
  0.040069758892059326,
  0.006437825970351696,
  -0.004330693744122982,
  -0.015195095911622047,
  -0.010382707230746746,
  -0.007757352199405432,
  0.0007904600352048874,
  -0.018089834600687027,
  -0.003205215558409691,
  -0.0183272585272789,
  0.016802269965410233,
  0.003565916558727622,
  0.0054835667833685875,
  -0.014501089230179787,
  -0.026134833693504333,
  0.0034335071686655283,
  -0.03057282604277134,
  0.024874664843082428,
  -0.048361316323280334,
  -0.010601866990327835,
  0.019249556586146355,
  -0.009318867698311806,
  -0.037257205694913864,
  0.010894080623984337,
  -0.015167701058089733,
  0.01530467625707388,
  0.023176174610853195,
  0.003771378891542554,
  0.0015192813007161021,
  -0.024381553754210472,
  0.013387025333940983,
  0.027760270982980728,
  0.00762950861826539,
  -0.0027395004872232676,
  -0.01730451174080372,
  -0.00973435863852501,
  0.012556043453514576,
  -0.003682345151901245,
  -0.013569658622145653,
  0.010419233702123165,
  -0.018089834600687027,
  0.0038992222398519516,
  -0.005086339078843594,
  0.025422563776373863,
  0.02118547074496746,
  0.02120373398065567,
  -0.0071363989263772964,
  -0.0020135329104959965,
  -0.016053473576903343,
  -0.006675249431282282,
  0.013131339102983475,
  -0.019888773560523987,
  0.005561186000704765,
  0.025605197995901108,
  0.018957342952489853,
  -0.04229788854718208,
  -0.0359787717461586,
  0.016784006729722023,
  0.0016277198446914554,
  0.009588251821696758,
  -0.013414420187473297,
  0.0018114947015419602,
  -0.030024925246834755,
  -0.0038193203508853912,
  0.0036663648206740618,
  0.002903870539739728,
  -0.007647772319614887,
  -0.0172223262488842,
  0.01006309874355793,
  0.012665623798966408,
  0.007031384389847517,
  -0.021386366337537766,
  -0.010674919933080673,
  0.004499629605561495,
  -0.0005929876933805645,
  -0.01277520414441824,
  0.004346674308180809,
  0.048397842794656754,
  0.0038444322999566793,
  -0.03680062294006348,
  -0.027760270982980728,
  -0.0023148779291659594,
  -0.03817037492990494,
  0.004164040554314852,
  -0.016674425452947617,
  0.014318455941975117,
  -0.005533791147172451,
  0.03228957951068878,
  -0.0172223262488842,
  -0.004586380440741777,
  0.050516389310359955,
  -0.019614823162555695,
  -0.03480992093682289,
  0.014738512225449085,
  0.007159227970987558,
  0.019121713936328888,
  -0.010492286644876003,
  -0.023139648139476776,
  -0.02266480028629303,
  -0.008844020776450634,
  -0.00936909206211567,
  0.015313807874917984,
  0.0044402736239135265,
  0.003638969734311104,
  0.007346427068114281,
  0.019139977172017097,
  -0.022883960977196693,
  0.003618423594161868,
  -0.008670519106090069,
  0.03528476506471634,
  0.02268306352198124,
  0.011213688179850578,
  -0.015103779733181,
  0.012802598997950554,
  0.00789889320731163,
  0.04861700162291527,
  0.03700152039527893,
  -0.008049565367400646,
  -0.002454135799780488,
  0.03634404018521309,
  0.03893743455410004,
  -0.020016616210341454,
  -5.486134978127666e-05,
  -0.018774710595607758,
  0.0006677532219327986,
  0.009232116863131523,
  -0.03988712653517723,
  0.017386697232723236,
  0.004362654406577349,
  -0.02441808022558689,
  -0.010839289985597134,
  -0.008862284012138844,
  0.024911191314458847,
  -0.012135987170040607,
  0.006250626873224974,
  -0.017459750175476074,
  ...],
 [0.007534371688961983,
  0.005438145250082016,
  -0.019238615408539772,
  0.009694576263427734,
  -0.005520940758287907,
  0.013721438124775887,
  -0.02991168014705181,
  0.011342954821884632,
  0.014835410751402378,
  0.031010599806904793,
  -0.020984843373298645,
  0.03110092133283615,
  -0.018771950155496597,
  0.009408555924892426,
  0.008836515247821808,
  -0.013608534820377827,
  -0.03817615658044815,
  0.011726823635399342,
  0.020984843373298645,
  -0.01193004846572876,
  0.06449000537395477,
  -0.015610676258802414,
  -0.004768256098031998,
  -0.005054275970906019,
  0.03901916369795799,
  0.01336014922708273,
  0.017793461680412292,
  0.03362994268536568,
  -0.03197403624653816,
  0.024627836421132088,
  -0.021617097780108452,
  0.02482353337109089,
  0.0012814452638849616,
  0.027066532522439957,
  0.0001524186081951484,
  0.03781486675143242,
  0.008693506009876728,
  -0.0028244485147297382,
  0.003283585887402296,
  -0.017703138291835785,
  -0.02116548642516136,
  -0.030303075909614563,
  -0.037905190140008926,
  0.019961191341280937,
  0.01180209219455719,
  0.04007292166352272,
  0.006597279570996761,
  -0.0071053411811590195,
  -0.008896730840206146,
  0.01910313218832016,
  -0.024552566930651665,
  0.002798104425892234,
  0.023002037778496742,
  0.028963297605514526,
  -0.01607733964920044,
  -0.008603183552622795,
  0.04762987419962883,
  -0.033720262348651886,
  -0.006811794359236956,
  -0.009754790924489498,
  0.03284715116024017,
  -0.034141767770051956,
  0.011117149144411087,
  -0.017597762867808342,
  -0.0033005212899297476,
  0.0014827883569523692,
  0.0029448780696839094,
  0.01120747160166502,
  0.037483684718608856,
  0.03718261420726776,
  0.0023860097862780094,
  -0.002058591926470399,
  0.01643862947821617,
  0.012885957956314087,
  -0.0007004669751040637,
  0.005867175292223692,
  -0.023483755066990852,
  0.029610607773065567,
  0.026614923030138016,
  -0.00781286507844925,
  0.015414977446198463,
  0.01001823041588068,
  0.00026155784144066274,
  0.008211787790060043,
  -0.0030954149551689625,
  -0.001138435211032629,
  0.006747816223651171,
  -0.060184650123119354,
  0.0003873972746077925,
  0.030378345400094986,
  -0.00048265891382470727,
  -0.016860131174325943,
  0.0025516003370285034,
  -0.014233263209462166,
  -0.014353692531585693,
  0.020759036764502525,
  0.012968753464519978,
  -0.00401933491230011,
  -0.04094603657722473,
  0.026524601504206657,
  0.0029467595741152763,
  0.010627904906868935,
  0.0020830542780458927,
  -0.027548251673579216,
  0.01129026710987091,
  0.005415564868599176,
  0.006680074613541365,
  0.0383266918361187,
  0.0027585886418819427,
  0.026810619980096817,
  -0.03630949929356575,
  -0.019313883036375046,
  -0.028029968962073326,
  -0.008783827535808086,
  -0.018546145409345627,
  0.04425784572958946,
  0.006555881816893816,
  0.027518143877387047,
  -0.01010102592408657,
  -0.02965576760470867,
  0.0070639438927173615,
  -0.026057936251163483,
  0.009619307704269886,
  -0.033750370144844055,
  0.001979560125619173,
  -0.017928944900631905,
  -0.010334357619285583,
  -0.002495148917660117,
  -0.012953699566423893,
  -0.0005593386595137417,
  -0.01868162862956524,
  -0.014639712870121002,
  -0.015640784054994583,
  0.03299768641591072,
  -0.011297794058918953,
  -0.011297794058918953,
  0.010755861178040504,
  0.015083796344697475,
  -0.0034115423914045095,
  0.005927389953285456,
  -0.028632117435336113,
  0.0320342518389225,
  0.0068757724948227406,
  -0.0070074922405183315,
  -0.028120292350649834,
  0.0013181386748328805,
  0.04230086877942085,
  -0.013646169565618038,
  0.0031010599341243505,
  0.028707385063171387,
  0.004282774403691292,
  0.01238165982067585,
  0.013864448294043541,
  -0.03444284200668335,
  0.028872976079583168,
  0.012336498126387596,
  -0.011666608974337578,
  0.02187301032245159,
  0.0036222939379513264,
  0.0035394986625760794,
  -0.036038532853126526,
  0.011779512278735638,
  -0.03835679963231087,
  -0.011403169482946396,
  0.0007338673458434641,
  -0.044077202677726746,
  -0.02167731337249279,
  -0.03263640031218529,
  -0.007538135163486004,
  0.06659752130508423,
  -0.02258053421974182,
  0.006548354867845774,
  0.0036524012684822083,
  0.023664399981498718,
  -0.016935400664806366,
  0.013495632447302341,
  0.013089182786643505,
  -0.007541898172348738,
  0.013924662955105305,
  -0.0014150467468425632,
  -0.004820943810045719,
  -0.007353727240115404,
  -0.011350481770932674,
  -0.033268652856349945,
  -0.00729351257905364,
  0.04762987419962883,
  -0.022460104897618294,
  0.01598701812326908,
  -0.008708558976650238,
  -0.02318268083035946,
  -0.00182337814476341,
  -0.022023547440767288,
  -0.008226840756833553,
  0.016905292868614197,
  0.06846418231725693,
  0.004918793216347694,
  -0.005727928597480059,
  -0.029640713706612587,
  0.00882898923009634,
  -0.0117192966863513,
  -0.008919310756027699,
  -0.0016323844902217388,
  -0.018290232867002487,
  -0.012005317024886608,
  0.003951593302190304,
  -0.02057839371263981,
  0.01053005550056696,
  0.008520388044416904,
  0.00791071355342865,
  -0.02212892286479473,
  -0.011275213211774826,
  -0.027232123538851738,
  -0.007026309613138437,
  -0.003992991056293249,
  0.02730739302933216,
  0.004470945801585913,
  -0.00976231787353754,
  -0.001887356280349195,
  0.02032248117029667,
  0.016860131174325943,
  -0.0003081301983911544,
  -0.022204192355275154,
  -0.003016382921487093,
  0.023920312523841858,
  -0.02858695574104786,
  -0.00033423895365558565,
  0.01144833117723465,
  0.02980630472302437,
  -0.025862237438559532,
  -0.01621282286942005,
  -0.02759341336786747,
  -0.013601007871329784,
  -0.020217103883624077,
  0.011373062618076801,
  -0.03965141624212265,
  0.008234367705881596,
  -0.003947829827666283,
  -0.01453433744609356,
  0.017477333545684814,
  0.014918206259608269,
  -0.0020849360153079033,
  0.00803866982460022,
  -0.024386975914239883,
  -0.022821392863988876,
  0.012020370922982693,
  0.008136519230902195,
  -0.051754582673311234,
  -0.012622518464922905,
  -0.004312882199883461,
  0.019449366256594658,
  -0.0026964920107275248,
  0.009318233467638493,
  0.010138659738004208,
  0.016965508460998535,
  -0.0024631598498672247,
  0.029866520315408707,
  0.03959120437502861,
  -0.01412036083638668,
  -0.02881276048719883,
  -0.0014733797870576382,
  0.035677243024110794,
  0.007346200291067362,
  -0.023363325744867325,
  0.0032421883661299944,
  0.04416752606630325,
  -0.0010481131030246615,
  0.019539689645171165,
  0.0016700186533853412,
  -0.008558022789657116,
  0.014255844056606293,
  0.04296322911977768,
  -0.025681594386696815,
  0.046425577253103256,
  0.00786555279046297,
  0.013841867446899414,
  -0.0008566490141674876,
  -0.012682733125984669,
  -0.031401995569467545,
  0.012878431007266045,
  -0.002474450273439288,
  0.02270096354186535,
  0.03022780828177929,
  -0.044137418270111084,
  0.00802361685782671,
  -0.0031048234086483717,
  0.007696198765188456,
  -0.033660050481557846,
  -0.008098884485661983,
  -0.017552601173520088,
  -0.005423091351985931,
  -0.09923391789197922,
  0.0018516037380322814,
  0.011568760499358177,
  -0.0044972896575927734,
  0.030754687264561653,
  -0.016634326428174973,
  0.04398687928915024,
  0.024100957438349724,
  -0.028707385063171387,
  0.018967648968100548,
  -0.020623553544282913,
  -0.01204295177012682,
  -0.019343990832567215,
  -0.020819252356886864,
  -0.017522495239973068,
  0.022550426423549652,
  0.02810523845255375,
  -0.00905479397624731,
  -0.013081655837595463,
  -0.009739737026393414,
  0.03926002234220505,
  -0.01058274321258068,
  -0.004342989530414343,
  -0.01522680651396513,
  -0.04094603657722473,
  0.015023581683635712,
  0.013299934566020966,
  0.005189759191125631,
  0.03775465115904808,
  0.01884721964597702,
  0.00816662609577179,
  0.010048337280750275,
  -0.01836550049483776,
  0.048894383013248444,
  -0.009664468467235565,
  -0.0001768808433553204,
  -0.008889203891158104,
  -0.00980747863650322,
  0.015083796344697475,
  0.0015430031344294548,
  -0.02331816405057907,
  -0.03865787386894226,
  0.018666574731469154,
  -0.0074628666043281555,
  0.037574008107185364,
  -0.015475193038582802,
  -0.04871373996138573,
  0.0026080517563968897,
  0.020894519984722137,
  -0.008106411434710026,
  -0.02000635303556919,
  -0.03405144438147545,
  -0.02434181608259678,
  0.00718813668936491,
  -0.0016276801470667124,
  -0.03293747082352638,
  0.0327267199754715,
  0.005671477410942316,
  0.033720262348651886,
  0.011154783889651299,
  0.0005758036277256906,
  -0.030890170484781265,
  0.03393101692199707,
  0.021180540323257446,
  0.023077305406332016,
  0.010650484822690487,
  -0.03161274641752243,
  -0.028993405401706696,
  -0.004583848174661398,
  0.020653661340475082,
  -0.014135414734482765,
  -0.04091592878103256,
  -0.0202923733741045,
  -0.010191347450017929,
  0.002909125527366996,
  0.010078445076942444,
  0.024236440658569336,
  -0.011809619143605232,
  0.021105272695422173,
  -0.011659082025289536,
  -0.03591810166835785,
  0.009611780755221844,
  -0.04868363216519356,
  0.015535407699644566,
  -0.03820626437664032,
  -0.001242870232090354,
  -0.0027454164810478687,
  0.010191347450017929,
  -0.0454019270837307,
  0.003876324975863099,
  0.007730069570243359,
  0.004361806437373161,
  0.01120747160166502,
  -0.010319303721189499,
  0.024236440658569336,
  -0.023137520998716354,
  0.00308788800612092,
  0.01669454202055931,
  -0.012554776854813099,
  -0.004753202199935913,
  -0.016709595918655396,
  0.014105306938290596,
  -0.021797742694616318,
  -0.03919980674982071,
  0.02103000320494175,
  -0.0031876186840236187,
  0.03399122878909111,
  -0.0014225736958906054,
  0.008821462281048298,
  -0.009190277196466923,
  -0.006759106647223234,
  -0.03769443929195404,
  0.0005828600260429084,
  0.011147256940603256,
  -0.00802361685782671,
  -0.026072990149259567,
  -0.04064496234059334,
  -0.024402029812335968,
  0.002839502179995179,
  0.004534923937171698,
  0.02224935218691826,
  -0.006890826392918825,
  0.03868798166513443,
  -0.026569761335849762,
  0.017838621512055397,
  -0.020337535068392754,
  -0.0073048025369644165,
  0.011425750330090523,
  0.024778373539447784,
  -0.020789144560694695,
  0.002169613027945161,
  0.014090253040194511,
  0.006484376732259989,
  0.024672996252775192,
  -0.01881711184978485,
  0.054976072162389755,
  -0.006367710418999195,
  0.008941891603171825,
  0.003283585887402296,
  -0.0025120843201875687,
  0.006153195630759001,
  -0.03432241082191467,
  -0.004708041436970234,
  0.03387080132961273,
  0.0015646428801119328,
  -0.007873079739511013,
  0.016378413885831833,
  0.003268532222136855,
  -0.01141822338104248,
  -0.008964472450315952,
  0.01646873541176319,
  -0.0012014725944027305,
  0.022294513881206512,
  0.015460139140486717,
  -0.0069999657571315765,
  0.006604806054383516,
  0.016062285751104355,
  -0.011041881516575813,
  -0.012547249905765057,
  -0.019900977611541748,
  -0.030724579468369484,
  0.01604723371565342,
  -0.06009432673454285,
  -0.020081620663404465,
  0.00420374283567071,
  0.0024198805913329124,
  0.021586989983916283,
  0.046034183353185654,
  -0.01698056235909462,
  -0.006879535969346762,
  0.03131167218089104,
  0.03579767420887947,
  0.023694507777690887,
  -0.0040343888103961945,
  0.0007715015672147274,
  -0.0413675382733345,
  0.0032704139593988657,
  0.03236543387174606,
  -0.01294617261737585,
  0.008083831518888474,
  -0.03206435963511467,
  -0.0017960933037102222,
  -0.0551266111433506,
  0.02611815184354782,
  -0.014428961090743542,
  -0.020608501508831978,
  -0.06954804807901382,
  -0.007632220629602671,
  -0.04669654369354248,
  -0.03528584912419319,
  0.01497089397162199,
  0.002879018196836114,
  -0.0061117978766560555,
  -0.0165590588003397,
  0.017507441341876984,
  0.007222007494419813,
  0.0008373614982701838,
  0.02730739302933216,
  -0.0004958309000357985,
  -0.0076623279601335526,
  0.01022898219525814,
  -0.018967648968100548,
  0.005656423978507519,
  -0.009927907958626747,
  0.004640299826860428,
  -0.01205800473690033,
  -0.022746125236153603,
  -0.033780477941036224,
  0.03302779421210289,
  0.019012808799743652,
  0.030468666926026344,
  -0.01598701812326908,
  0.015317128971219063,
  0.023077305406332016,
  0.002991920802742243,
  -0.004572558216750622,
  -0.004884922411292791,
  0.010319303721189499,
  -0.012908538803458214,
  0.025485895574092865,
  0.048412665724754333,
  -0.00020981079433113337,
  0.003005092730745673,
  -0.01975044049322605,
  -0.01633325219154358,
  0.013563374057412148,
  0.00816662609577179,
  0.012833270244300365,
  0.012374132871627808,
  -0.04196968674659729,
  0.027472982183098793,
  0.004297828301787376,
  -0.0025403101462870836,
  0.007124158553779125,
  -0.0013661222765222192,
  0.015490246005356312,
  -0.013691330328583717,
  0.029761143028736115,
  0.005295135080814362,
  -0.01607733964920044,
  -0.02405579574406147,
  0.030950386077165604,
  0.033298760652542114,
  0.025802023708820343,
  0.010123605839908123,
  -0.001065048505552113,
  -0.0225052647292614,
  0.003962883725762367,
  -0.012968753464519978,
  0.032275110483169556,
  -0.021135380491614342,
  0.029309533536434174,
  0.012306391261518002,
  -0.030077271163463593,
  0.0016775454860180616,
  0.01014618668705225,
  0.009732210077345371,
  -0.03627939149737358,
  -0.0010669302428141236,
  -0.010025757364928722,
  0.013044022023677826,
  -0.003405897179618478,
  0.0003151866258122027,
  0.01556551456451416,
  0.04160839691758156,
  0.00934081431478262,
  -0.01123757939785719,
  -0.0008646462811157107,
  -0.021857956424355507,
  -0.026705244556069374,
  0.02955039218068123,
  0.029354693368077278,
  0.03495466709136963,
  0.04495031759142876,
  0.00892683770507574,
  0.012185961939394474,
  -0.021602043882012367,
  -0.002209128811955452,
  -0.005569864995777607,
  0.014127887785434723,
  0.006811794359236956,
  0.012614991515874863,
  0.01649884320795536,
  -0.02955039218068123,
  0.0016060405177995563,
  0.057836275547742844,
  0.03486434370279312,
  -0.03597831726074219,
  0.0028338569682091475,
  -0.006864482536911964,
  -0.01565583609044552,
  0.011019300669431686,
  0.03721271827816963,
  -0.01341283693909645,
  -0.005983841605484486,
  -0.0193590447306633,
  0.023242896422743797,
  -0.011681662872433662,
  -0.014444014988839626,
  -0.016227876767516136,
  0.02447729930281639,
  0.03251596912741661,
  0.019735386595129967,
  0.0039704106748104095,
  0.013984877616167068,
  0.026283740997314453,
  -0.041217003017663956,
  0.016197768971323967,
  -0.011809619143605232,
  -0.001350127742625773,
  0.01320208515971899,
  -0.008761247619986534,
  0.01604723371565342,
  -0.001405638293363154,
  -0.0006858837441541255,
  0.009197804145514965,
  0.017552601173520088,
  0.007673617918044329,
  -0.015700997784733772,
  -0.016137555241584778,
  0.0153547627851367,
  0.0018356092041358352,
  -0.036580465734004974,
  0.010936505161225796,
  -0.023498808965086937,
  -0.0268256738781929,
  0.016875185072422028,
  0.055487897247076035,
  -0.00040480311145074666,
  0.006059110164642334,
  -0.011508545838296413,
  -0.025877291336655617,
  0.021526776254177094,
  -0.005140834953635931,
  0.011327900923788548,
  0.019946137443184853,
  0.01193004846572876,
  -0.006668784189969301,
  -0.005441908724606037,
  0.019057970494031906,
  -0.025576218962669373,
  -0.013548320159316063,
  0.0009258018690161407,
  -0.0014169284841045737,
  -0.025034286081790924,
  -0.0051596518605947495,
  0.0036862720735371113,
  0.019313883036375046,
  0.009980595670640469,
  0.009152643382549286,
  0.031401995569467545,
  0.009175223298370838,
  0.061087872833013535,
  0.03317832946777344,
  -0.02190311811864376,
  -0.0027585886418819427,
  -0.010236509144306183,
  -0.01604723371565342,
  -0.00669889198616147,
  0.04636536166071892,
  -0.040855713188648224,
  0.018952595070004463,
  0.011199944652616978,
  -0.005404274445027113,
  0.05723412707448006,
  -0.008475227281451225,
  0.01678486354649067,
  -0.010131132788956165,
  -0.037965405732393265,
  0.003268532222136855,
  -0.0022938058245927095,
  0.004523633513599634,
  -0.018380554392933846,
  -0.030378345400094986,
  -0.01678486354649067,
  0.004275247920304537,
  -0.011087042279541492,
  -0.007752649951726198,
  -0.004211269784718752,
  0.010793494991958141,
  -0.00908490177243948,
  -0.03600842505693436,
  0.03045361302793026,
  -0.01621282286942005,
  0.026569761335849762,
  -0.027668680995702744,
  -0.03886862471699715,
  -0.003336273832246661,
  -0.03718261420726776,
  0.021722473204135895,
  0.003146221162751317,
  -0.027834272012114525,
  -0.005720402114093304,
  -0.011832199990749359,
  0.005983841605484486,
  -0.030769741162657738,
  -0.04040410369634628,
  -0.006672547664493322,
  -0.016152609139680862,
  0.007880606688559055,
  0.010138659738004208,
  0.00602147588506341,
  -0.0008557081455364823,
  -0.004617718979716301,
  -0.0003565842634998262,
  0.003131167497485876,
  0.029128888621926308,
  -0.023017091676592827,
  -0.02051817812025547,
  -0.004519870039075613,
  0.016995614394545555,
  0.057896487414836884,
  -0.017793461680412292,
  0.00908490177243948,
  -0.008821462281048298,
  -0.0055247037671506405,
  -0.0026833200827240944,
  0.003744605230167508,
  -0.028210613876581192,
  0.013292407616972923,
  0.0284966342151165,
  -0.004169872030615807,
  -0.02949017658829689,
  0.018531091511249542,
  0.026554707437753677,
  0.026765460148453712,
  -0.030498774722218513,
  -0.007191900163888931,
  -0.013563374057412148,
  0.00215455936267972,
  -0.009702103212475777,
  0.007094051223248243,
  -0.002948641311377287,
  0.02659986913204193,
  0.03697185963392258,
  0.01785367541015148,
  0.018290232867002487,
  0.017251528799533844,
  -0.023498808965086937,
  0.011041881516575813,
  -0.0037144978996366262,
  -0.010891344398260117,
  -0.011071988381445408,
  -0.0019682699348777533,
  -0.004561267793178558,
  -0.012471981346607208,
  0.010033284313976765,
  0.017868729308247566,
  0.0004589023010339588,
  -0.002427407307550311,
  0.010552636347711086,
  0.01823001727461815,
  0.028029968962073326,
  -0.004422021098434925,
  -0.01159134041517973,
  0.01949452795088291,
  0.00905479397624731,
  -0.023273004218935966,
  0.010605324059724808,
  0.008964472450315952,
  -0.018335392698645592,
  -0.026765460148453712,
  0.02968587540090084,
  -0.024176225066184998,
  -0.009942961856722832,
  0.011041881516575813,
  -0.023559024557471275,
  0.015347235836088657,
  -0.03877830505371094,
  -0.002783050760626793,
  -0.007933294400572777,
  0.014654766768217087,
  0.008670925162732601,
  0.005855885334312916,
  -0.025350412353873253,
  0.0009032213711179793,
  0.0005188818322494626,
  -0.03862776607275009,
  0.043625593185424805,
  0.0046967510133981705,
  -0.008580602705478668,
  -0.019148292019963264,
  -0.03314822539687157,
  0.009370921179652214,
  0.010853709653019905,
  -0.024898802861571312,
  0.0019626247230917215,
  -0.022264406085014343,
  -0.0027661153580993414,
  -0.013924662955105305,
  -0.007398888468742371,
  -0.03161274641752243,
  -0.04022345691919327,
  0.00045796146150678396,
  -0.023754721507430077,
  -0.02312246710062027,
  0.012374132871627808,
  -0.006074163597077131,
  0.008512861095368862,
  0.015911750495433807,
  -0.008211787790060043,
  -0.00674405274912715,
  0.02563643269240856,
  0.02643427811563015,
  0.020111728459596634,
  -0.012155854143202305,
  -0.002101871417835355,
  -0.016845079138875008,
  -0.0072370609268546104,
  -0.04826212674379349,
  -0.02106011100113392,
  -0.03676110878586769,
  -0.02575686201453209,
  0.007639747112989426,
  -0.019885923713445663,
  0.030829954892396927,
  0.007730069570243359,
  0.0017575182719156146,
  -0.05578897148370743,
  0.01618271693587303,
  0.01205800473690033,
  -0.0022542900405824184,
  0.013616061769425869,
  -0.006401581224054098,
  -0.01323219295591116,
  0.003117995336651802,
  -0.0023991817142814398,
  0.017251528799533844,
  0.0024048269260674715,
  -0.01061285100877285,
  -0.0004864223301410675,
  0.02611815184354782,
  0.041728828102350235,
  -0.011952629312872887,
  -0.008791354484856129,
  0.03483423590660095,
  0.001208999427035451,
  -0.02498912438750267,
  0.011470911093056202,
  -0.027217069640755653,
  -0.005106964148581028,
  0.0017688084626570344,
  0.024688050150871277,
  0.03384069353342056,
  -0.03525574132800102,
  0.012667679227888584,
  -0.011132203042507172,
  -0.01836550049483776,
  0.01556551456451416,
  0.04955674707889557,
  -0.003225252963602543,
  0.010070918127894402,
  -0.012088112533092499,
  -0.004403204191476107,
  -0.03384069353342056,
  0.005637606605887413,
  0.00446341885253787,
  0.009137589484453201,
  0.01627303846180439,
  -0.033298760652542114,
  0.015866588801145554,
  -0.009641887620091438,
  -0.03694175183773041,
  0.002092462731525302,
  -0.0279095396399498,
  -0.0010612851474434137,
  -0.022625695914030075,
  0.021045057103037834,
  0.015294548124074936,
  0.012833270244300365,
  0.003575251204892993,
  -0.0012842678697779775,
  -0.01762787066400051,
  -0.027262231335043907,
  0.01737195812165737,
  -0.001853485475294292,
  0.019855815917253494,
  0.007440285757184029,
  -0.034232091158628464,
  0.020111728459596634,
  0.01107951533049345,
  -0.03236543387174606,
  0.02337837964296341,
  -0.02988157421350479,
  -0.013593481853604317,
  0.0073311468586325645,
  0.015460139140486717,
  0.012201014906167984,
  0.022264406085014343,
  0.024386975914239883,
  -0.019283775240182877,
  0.02515471540391445,
  -0.01352574024349451,
  -0.01658916473388672,
  0.01669454202055931,
  -0.013939715921878815,
  0.017191313207149506,
  -0.023197734728455544,
  -0.020247211679816246,
  0.010710700415074825,
  -0.0069999657571315765,
  0.008685979060828686,
  -0.012359078973531723,
  -0.013352622278034687,
  0.017311742529273033,
  0.0029467595741152763,
  -0.008896730840206146,
  0.006570935249328613,
  -0.015369816683232784,
  -0.022806338965892792,
  0.009694576263427734,
  -0.025064392015337944,
  -0.010710700415074825,
  -0.004851051606237888,
  -0.019690224900841713,
  -0.014985947869718075,
  0.013578427955508232,
  0.00638276431709528,
  0.009815005585551262,
  0.007319856435060501,
  0.014639712870121002,
  -0.006341366562992334,
  -0.0024085904005914927,
  0.004083313047885895,
  0.019238615408539772,
  -0.0033400373067706823,
  -0.026389118283987045,
  0.0052838451229035854,
  0.03549659997224808,
  0.010793494991958141,
  -0.029640713706612587,
  -0.0019127593841403723,
  -0.010304250754415989,
  0.03248586133122444,
  0.024266546592116356,
  -0.0032948763109743595,
  -0.0320342518389225,
  -0.0012456928379833698,
  -0.0027360080275684595,
  0.029113834723830223,
  -0.020563339814543724,
  -0.002756706904619932,
  0.007820392027497292,
  0.0049940613098442554,
  -0.01746227964758873,
  -0.001295558176934719,
  0.03998259827494621,
  -0.007508027367293835,
  0.0011901822872459888,
  -0.008527914993464947,
  -0.006160722579807043,
  -0.02357407659292221,
  0.029866520315408707,
  -0.01318703219294548,
  0.02733749896287918,
  -0.022264406085014343,
  -0.010838656686246395,
  0.010733280330896378,
  -0.005201049614697695,
  0.005223629996180534,
  -0.012163381092250347,
  0.012118220329284668,
  0.0503094308078289,
  -0.0021827849559485912,
  -0.007918240502476692,
  -0.004060732666403055,
  -0.03387080132961273,
  -0.006815557833760977,
  -0.018952595070004463,
  0.00887414999306202,
  0.00921285804361105,
  -0.009039740078151226,
  0.007647274062037468,
  -0.0020115491934120655,
  -0.016890238970518112,
  0.0117192966863513,
  0.024462245404720306,
  -0.015971964225172997,
  0.004038152284920216,
  0.0026287504006177187,
  0.021300969645380974,
  -0.02187301032245159,
  0.03739336505532265,
  0.0051859961822628975,
  0.011613921262323856,
  0.03320843726396561,
  -0.0057166386395692825,
  -0.008241894654929638,
  -0.025877291336655617,
  0.015144011937081814,
  0.038025617599487305,
  0.010213928297162056,
  0.016830025240778923,
  -0.008053723722696304,
  0.014963367022573948,
  -0.023363325744867325,
  -0.0004229145706631243,
  -0.019961191341280937,
  0.010793494991958141,
  0.01762787066400051,
  -0.004109657369554043,
  0.009920381009578705,
  -0.02742782235145569,
  -0.016875185072422028,
  0.010454786941409111,
  -0.002252408303320408,
  0.002997565781697631,
  -0.008768773637712002,
  0.011621448211371899,
  -0.02578696981072426,
  ...],
 [0.03536432236433029,
  -0.00970013253390789,
  -0.013207930140197277,
  0.002378053730353713,
  0.011847763322293758,
  -0.027650747448205948,
  0.005002190358936787,
  0.08876874297857285,
  -0.011981991119682789,
  0.05977572500705719,
  -0.007485388312488794,
  0.0031655183993279934,
  0.001218109391629696,
  0.009190070442855358,
  -0.0018858883995562792,
  0.011221371591091156,
  -0.026183199137449265,
  0.003100642003118992,
  -0.006420521531254053,
  0.009279554709792137,
  0.04173562675714493,
  -0.0016755994874984026,
  -0.01941816322505474,
  0.04184300824999809,
  -0.0216731745749712,
  0.004930602386593819,
  -0.017234737053513527,
  -0.007127449847757816,
  -0.017762696370482445,
  0.01479180809110403,
  -0.0004247722972650081,
  0.0016219087410718203,
  0.010890278033912182,
  -0.011981991119682789,
  -0.04470651596784592,
  -0.006384727545082569,
  0.03711821883916855,
  -0.020849917083978653,
  -0.043990638107061386,
  0.004371323622763157,
  0.006335511337965727,
  0.013234775513410568,
  -0.010952917858958244,
  0.010684464126825333,
  0.047426849603652954,
  0.04309579357504845,
  0.004362375009804964,
  -0.006930583622306585,
  0.010451803915202618,
  -0.008773966692388058,
  0.0035234568640589714,
  0.02328389883041382,
  0.003921663388609886,
  0.0672566369175911,
  0.0009194544982165098,
  0.001134217600338161,
  -0.006648707203567028,
  -0.02997734770178795,
  0.01687679998576641,
  -0.0010945087997242808,
  0.00033780443482100964,
  -0.029154090210795403,
  -0.025699982419610023,
  0.024339817464351654,
  -0.04016069695353508,
  -0.021064680069684982,
  0.025932643562555313,
  0.017136305570602417,
  0.011257165111601353,
  0.016849953681230545,
  0.0011554701486602426,
  0.015418199822306633,
  0.03616968169808388,
  0.025485219433903694,
  -0.007641986478120089,
  0.017485294491052628,
  0.04735526069998741,
  0.02535994164645672,
  -0.011937248520553112,
  0.0461740642786026,
  0.0200982466340065,
  0.018057996407151222,
  0.003767302492633462,
  0.03593702241778374,
  -0.009431678801774979,
  0.001891481108032167,
  -0.03761933371424675,
  -0.09027208387851715,
  0.02006245218217373,
  0.029207780957221985,
  0.020849917083978653,
  -0.003313167952001095,
  -0.0368676632642746,
  2.0850615328527056e-05,
  0.033807288855314255,
  0.024321919307112694,
  0.013324259780347347,
  -0.012152011506259441,
  -0.030174214392900467,
  0.027507571503520012,
  -0.024089260026812553,
  -0.016849953681230545,
  4.774843546329066e-05,
  -0.008599472232162952,
  0.0350421778857708,
  0.04037545993924141,
  0.03609809651970863,
  0.011355598457157612,
  0.011436134576797485,
  -0.017512140795588493,
  -0.001149877323769033,
  -0.009512214921414852,
  -0.02519886940717697,
  -0.012304135598242283,
  -0.0063713048584759235,
  -0.0368676632642746,
  -0.04155665636062622,
  0.028080273419618607,
  -0.009413782507181168,
  -0.015239231288433075,
  -0.006586067844182253,
  0.007257202640175819,
  0.007561450358480215,
  -0.0216731745749712,
  0.007230357266962528,
  -0.023838702589273453,
  -0.005758335348218679,
  0.006863470189273357,
  0.008201265707612038,
  -0.0036196529399603605,
  -0.019579235464334488,
  -0.004449622705578804,
  -0.008049141615629196,
  -0.014478611759841442,
  0.0032997452653944492,
  0.037762511521577835,
  0.03049635887145996,
  0.012178856879472733,
  0.016867851838469505,
  0.008800812065601349,
  -0.025234663859009743,
  0.0014843260869383812,
  0.0013109496794641018,
  0.007897017523646355,
  -0.04757002368569374,
  0.008635265752673149,
  0.005592788569629192,
  0.019901379942893982,
  0.0070648109540343285,
  -0.017646366730332375,
  0.0198476891964674,
  -0.027650747448205948,
  -0.04313158616423607,
  0.02401767298579216,
  -0.02052777260541916,
  -0.0008176657138392329,
  0.029225677251815796,
  0.0017360015772283077,
  -0.019042328000068665,
  -0.046997323632240295,
  0.013127394020557404,
  0.047212086617946625,
  -0.00749881099909544,
  -0.03115854412317276,
  -0.024876724928617477,
  -3.5059794754488394e-05,
  0.02698856219649315,
  0.026612726971507072,
  -0.0210109893232584,
  0.03815624117851257,
  -0.016536757349967957,
  -0.014335435815155506,
  0.05029035732150078,
  -0.001023480319418013,
  0.0373687781393528,
  -0.008362337946891785,
  0.0030872193165123463,
  -0.044241197407245636,
  0.021404720842838287,
  0.013064754195511341,
  0.023659734055399895,
  -0.02235325798392296,
  0.0058522941544651985,
  -0.015785086899995804,
  -0.0325545035302639,
  -0.0024921465665102005,
  0.019453955814242363,
  -0.003386992961168289,
  -0.01685890182852745,
  -0.028921429067850113,
  0.0010520035866647959,
  0.002796394284814596,
  0.028295036405324936,
  0.009565905667841434,
  0.0025704456493258476,
  -0.007172192446887493,
  0.004469756968319416,
  0.010532340034842491,
  -0.016805211082100868,
  -0.0137448376044631,
  -0.03531062975525856,
  -0.01821906864643097,
  -0.014245951548218727,
  0.0070871817879378796,
  -0.03026369772851467,
  -0.02147630974650383,
  0.007932811975479126,
  0.013118444941937923,
  0.040268078446388245,
  0.0015917076962068677,
  0.05107782036066055,
  -0.01763741858303547,
  -0.011024504899978638,
  -0.0016577026108279824,
  0.018881255760788918,
  0.049252334982156754,
  -0.02680959179997444,
  0.04610247537493706,
  0.00832206942141056,
  0.012429413385689259,
  0.015355560928583145,
  0.002353445626795292,
  0.011445082724094391,
  0.02099309116601944,
  0.012939476408064365,
  -0.017932718619704247,
  0.0233375895768404,
  0.06256764382123947,
  -0.02144051529467106,
  -0.006791882682591677,
  -0.0006403743172995746,
  0.01757477968931198,
  -0.0242503322660923,
  0.0228364747017622,
  -0.02772233448922634,
  0.007431697566062212,
  0.015802983194589615,
  -0.04377587512135506,
  -0.011185577139258385,
  0.019471853971481323,
  -0.01639358326792717,
  -0.007073759101331234,
  -0.04449175298213959,
  -0.033127207309007645,
  0.0166083462536335,
  -0.015883520245552063,
  -0.021064680069684982,
  0.020420391112565994,
  0.00968223623931408,
  -0.03629496321082115,
  0.009861204773187637,
  0.01831750199198723,
  -0.010693412274122238,
  0.006505531724542379,
  0.0006744903512299061,
  0.027543365955352783,
  0.02194162830710411,
  0.06167279928922653,
  -0.03414732962846756,
  -0.00037331864587031305,
  0.029458336532115936,
  -0.028223449364304543,
  -0.036796074360609055,
  0.027418088167905807,
  0.020169833675026894,
  0.005628582555800676,
  0.01325267180800438,
  -0.027060149237513542,
  -0.01466652937233448,
  -0.00887687411159277,
  0.021136267110705376,
  -0.01386116724461317,
  0.025037797167897224,
  0.029368853196501732,
  0.05336862802505493,
  -0.03808465600013733,
  -0.01395960059016943,
  -0.005597263108938932,
  0.013950652442872524,
  -0.021189957857131958,
  0.022299567237496376,
  0.05662586912512779,
  0.008398131467401981,
  -0.0233375895768404,
  -0.031230133026838303,
  -0.010156504809856415,
  0.010433906689286232,
  -0.011686691083014011,
  -0.010550236329436302,
  -0.058737706393003464,
  -0.005968624260276556,
  -0.016151973977684975,
  -0.009762772358953953,
  -0.015668757259845734,
  -0.00675161462277174,
  0.016017746180295944,
  0.06936847418546677,
  -0.030192110687494278,
  -0.004805324133485556,
  0.013610610738396645,
  -0.010317577049136162,
  -0.011212422512471676,
  0.005230376031249762,
  -0.06535957008600235,
  0.004004436545073986,
  0.025252560153603554,
  0.05372656509280205,
  -0.004930602386593819,
  -0.0636056661605835,
  0.0002326600078959018,
  0.018066944554448128,
  -0.04409801959991455,
  -0.015856673941016197,
  -0.06027683988213539,
  -0.019579235464334488,
  -0.02149420604109764,
  -0.014899189583957195,
  -0.035185351967811584,
  -0.03024580143392086,
  0.014684426598250866,
  -0.006841099355369806,
  0.015749292448163033,
  -0.01687679998576641,
  0.013315311633050442,
  0.030639534816145897,
  0.020474081858992577,
  -0.024125054478645325,
  0.0004669978516176343,
  0.0008210213854908943,
  -0.011946196667850018,
  -0.05494355782866478,
  -0.030371079221367836,
  0.024142950773239136,
  0.0037561168428510427,
  -0.038657356053590775,
  -0.047212086617946625,
  0.002036893740296364,
  -0.014308590441942215,
  0.007534604985266924,
  0.008845554664731026,
  0.036366548389196396,
  0.04341793805360794,
  0.023946084082126617,
  0.03923005610704422,
  0.00982541125267744,
  0.003129724645987153,
  0.016733624041080475,
  -0.035650674253702164,
  0.008818709291517735,
  0.01558822114020586,
  0.00030480697751045227,
  -0.017395809292793274,
  0.029225677251815796,
  0.003505560103803873,
  0.020366700366139412,
  -0.01040706131607294,
  0.0027673118747770786,
  -0.038943707942962646,
  0.030388977378606796,
  -0.018344346433877945,
  -0.0006040211883373559,
  0.020187729969620705,
  0.013852219097316265,
  0.012894733808934689,
  0.04248729720711708,
  -0.027346499264240265,
  0.03049635887145996,
  -0.0230870321393013,
  0.019668718799948692,
  -0.009995431639254093,
  -0.005301963537931442,
  0.021189957857131958,
  -0.013547971844673157,
  0.02770443819463253,
  -0.026594828814268112,
  0.003230394795536995,
  0.026612726971507072,
  -0.00037443722249008715,
  -0.0019272749777883291,
  0.00924376118928194,
  -0.017941666767001152,
  0.009234813041985035,
  -0.03364621847867966,
  -0.0348811037838459,
  -0.00979856587946415,
  -0.0067963567562401295,
  0.0225859172642231,
  -0.018013253808021545,
  0.0008199028088711202,
  0.010577081702649593,
  -0.007252728566527367,
  -0.020223524421453476,
  0.01164194941520691,
  0.025807363912463188,
  0.04893019050359726,
  0.021351030096411705,
  0.030192110687494278,
  -0.0061744386330246925,
  0.016894696280360222,
  -0.009852256625890732,
  -0.020402492955327034,
  -0.038872119039297104,
  -0.04943130537867546,
  -0.001763965585269034,
  0.029028810560703278,
  0.01282314583659172,
  0.0048142727464437485,
  -0.01086343266069889,
  0.01870228536427021,
  0.017476346343755722,
  -0.03248291835188866,
  -0.03622337430715561,
  -0.013494281098246574,
  -0.032447122037410736,
  0.03226815536618233,
  0.021816350519657135,
  0.03597281873226166,
  0.015427148900926113,
  -0.018845461308956146,
  -0.04706890881061554,
  -0.012375722639262676,
  0.021995319053530693,
  -0.008053615689277649,
  -0.027758128941059113,
  0.0015190014382824302,
  0.00566437654197216,
  0.008505512960255146,
  -0.0032907968852669,
  0.0031767040491104126,
  0.0033623846247792244,
  -0.00641157291829586,
  0.00994174089282751,
  0.0013545734109357,
  0.07012014836072922,
  -0.004288550466299057,
  -0.010791845619678497,
  0.0198476891964674,
  0.004572663921862841,
  -0.006107325199991465,
  -0.0017002078238874674,
  0.028724562376737595,
  -0.0043690865859389305,
  -0.011409289203584194,
  0.009610648266971111,
  -0.04771319776773453,
  0.01175827905535698,
  -0.022496433928608894,
  -0.021834246814250946,
  0.002583868335932493,
  0.03740457072854042,
  0.005167736671864986,
  -0.04943130537867546,
  -0.03226815536618233,
  -0.028312934562563896,
  -0.0014854447217658162,
  0.02281857840716839,
  0.007852275855839252,
  0.021351030096411705,
  -0.013145290315151215,
  -0.0011280655162408948,
  0.013655353337526321,
  0.025091487914323807,
  0.0110871447250247,
  -0.014228054322302341,
  0.0025883426424115896,
  -0.005561469122767448,
  -0.05039773881435394,
  -0.0006062582833692431,
  -0.03829941898584366,
  -0.007221408654004335,
  -0.006930583622306585,
  0.017709005624055862,
  -0.019310781732201576,
  0.0006549155805259943,
  -0.007995450869202614,
  -0.004431725945323706,
  -0.01815642975270748,
  -0.031731247901916504,
  -0.01775374822318554,
  -0.028921429067850113,
  -0.02054566890001297,
  -0.022496433928608894,
  0.004966396372765303,
  0.010890278033912182,
  0.028903532773256302,
  0.05029035732150078,
  0.03690345585346222,
  0.0007114027393981814,
  0.025449426844716072,
  0.046961527317762375,
  0.021386824548244476,
  0.0004997156793251634,
  -0.013950652442872524,
  0.01961502805352211,
  -0.019704513251781464,
  -0.03815624117851257,
  -0.03847838565707207,
  -0.06120748072862625,
  0.0043534268625080585,
  0.014004343189299107,
  0.016679933294653893,
  -0.017485294491052628,
  0.0024093734100461006,
  0.028348727151751518,
  0.00524827279150486,
  -0.006062583066523075,
  0.011624052189290524,
  0.005695695988833904,
  -0.005574891809374094,
  0.009140853770077229,
  0.01387011632323265,
  -0.021977422758936882,
  -0.004149849060922861,
  -0.006187861319631338,
  -0.015212385915219784,
  -0.025252560153603554,
  0.01130190771073103,
  4.180609539616853e-05,
  0.011212422512471676,
  -0.019060224294662476,
  -0.008416028693318367,
  0.025037797167897224,
  -0.01038021594285965,
  0.02976258471608162,
  0.005476458929479122,
  0.0034004156477749348,
  0.01913181133568287,
  -0.021834246814250946,
  0.07226777821779251,
  0.032178670167922974,
  0.00829522404819727,
  0.028223449364304543,
  -0.0007074877503328025,
  0.0068903155624866486,
  -0.008867925964295864,
  -0.024823034182190895,
  -0.022281670942902565,
  0.024196641519665718,
  -0.02263960801064968,
  0.0017896924400702119,
  0.006550274323672056,
  0.002814291277900338,
  -0.0012158722383901477,
  -0.015865623950958252,
  -0.011964093893766403,
  0.040518637746572495,
  0.0008993204100988805,
  -0.02979837916791439,
  -0.02102888561785221,
  -0.015096055343747139,
  0.02031300961971283,
  0.04406222701072693,
  0.021995319053530693,
  -0.012411517091095448,
  0.038013067096471786,
  -0.0019104966195300221,
  -0.008765018545091152,
  0.009087163023650646,
  -0.0042035398073494434,
  -0.006966377608478069,
  0.036348652094602585,
  0.037046633660793304,
  -0.010272834450006485,
  -0.0060849543660879135,
  0.006089428439736366,
  -0.0005584399332292378,
  -0.006129696499556303,
  0.021637381985783577,
  -0.021565793082118034,
  -0.023409176617860794,
  0.007261676713824272,
  0.003297508228570223,
  -0.035203248262405396,
  -0.008581575006246567,
  -0.0004465841921046376,
  0.0013500992208719254,
  0.033789392560720444,
  0.021780556067824364,
  -0.008290749974548817,
  0.022317463532090187,
  -0.008943988010287285,
  -0.017941666767001152,
  0.019722409546375275,
  0.00767330639064312,
  -0.03665290027856827,
  -0.014979725703597069,
  0.012304135598242283,
  -0.04255888611078262,
  -0.051364172250032425,
  0.02011614292860031,
  0.02999524399638176,
  0.0149081377312541,
  0.013977497816085815,
  0.001321016694419086,
  0.014684426598250866,
  0.03765513002872467,
  -0.011713536456227303,
  0.014639683999121189,
  -0.016867851838469505,
  -0.003152095712721348,
  -0.03531062975525856,
  0.01279630046337843,
  0.009807514026761055,
  -0.03233974054455757,
  -0.01533766370266676,
  0.024590373039245605,
  0.022782783955335617,
  0.009351142682135105,
  -0.015883520245552063,
  -0.002131971064954996,
  0.00024762071552686393,
  -0.004890334326773882,
  0.02378501184284687,
  0.01558822114020586,
  -0.03416522964835167,
  0.003648735349997878,
  0.007700151763856411,
  0.0013266095193102956,
  -0.020026657730340958,
  0.0019194451160728931,
  0.030335286632180214,
  -0.03815624117851257,
  0.016134077683091164,
  0.043990638107061386,
  -0.03369990736246109,
  0.0032684258185327053,
  0.016232511028647423,
  -0.016089335083961487,
  0.01988348178565502,
  0.009431678801774979,
  -0.025073589757084846,
  -0.033377762883901596,
  -0.007659883704036474,
  0.005825448781251907,
  0.0030827452428638935,
  -0.014988673850893974,
  0.045136041939258575,
  -0.03301982581615448,
  -0.014836549758911133,
  0.013181084766983986,
  0.0021174298599362373,
  0.003999962471425533,
  0.039158470928668976,
  -0.02836662530899048,
  -0.027095943689346313,
  0.01117662899196148,
  -0.022424845024943352,
  0.02859928458929062,
  0.0017236975254490972,
  0.037941478192806244,
  0.015668757259845734,
  -0.015552426688373089,
  0.016304098069667816,
  0.010487597435712814,
  0.01592826284468174,
  -0.009628545492887497,
  0.018031150102615356,
  -0.005436190869659185,
  0.00737800681963563,
  -0.005033509805798531,
  -0.012617331929504871,
  -0.0242503322660923,
  0.025699982419610023,
  -0.02170896902680397,
  0.017619522288441658,
  0.009136379696428776,
  -0.003230394795536995,
  -0.027113839983940125,
  -0.02305123768746853,
  0.014478611759841442,
  -0.004756107460707426,
  -0.01965082250535488,
  0.02147630974650383,
  -0.006134170573204756,
  -0.005946252960711718,
  -0.025968436151742935,
  -0.011042402125895023,
  0.00970908161252737,
  0.009977535344660282,
  -0.015534530393779278,
  -0.021565793082118034,
  0.018845461308956146,
  0.02382080629467964,
  0.026380065828561783,
  -0.011552464216947556,
  -0.027346499264240265,
  -0.012715764343738556,
  0.02426822856068611,
  -0.0027583634946495295,
  0.012697868049144745,
  -0.0019228007877245545,
  0.015883520245552063,
  -0.018755976110696793,
  0.018147481605410576,
  0.0040983958169817924,
  0.022729093208909035,
  0.02514517866075039,
  0.019937172532081604,
  -0.006841099355369806,
  -0.009288503788411617,
  -0.02704225294291973,
  -0.01086343266069889,
  -0.016330942511558533,
  -0.01627725176513195,
  0.014648632146418095,
  0.024787239730358124,
  0.00788806937634945,
  0.011149783618748188,
  0.003961931448429823,
  -0.035203248262405396,
  0.0189349465072155,
  0.021601587533950806,
  -0.005480933003127575,
  0.01475601363927126,
  0.009807514026761055,
  -0.013530074618756771,
  -0.012608382850885391,
  -0.016429375857114792,
  0.006487634964287281,
  0.011346650309860706,
  0.004653200041502714,
  0.005601737182587385,
  0.010326525196433067,
  -0.006219181232154369,
  0.008277327753603458,
  0.00017896924691740423,
  0.01941816322505474,
  0.019525544717907906,
  0.018630698323249817,
  -0.0089260907843709,
  -0.023606043308973312,
  0.0060133663937449455,
  -0.005843345541507006,
  0.03330617398023605,
  0.005646479316055775,
  0.051113616675138474,
  0.02240694873034954,
  0.022460639476776123,
  0.011561413295567036,
  -0.010970814153552055,
  0.03393256664276123,
  0.012375722639262676,
  0.006040211766958237,
  0.00832206942141056,
  -0.06765037029981613,
  0.005051407031714916,
  0.011829867027699947,
  0.00890371948480606,
  0.01589246839284897,
  -0.007149821147322655,
  0.04817852005362511,
  -0.027078045532107353,
  -0.001202449551783502,
  -0.04370428994297981,
  0.014621786773204803,
  0.02260381542146206,
  0.01649201475083828,
  -0.011382443830370903,
  -0.01848752237856388,
  0.03024580143392086,
  0.005239324644207954,
  0.03534642606973648,
  -0.01013860758394003,
  -0.020724637433886528,
  -0.010469700209796429,
  -0.03207128867506981,
  -0.024304023012518883,
  -0.016402531415224075,
  0.02235325798392296,
  0.015042364597320557,
  0.01765531487762928,
  0.0210109893232584,
  0.0013075940078124404,
  0.006008892320096493,
  -0.005861242767423391,
  0.008568152785301208,
  -0.026916973292827606,
  -0.03697504475712776,
  -0.03364621847867966,
  -0.0027404665015637875,
  -0.023480763658881187,
  -0.020008761435747147,
  0.007973079569637775,
  0.0016476355958729982,
  0.010514442808926105,
  -0.021565793082118034,
  0.0034227867145091295,
  0.0020234710536897182,
  0.025019899010658264,
  0.0043534268625080585,
  0.005145365837961435,
  0.020724637433886528,
  -0.011498773470520973,
  0.008881348185241222,
  0.011525618843734264,
  0.00028173672035336494,
  -0.01454125065356493,
  0.0037046633660793304,
  0.007078233640640974,
  -0.02859928458929062,
  0.031051162630319595,
  -0.030567945912480354,
  0.019274987280368805,
  0.049323923885822296,
  -0.0163577888160944,
  -0.022711196914315224,
  -0.012572589330375195,
  -0.018415935337543488,
  -0.024661961942911148,
  -0.009512214921414852,
  -0.012697868049144745,
  -0.006756088696420193,
  -0.008845554664731026,
  0.022925959900021553,
  -0.008008873090147972,
  0.025896849110722542,
  0.008084935136139393,
  -0.008326543495059013,
  0.004067075904458761,
  0.01606248877942562,
  0.01326162088662386,
  0.016518861055374146,
  0.04116292670369148,
  -0.010460752062499523,
  0.0006001062574796379,
  0.04059022292494774,
  0.009753823280334473,
  -0.02489462122321129,
  0.014550198800861835,
  0.021064680069684982,
  0.02700645849108696,
  -0.007324316073209047,
  0.019328678026795387,
  0.017297377809882164,
  0.034952692687511444,
  0.004442911595106125,
  -0.036348652094602585,
  -0.006259449291974306,
  0.023892393335700035,
  0.03140910342335701,
  0.011561413295567036,
  -0.016402531415224075,
  0.032429225742816925,
  -0.005409345496445894,
  0.000946299871429801,
  -0.0207604318857193,
  -0.007565924432128668,
  -0.006268397439271212,
  0.0189349465072155,
  0.00582992285490036,
  0.0368676632642746,
  0.05408450588583946,
  0.009610648266971111,
  0.03448737412691116,
  -0.035901229828596115,
  0.015606117434799671,
  -0.02195952646434307,
  0.009190070442855358,
  0.025306250900030136,
  0.008867925964295864,
  0.011266113258898258,
  -0.010970814153552055,
  -0.019239192828536034,
  -0.05122099816799164,
  0.006505531724542379,
  -0.00779411057010293,
  -0.02494831196963787,
  -0.03973117098212242,
  -0.016670985147356987,
  -0.0016330942744389176,
  -0.043274763971567154,
  -0.0077090999111533165,
  0.004201302770525217,
  0.009288503788411617,
  0.015060261823236942,
  -0.018129583448171616,
  -0.003393704304471612,
  -0.005109571851789951,
  -0.0005844464176334441,
  -0.0011319804470986128,
  0.0137448376044631,
  -0.005677799228578806,
  -0.036133889108896255,
  0.03001314215362072,
  0.0035077971406280994,
  0.005400396883487701,
  0.007852275855839252,
  0.011042402125895023,
  0.008751596324145794,
  -0.003803096478804946,
  0.004395931959152222,
  0.025538910180330276,
  0.002346734283491969,
  -0.02170896902680397,
  0.007404852192848921,
  0.030388977378606796,
  0.002267316682264209,
  0.02473354898393154,
  -0.025002002716064453,
  -0.007879121229052544,
  -0.023140722885727882,
  0.01846962608397007,
  0.006263923365622759,
  -0.007315367925912142,
  -0.0018892440712079406,
  -0.01396854966878891,
  -0.004653200041502714,
  0.029905760660767555,
  -0.006111799739301205,
  0.014156467281281948,
  0.017592675983905792,
  -0.013091599568724632,
  0.00031151832081377506,
  -0.019006533548235893,
  -0.009377988055348396,
  -0.024304023012518883,
  -0.033127207309007645,
  0.013932755216956139,
  0.011731433682143688,
  -0.01744055189192295,
  -0.029673099517822266,
  0.007306419312953949,
  0.020474081858992577,
  0.04044704884290695,
  -0.012205702252686024,
  -0.03572225943207741,
  -0.05172210931777954,
  0.044241197407245636,
  0.012035681866109371,
  0.03214287385344505,
  -0.01647411845624447,
  0.02033090591430664,
  0.02978048101067543,
  -0.005561469122767448,
  -0.030639534816145897,
  -0.010514442808926105,
  -0.014380178414285183,
  0.011105041019618511,
  -0.016116179525852203,
  -0.010156504809856415,
  -0.007015594281256199,
  0.044455960392951965,
  0.029637306928634644,
  -0.03647392988204956,
  0.005735964048653841,
  -0.028348727151751518,
  -0.013807476498186588,
  0.016420427709817886,
  0.005275118164718151,
  0.020724637433886528,
  -0.009735926985740662,
  -0.02058146335184574,
  -0.019453955814242363,
  -0.005208004731684923,
  0.01190145406872034,
  0.04459913447499275,
  -0.039337437599897385,
  0.04864383861422539,
  -0.0207604318857193,
  0.012196754105389118,
  0.0015883520245552063,
  -0.01028178259730339,
  -0.0110871447250247,
  -0.00812967773526907,
  -0.013521126471459866,
  0.018344346433877945,
  0.023444971069693565,
  0.03158807009458542,
  0.0014809704152867198,
  -0.01512290071696043,
  -0.03550749644637108,
  -0.01821906864643097,
  0.018755976110696793,
  0.010568133555352688,
  0.01685890182852745,
  0.012160959653556347,
  -0.015758242458105087,
  0.01808484084904194,
  0.009530112147331238,
  0.040482841432094574,
  0.012885785661637783,
  -0.0002961381687782705,
  0.006496583577245474,
  -0.04059022292494774,
  0.006219181232154369,
  -0.036599211394786835,
  0.0015066972700878978,
  0.010594978928565979,
  0.03391467034816742,
  0.023928187787532806,
  -0.00010227812163066119,
  -0.022299567237496376,
  0.01742265559732914,
  -0.01889915205538273,
  0.020491978153586388,
  0.014675477519631386,
  0.0007662120624445379,
  -0.0008775085443630815,
  0.014711271971464157,
  -0.014711271971464157,
  ...],
 [0.011899227276444435,
  0.025202829390764236,
  0.00041751674143597484,
  0.011234995909035206,
  0.012857617810368538,
  -0.007671870291233063,
  0.01743132434785366,
  0.06437349319458008,
  -0.0013118566712364554,
  0.03575461730360985,
  -0.03214878961443901,
  0.009076245129108429,
  -0.017032785341143608,
  -0.046116624027490616,
  -0.050633396953344345,
  0.04072686284780502,
  -0.03626702353358269,
  0.004032832104712725,
  -0.021672915667295456,
  -0.006210561841726303,
  0.022223278880119324,
  -0.015097026713192463,
  -0.02237510308623314,
  0.015286806970834732,
  -0.0017697017174214125,
  -0.0008385918918065727,
  -0.008929165080189705,
  -0.0003840086574200541,
  -0.032338570803403854,
  -0.0075200460851192474,
  0.05332827568054199,
  0.00042730229324661195,
  0.006741946563124657,
  0.016463445499539375,
  -0.008340845815837383,
  0.0031906820368021727,
  0.03256630524992943,
  -0.014110168442130089,
  -0.050026100128889084,
  -0.002773165237158537,
  -0.013958344236016273,
  0.0041656787507236,
  -0.01469848770648241,
  0.02810646966099739,
  0.009830622002482414,
  0.017535703256726265,
  0.04759690910577774,
  -0.020249562337994576,
  -0.021008683368563652,
  -0.017032785341143608,
  0.010969303548336029,
  0.0014826589031144977,
  0.002393604489043355,
  0.0018112161196768284,
  -0.0010609907330945134,
  0.04205532371997833,
  0.0005055866786278784,
  -0.020799925550818443,
  0.019907958805561066,
  0.0193291287869215,
  -0.0016937896143645048,
  -0.004329364281147718,
  -0.015220383182168007,
  0.03102908656001091,
  -0.008416757918894291,
  -0.004398159682750702,
  -0.031351711601018906,
  0.024007214233279228,
  -0.002452910877764225,
  0.07655739039182663,
  -0.03550790250301361,
  0.0011191109661012888,
  0.02220430038869381,
  -0.0036461548879742622,
  -0.0016510889399796724,
  0.03558381274342537,
  0.0687384381890297,
  0.027347346767783165,
  0.001211628899909556,
  0.015742279589176178,
  0.019243726506829262,
  -0.01493571326136589,
  0.00538027286529541,
  0.025354653596878052,
  0.004018598701804876,
  -0.00030009017791599035,
  -0.013389003463089466,
  -0.07606396079063416,
  0.00490582175552845,
  0.011348864994943142,
  -0.00925179198384285,
  0.0030910472851246595,
  -0.005313849542289972,
  0.00444323243573308,
  0.015106515027582645,
  0.006186839193105698,
  0.0004059520142618567,
  -0.00847843661904335,
  -0.016501400619745255,
  0.021008683368563652,
  -0.017715994268655777,
  0.003987759351730347,
  -0.016036439687013626,
  -0.021843718364834785,
  0.010058358311653137,
  -0.00634815264493227,
  0.027726907283067703,
  -0.0023342983331531286,
  -0.028599897399544716,
  0.0074109225533902645,
  -0.0070361061953008175,
  -0.011633534915745258,
  -0.0103050721809268,
  0.02190065197646618,
  0.009925511665642262,
  0.007852161303162575,
  -0.022963421419262886,
  0.0046282680705189705,
  0.00974522065371275,
  -0.009878067299723625,
  0.023817433044314384,
  0.029548799619078636,
  0.011481710709631443,
  -0.02573421411216259,
  -0.014745933003723621,
  0.009953979402780533,
  -0.029321063309907913,
  -0.0036295491736382246,
  0.01627366431057453,
  -0.014584619551897049,
  -0.007890117354691029,
  0.008217488415539265,
  -0.0010372682008892298,
  0.009972956962883472,
  0.00672296853736043,
  0.03452104330062866,
  0.045812975615262985,
  0.008900698274374008,
  -0.03740570694208145,
  0.012905063107609749,
  -0.04797647148370743,
  -0.004101627506315708,
  0.015239361673593521,
  0.03334440663456917,
  -0.05750344321131706,
  -0.015808703377842903,
  0.0004018005565740168,
  0.020648101344704628,
  -0.01708023063838482,
  -0.004265313036739826,
  0.02288750931620598,
  -0.04915310814976692,
  0.02379845455288887,
  0.0007893675938248634,
  -0.009840111248195171,
  0.0009204346570186317,
  0.012800684198737144,
  0.007648147642612457,
  0.008791574276983738,
  -0.020667079836130142,
  -0.006998150143772364,
  0.03038383275270462,
  -0.018560517579317093,
  0.021293355152010918,
  -0.023039333522319794,
  0.0031764484010636806,
  0.010978792794048786,
  -0.0011469849850982428,
  -0.005038667935878038,
  0.07325521111488342,
  -0.001358115579932928,
  0.012297766283154488,
  0.022128388285636902,
  0.012686816044151783,
  -0.0007371780229732394,
  0.0273283701390028,
  -0.0004952080780640244,
  -0.0479385145008564,
  -0.004687574692070484,
  -0.02681596204638481,
  -0.01846562698483467,
  -0.0178108848631382,
  -0.01872183009982109,
  -0.0219955425709486,
  0.00925179198384285,
  -0.031389668583869934,
  0.06247568875551224,
  0.006523699499666691,
  -0.00480855954810977,
  -0.049039240926504135,
  -0.022261234000325203,
  -0.03408455103635788,
  0.027650995180010796,
  -0.0074726007878780365,
  0.028599897399544716,
  -0.0034065572544932365,
  -0.004267685580998659,
  -0.0023129479959607124,
  -0.03214878961443901,
  -0.025278741493821144,
  -0.0002827430726028979,
  -0.03958817943930626,
  0.018645917996764183,
  0.008853252977132797,
  0.0034113016445189714,
  0.0054229735396802425,
  0.028808655217289925,
  -0.012070029973983765,
  0.021634958684444427,
  -0.030497701838612556,
  0.020553210750222206,
  0.03622906655073166,
  0.04285240173339844,
  0.012838640250265598,
  0.037842199206352234,
  0.0038691468071192503,
  -0.029662666842341423,
  0.013721118681132793,
  -0.01578972488641739,
  -0.03296484425663948,
  0.025506477802991867,
  -0.011927695013582706,
  -0.01224083174020052,
  0.03467286750674248,
  -0.0437253899872303,
  -0.008535371161997318,
  0.021578025072813034,
  0.02620866522192955,
  0.02220430038869381,
  0.006775158457458019,
  -0.017630593851208687,
  0.006922238040715456,
  -0.05040565878152847,
  0.033667031675577164,
  -0.05135456100106239,
  -0.02005978301167488,
  -0.01927219331264496,
  -0.035659726709127426,
  -0.004203634802252054,
  0.04763486608862877,
  -0.020382408052682877,
  0.01785833016037941,
  -0.04835603013634682,
  -0.022564882412552834,
  -0.0019488069228827953,
  0.003852540859952569,
  0.00871566217392683,
  0.036039285361766815,
  0.0038383074570447206,
  -0.006144138518720865,
  -0.01772548444569111,
  -0.01677658222615719,
  0.008616027422249317,
  0.013104332610964775,
  -0.022697729989886284,
  -0.009228069335222244,
  0.029036393389105797,
  0.04281444475054741,
  -0.05503629893064499,
  -0.006504721473902464,
  0.025620346888899803,
  0.009398871101439,
  -0.04987427592277527,
  0.015590455383062363,
  -0.0017542820423841476,
  -0.05279689282178879,
  0.00979266595095396,
  -0.005660198628902435,
  -0.04573706164956093,
  -0.0070645734667778015,
  0.02848603017628193,
  -0.007700337562710047,
  0.015210894867777824,
  -0.00961711909621954,
  0.007947051897644997,
  -0.019205771386623383,
  -0.0036485271994024515,
  0.005162025336176157,
  0.03343929722905159,
  0.010798501782119274,
  -0.0031574703752994537,
  0.041979413479566574,
  0.03192105516791344,
  0.01028609462082386,
  0.018275847658514977,
  -0.003380462294444442,
  -0.029415953904390335,
  0.0014091189950704575,
  -0.0035892208106815815,
  -0.02522180788218975,
  0.010751056484878063,
  -0.03237652778625488,
  0.004466954618692398,
  0.005427717696875334,
  0.021122552454471588,
  0.03947431221604347,
  0.04118233546614647,
  -0.027613040059804916,
  0.011652513407170773,
  -0.0017839352367445827,
  0.009972956962883472,
  -0.024007214233279228,
  -0.018532050773501396,
  -0.03894292563199997,
  -0.00720690842717886,
  0.014024767093360424,
  0.029453909024596214,
  0.0283152274787426,
  -0.04235897213220596,
  -0.021008683368563652,
  0.04281444475054741,
  -0.015438631176948547,
  -0.05374579504132271,
  -0.05697206035256386,
  -0.036380890756845474,
  -0.03886701539158821,
  0.03283200040459633,
  -0.005313849542289972,
  -0.03321155905723572,
  -0.016843006014823914,
  -0.018902122974395752,
  -0.006367130670696497,
  -0.027726907283067703,
  0.0319400317966938,
  0.00501494575291872,
  0.028334205970168114,
  -0.0026806474197655916,
  -0.012013095431029797,
  -0.004841770976781845,
  -0.015476587228477001,
  -0.05253119766712189,
  -0.023570718243718147,
  0.005247426684945822,
  0.0005387982819229364,
  0.018959056586027145,
  -0.04042321443557739,
  0.010343029163777828,
  0.02569625899195671,
  -0.04501589760184288,
  -0.00453100586310029,
  0.00914266798645258,
  -0.018446648493409157,
  0.012202875688672066,
  -0.005408739671111107,
  -0.005119324661791325,
  0.022432036697864532,
  0.04638231545686722,
  -0.037121035158634186,
  0.007595958188176155,
  0.03490060567855835,
  0.0005328676197677851,
  -0.011870760470628738,
  0.048431944102048874,
  -0.006803625263273716,
  0.034748781472444534,
  -0.017184609547257423,
  0.005702899303287268,
  -0.03288893401622772,
  0.017355412244796753,
  -0.02057218924164772,
  0.005408739671111107,
  0.02087583765387535,
  -0.003828818444162607,
  -0.015675855800509453,
  -0.0018479861319065094,
  0.006628078408539295,
  0.0010627699084579945,
  -0.016833515837788582,
  -0.030402811244130135,
  -0.009294492192566395,
  -0.04892537370324135,
  -0.00549888564273715,
  -0.017839351668953896,
  0.010485364124178886,
  -0.029833469539880753,
  -0.025164874270558357,
  0.01880723237991333,
  0.01023864932358265,
  -0.0029605734162032604,
  0.029795514419674873,
  0.002404279774054885,
  0.04376334697008133,
  -0.0022749919444322586,
  -0.05735161900520325,
  0.03490060567855835,
  0.030630547553300858,
  0.005641220603138208,
  0.0011256346479058266,
  -0.0029985294677317142,
  0.02875172160565853,
  -0.007164207752794027,
  -0.04049912467598915,
  -0.018446648493409157,
  0.033420316874980927,
  0.030877262353897095,
  0.00360345421358943,
  -0.01854153908789158,
  0.013531338423490524,
  0.010466385632753372,
  -0.006016036961227655,
  -0.002247710945084691,
  -0.03459695726633072,
  -0.007648147642612457,
  -0.006537932902574539,
  0.01622621901333332,
  0.014755421318113804,
  -0.05313849449157715,
  -0.0029083837289363146,
  0.016330597922205925,
  0.0229823999106884,
  -0.03296484425663948,
  -0.021843718364834785,
  -0.028467051684856415,
  -0.014641553163528442,
  0.023532763123512268,
  -0.017232054844498634,
  0.040385257452726364,
  0.02431086264550686,
  -0.020686058327555656,
  -0.025278741493821144,
  0.034445133060216904,
  0.01600797288119793,
  -0.027707930654287338,
  -0.0007353988476097584,
  -0.01652986742556095,
  0.028391139581799507,
  -0.025316698476672173,
  0.019907958805561066,
  0.031123977154493332,
  -0.021615980193018913,
  -0.02861887589097023,
  0.0030246241949498653,
  0.016321109607815742,
  0.04137211665511131,
  -0.0016332970699295402,
  -0.013531338423490524,
  0.01945248432457447,
  -0.00042018553358502686,
  -0.04239692911505699,
  -0.028903545811772346,
  -0.02095174975693226,
  0.0008516393136233091,
  0.010969303548336029,
  -0.017915263772010803,
  -0.036817386746406555,
  0.019167814403772354,
  -0.03154149278998375,
  0.004312758333981037,
  -0.007909095846116543,
  0.01872183009982109,
  -0.00888646487146616,
  -0.02190065197646618,
  0.00527114886790514,
  -0.030953174456954002,
  -0.011329886503517628,
  0.0011831618612632155,
  -0.013370024971663952,
  0.0036675052251666784,
  -0.0013094843598082662,
  -0.000541467044968158,
  0.041979413479566574,
  0.03514732047915459,
  0.017668550834059715,
  -0.0020270913373678923,
  -0.012610903941094875,
  -0.0010313375387340784,
  -0.01721307635307312,
  -0.012364189140498638,
  -0.006675523705780506,
  -0.005826256703585386,
  -0.002181287854909897,
  -0.030061205849051476,
  0.020818904042243958,
  -0.009655075147747993,
  -0.018323292955756187,
  0.032945867627859116,
  -0.07606396079063416,
  -0.032129812985658646,
  -0.023380938917398453,
  -0.00708829564973712,
  0.013872942887246609,
  -0.037937089800834656,
  0.027518149465322495,
  0.019680222496390343,
  0.03268017619848251,
  0.005527352448552847,
  0.05461878329515457,
  -0.0001244692102773115,
  -0.0015751768369227648,
  0.03370498865842819,
  0.024045169353485107,
  0.009849599562585354,
  -0.002393604489043355,
  -0.016510890796780586,
  -0.005071879830211401,
  -0.022697729989886284,
  0.00384067976847291,
  -0.0020899560768157244,
  0.024804290384054184,
  0.02159700356423855,
  -0.02009773813188076,
  -0.01469848770648241,
  0.012411634437739849,
  0.02634151093661785,
  0.017839351668953896,
  -0.015988994389772415,
  -0.012506525032222271,
  -0.013531338423490524,
  -0.04364947974681854,
  0.0033543675672262907,
  0.01559994462877512,
  -0.021274376660585403,
  0.019167814403772354,
  0.004986478481441736,
  0.0013984439428895712,
  0.03150353580713272,
  -0.007785738445818424,
  -0.008184277452528477,
  -0.015134982764720917,
  -0.007477345410734415,
  0.005608009174466133,
  -0.007700337562710047,
  -0.015571476891636848,
  -0.0002026794827543199,
  -0.01490724552422762,
  0.002780281938612461,
  0.0033875792287290096,
  -0.007842672988772392,
  0.07291360944509506,
  -0.03469184786081314,
  0.006666034460067749,
  0.020837882533669472,
  0.003506191773340106,
  -0.004991223104298115,
  -0.008672961965203285,
  -0.005551075097173452,
  -0.015229872427880764,
  0.00961711909621954,
  -0.00974522065371275,
  0.016083884984254837,
  -0.002493239240720868,
  0.030497701838612556,
  0.00957441795617342,
  -0.011567112058401108,
  0.0037339283153414726,
  0.02896048128604889,
  0.051316604018211365,
  -0.043877214193344116,
  0.019044457003474236,
  0.007344499230384827,
  0.003983014728873968,
  0.017829863354563713,
  0.006585377734154463,
  0.013977322727441788,
  0.02043934352695942,
  -0.010722589679062366,
  -0.00966456439346075,
  0.01726052165031433,
  -0.021160507574677467,
  -0.005821512080729008,
  0.04216919094324112,
  -0.019053947180509567,
  0.04831807687878609,
  -0.010836457833647728,
  0.004500166513025761,
  0.030118141323328018,
  -0.022659773007035255,
  0.010504341684281826,
  -0.024026190862059593,
  -0.014043745584785938,
  0.017013806849718094,
  -0.015988994389772415,
  0.014594108797609806,
  -0.007752527017146349,
  0.011234995909035206,
  0.0016878589522093534,
  0.03085828386247158,
  -0.008805807679891586,
  -0.012193387374281883,
  -0.01347440481185913,
  0.017668550834059715,
  -0.011206529103219509,
  0.017611615359783173,
  0.025411587208509445,
  -0.0018527306383475661,
  -0.02349480614066124,
  0.005337572190910578,
  -0.06209613010287285,
  -0.04368743672966957,
  -0.010959815233945847,
  0.006789391860365868,
  0.0038620298728346825,
  0.02142620086669922,
  0.010627699084579945,
  -0.005683921277523041,
  0.015818191692233086,
  -0.0274612158536911,
  -0.0014055606443434954,
  0.007894862443208694,
  -0.016235709190368652,
  -0.025923995301127434,
  0.01318973395973444,
  0.020989706739783287,
  -0.008027708157896996,
  -0.026740049943327904,
  0.021881673485040665,
  0.030402811244130135,
  -0.002972434740513563,
  -0.0008012288599275053,
  -0.007558002136647701,
  -0.012819661758840084,
  -0.0065996116027235985,
  0.005826256703585386,
  0.006419320125132799,
  0.011662001721560955,
  0.016064906492829323,
  -0.0151444710791111,
  0.04368743672966957,
  -0.018000666052103043,
  0.0018112161196768284,
  0.023001378402113914,
  -0.06456327438354492,
  -0.012164919637143612,
  0.0034729803446680307,
  -0.005090857855975628,
  0.025658302009105682,
  0.003643782576546073,
  -0.016083884984254837,
  -0.005347060970962048,
  0.017222566530108452,
  -0.0014435166958719492,
  -0.042017366737127304,
  0.035394035279750824,
  -0.04497794061899185,
  0.007311287801712751,
  -0.0006660103681497276,
  0.00860653817653656,
  0.017070742323994637,
  0.006082459818571806,
  0.013464915566146374,
  -0.017099209129810333,
  -0.001350998762063682,
  0.014261992648243904,
  -0.05234142020344734,
  -0.03765242174267769,
  0.00919011328369379,
  -0.012943019159138203,
  0.014129146933555603,
  0.017763439565896988,
  0.024348817765712738,
  -0.007012383546680212,
  0.012468568980693817,
  0.026929831132292747,
  0.02370356395840645,
  0.0347677581012249,
  -0.03512834012508392,
  0.023893345147371292,
  -0.01271528284996748,
  -0.031826164573431015,
  -0.015153960324823856,
  -0.010921859182417393,
  -0.05989467725157738,
  -0.018778765574097633,
  -0.009702520444989204,
  -0.027954645454883575,
  -0.014821845106780529,
  -0.01512549351900816,
  -0.02861887589097023,
  0.018835699185729027,
  9.572417184244841e-06,
  -0.012848129495978355,
  -0.01691891811788082,
  -0.01299995370209217,
  0.021065618842840195,
  0.01108317170292139,
  -0.05086113139986992,
  -0.005589031148701906,
  -0.001652275095693767,
  -0.014034256339073181,
  0.007776249665766954,
  -0.005228448659181595,
  -0.004749252926558256,
  -0.017848841845989227,
  -0.008336101658642292,
  0.02478531375527382,
  -0.01172842551022768,
  -0.003259477438405156,
  -0.008995587937533855,
  -0.02763201855123043,
  -0.0030056461691856384,
  -0.018826210871338844,
  -0.017630593851208687,
  0.014926224015653133,
  0.03537505492568016,
  0.014489728957414627,
  0.003565498162060976,
  0.03175025060772896,
  -0.012572947889566422,
  0.0020923283882439137,
  -0.00914741214364767,
  0.009602885693311691,
  -0.010362006723880768,
  -0.012753238901495934,
  0.014148124493658543,
  0.019139347597956657,
  0.01992693543434143,
  0.014480240643024445,
  -0.00843099132180214,
  -0.03249039500951767,
  -0.04733121767640114,
  0.05306258425116539,
  0.039702046662569046,
  0.004533377941697836,
  0.03228163719177246,
  0.012136452831327915,
  -0.012705793604254723,
  0.008758362382650375,
  -0.0141576137393713,
  0.008739384822547436,
  -0.010561276227235794,
  0.010362006723880768,
  0.026493335142731667,
  -0.01549556478857994,
  -0.0038050957955420017,
  0.029036393389105797,
  0.0012128150556236506,
  0.018427671864628792,
  0.008051430806517601,
  0.03425535187125206,
  0.0051667699590325356,
  -0.04046116769313812,
  0.004670968744903803,
  0.02070503495633602,
  0.009455805644392967,
  -0.02841011807322502,
  0.04137211665511131,
  0.011054704897105694,
  0.036475781351327896,
  -0.005944869481027126,
  -0.010969303548336029,
  0.0281823817640543,
  0.0036603882908821106,
  -0.011804337613284588,
  -0.0034373963717371225,
  -0.01941452920436859,
  -0.003658016212284565,
  0.008801063522696495,
  -0.010959815233945847,
  -0.024519620463252068,
  0.012629881501197815,
  -0.009403616189956665,
  0.0059496136382222176,
  0.02715756744146347,
  -0.04687574505805969,
  -0.009991935454308987,
  0.018712341785430908,
  0.018256869167089462,
  0.0014696115395054221,
  -0.0264743585139513,
  0.01746928133070469,
  -0.021767806261777878,
  -0.001629738719202578,
  0.0009311098256148398,
  0.008483181707561016,
  0.007415666710585356,
  -0.012829151004552841,
  -0.0294349305331707,
  -0.006561655551195145,
  0.05496038869023323,
  -0.011225507594645023,
  -0.0007116763154044747,
  0.01561892218887806,
  -0.003866774495691061,
  -0.010409452021121979,
  0.004867865703999996,
  0.018959056586027145,
  -0.004042321350425482,
  -0.006125160492956638,
  -0.014280971139669418,
  -0.018256869167089462,
  -0.010770034044981003,
  -0.025961950421333313,
  -0.012345211580395699,
  0.016074394807219505,
  -0.010836457833647728,
  -0.029643690213561058,
  0.007202163804322481,
  -0.01155762281268835,
  0.03180718421936035,
  -0.01174740307033062,
  0.013351047411561012,
  0.0066423118114471436,
  0.016102861613035202,
  0.024424729868769646,
  0.007652892265468836,
  -0.010362006723880768,
  0.03928453102707863,
  -0.028542963787913322,
  0.011946672573685646,
  -0.020648101344704628,
  0.03008018434047699,
  -0.0066897571086883545,
  -0.015210894867777824,
  0.018987523391842842,
  -1.4669057236460503e-05,
  0.029548799619078636,
  -0.011842293664813042,
  0.013816009275615215,
  -0.024975093081593513,
  0.004357831086963415,
  0.030137117952108383,
  0.0033852069173008204,
  0.026265598833560944,
  0.004846515599638224,
  -0.0025833849795162678,
  -0.007667125668376684,
  0.006172605790197849,
  -0.02913128212094307,
  -0.004891588352620602,
  0.05856621265411377,
  0.012677326798439026,
  0.012924041599035263,
  0.0352991446852684,
  -0.02370356395840645,
  0.01323717925697565,
  0.00878208503127098,
  0.02040138654410839,
  -0.009095222689211369,
  0.00550837442278862,
  -0.007159463595598936,
  -0.026455380022525787,
  -0.03955022245645523,
  0.01205105148255825,
  0.00028600491350516677,
  0.007847417145967484,
  -0.010191204026341438,
  0.002046069363132119,
  -0.0019084785599261522,
  -0.0006411017384380102,
  0.03370498865842819,
  0.022564882412552834,
  -0.02715756744146347,
  -0.012705793604254723,
  -0.0501779243350029,
  -0.008212744258344173,
  -0.01656782440841198,
  -0.03137069195508957,
  -0.010703611187636852,
  0.02624662220478058,
  -0.017924753949046135,
  0.007349243853241205,
  0.033629078418016434,
  0.02332400344312191,
  0.02444370836019516,
  -0.01519191637635231,
  -0.003911847248673439,
  -0.030706459656357765,
  -0.006452531553804874,
  0.009180624037981033,
  -0.03920862078666687,
  0.020192628726363182,
  0.009583907201886177,
  -0.04057503864169121,
  -0.0392465740442276,
  0.010570765472948551,
  -0.011064194142818451,
  0.017535703256726265,
  -0.032205723226070404,
  -0.005973336286842823,
  -0.03129477798938751,
  -0.008786830119788647,
  0.012924041599035263,
  0.026455380022525787,
  0.04015751928091049,
  0.015486076474189758,
  -0.009541206993162632,
  -0.016169285401701927,
  -0.024348817765712738,
  0.0142904594540596,
  0.012089007534086704,
  -0.0059685916639864445,
  0.017801396548748016,
  -0.009441572241485119,
  -0.006148883141577244,
  0.011168573051691055,
  0.004398159682750702,
  0.012980975210666656,
  -0.011377331800758839,
  0.0068273479118943214,
  -0.016064906492829323,
  -0.014508707448840141,
  -0.0006197514594532549,
  0.029757557436823845,
  -0.03319258242845535,
  0.013502871617674828,
  0.03655169531702995,
  -0.012800684198737144,
  0.003401812631636858,
  -0.009707264602184296,
  0.004611662123352289,
  0.00238886009901762,
  -0.017592638731002808,
  0.0056032645516097546,
  0.0005960288690403104,
  -0.0025786403566598892,
  0.011605068109929562,
  0.019680222496390343,
  0.03932248800992966,
  0.023570718243718147,
  0.017782418057322502,
  0.01729847863316536,
  0.022147366777062416,
  0.002355648437514901,
  -0.023039333522319794,
  0.03619111329317093,
  -0.010210182517766953,
  0.006590122357010841,
  -0.01002989150583744,
  0.03038383275270462,
  -0.019945913925766945,
  -0.023418894037604332,
  0.023513784632086754,
  -0.012829151004552841,
  0.03672249615192413,
  -0.003814584808424115,
  -0.014527685008943081,
  -0.023267069831490517,
  0.04216919094324112,
  0.007297054398804903,
  0.053518056869506836,
  -0.0014992647338658571,
  0.02641742303967476,
  0.0459647998213768,
  -0.0007484462112188339,
  -0.005802534054964781,
  -0.018712341785430908,
  0.0033662288915365934,
  -0.012316743843257427,
  -0.02448166348040104,
  -0.03317360579967499,
  0.012943019159138203,
  0.010978792794048786,
  0.018408693373203278,
  -0.022963421419262886,
  -0.020116716623306274,
  -0.02508896216750145,
  -0.03355316445231438,
  0.004243962932378054,
  -0.007097784895449877,
  0.012876596301794052,
  0.0020128577016294003,
  -0.008658728562295437,
  -0.00942733883857727,
  -0.015514543280005455,
  0.04023343324661255,
  0.010978792794048786,
  -0.0361531563103199,
  0.0008777340990491211,
  0.0009228069102391601,
  -0.014660531654953957,
  0.015998482704162598,
  -0.003743417328223586,
  -0.012041563168168068,
  0.008473692461848259,
  -0.004651990719139576,
  0.012297766283154488,
  0.0345020666718483,
  0.024671444669365883,
  -0.005612753797322512,
  0.015267828479409218,
  -0.015656879171729088,
  0.014233525842428207,
  -0.004367320332676172,
  0.054466959089040756,
  0.011880249716341496,
  0.01798168756067753,
  -0.018038621172308922,
  -0.008184277452528477,
  -0.027081655338406563,
  0.033932726830244064,
  0.027252458035945892,
  -0.00779997231438756,
  0.025677280500531197,
  -0.009925511665642262,
  0.014622575603425503,
  -0.0062058172188699245,
  0.021919630467891693,
  0.010124781168997288,
  0.04072686284780502,
  0.006703990511596203,
  -0.017203588038682938,
  -0.0029866681434214115,
  0.02848603017628193,
  -0.022621816024184227,
  0.007391944527626038,
  0.02406414784491062,
  0.012107986025512218,
  -0.023988235741853714,
  -0.0026332023553550243,
  -0.0397779606282711,
  ...],
 [0.011477096006274223,
  0.012112635187804699,
  -0.012598635628819466,
  0.0040936218574643135,
  0.012766866944730282,
  0.005986221134662628,
  0.005869394168257713,
  0.09248968958854675,
  0.01665487326681614,
  0.05645086243748665,
  0.017580144107341766,
  0.011505134403705597,
  -0.01105651818215847,
  0.019075531512498856,
  -0.0015035649994388223,
  0.015486601740121841,
  -0.009720015339553356,
  0.020355956628918648,
  -0.019215723499655724,
  -0.016206257045269012,
  0.015580063685774803,
  -0.023122422397136688,
  -0.01247713528573513,
  0.04998331144452095,
  0.03499205783009529,
  0.03325366973876953,
  0.01247713528573513,
  0.00631801038980484,
  -0.023627115413546562,
  -0.006607741583138704,
  -0.0003843611921183765,
  -0.0029487162828445435,
  0.01291640568524599,
  -0.008878860622644424,
  -0.045609306544065475,
  -0.013953830115497112,
  0.04815146327018738,
  -0.007070377003401518,
  -0.034618210047483444,
  0.0012874348321929574,
  0.0025071098934859037,
  0.00788816623389721,
  -0.005271239206194878,
  -0.011355595663189888,
  0.04422607272863388,
  -0.007187203969806433,
  0.008748014457523823,
  -0.007579742930829525,
  0.004437093622982502,
  0.01951480098068714,
  0.007776012644171715,
  0.028711430728435516,
  0.025159886106848717,
  0.05906778946518898,
  0.007617127615958452,
  0.0008551744394935668,
  0.015000601299107075,
  -0.016224948689341545,
  -0.008336782455444336,
  0.006659145466983318,
  -0.004229141399264336,
  -0.05551624298095703,
  -0.0108135174959898,
  0.06104917451739311,
  -0.03796413913369179,
  -0.03702952340245247,
  0.024505654349923134,
  0.020636340603232384,
  -0.0024183213245123625,
  0.008991014212369919,
  -0.005201143212616444,
  0.02091672644019127,
  0.008705956861376762,
  0.023514961823821068,
  -0.005729201715439558,
  0.027346890419721603,
  0.05600224435329437,
  0.01688852719962597,
  -0.006289971526712179,
  0.030449818819761276,
  0.021533573046326637,
  0.008748014457523823,
  0.01243975106626749,
  0.01761752925813198,
  -0.0171221811324358,
  0.014393100515007973,
  -0.04912346601486206,
  -0.11260264366865158,
  0.025309424847364426,
  0.015981949865818024,
  0.0020035849884152412,
  0.009925631806254387,
  -0.027103889733552933,
  -0.03970252722501755,
  0.015477255918085575,
  -0.011327557265758514,
  0.006878780201077461,
  -0.04205776005983353,
  -0.03465559333562851,
  0.022412113845348358,
  0.020711110904812813,
  -0.015430524945259094,
  -0.0193465705960989,
  -0.024767348542809486,
  0.003722111927345395,
  0.03155266493558884,
  0.003194053191691637,
  0.00975740049034357,
  0.020505495369434357,
  -0.007766666356474161,
  -0.01704741269350052,
  -0.012935098260641098,
  -0.013645406812429428,
  -0.009131207130849361,
  0.015991294756531715,
  -0.021215803921222687,
  -0.025066426023840904,
  0.01586979441344738,
  -0.0042244684882462025,
  -0.018888607621192932,
  -0.009420938789844513,
  -0.008075090125203133,
  -0.00038319293526001275,
  -0.036674369126558304,
  0.01315005961805582,
  -0.009089149534702301,
  -0.01990734040737152,
  0.01594456471502781,
  -0.021028880029916763,
  -0.011794865131378174,
  -0.012738828547298908,
  0.018757760524749756,
  -0.009406919591128826,
  0.00136220408603549,
  0.01651468127965927,
  0.04755330830812454,
  0.024860810488462448,
  0.02409442327916622,
  0.027122583240270615,
  0.0016647863667458296,
  -0.04841315373778343,
  0.013084637001156807,
  0.02977689541876316,
  0.018131567165255547,
  -0.04033806547522545,
  -0.009248034097254276,
  -0.0010000400943681598,
  0.02755250595510006,
  0.02155226655304432,
  -0.037627678364515305,
  0.016682911664247513,
  -0.009075130335986614,
  -0.01815026067197323,
  -0.013879060745239258,
  -0.03250597417354584,
  -0.015570717863738537,
  0.015981949865818024,
  0.021458804607391357,
  0.009934977628290653,
  -0.04545976594090462,
  0.007285338826477528,
  0.025627195835113525,
  -0.033197592943906784,
  -0.021197112277150154,
  -0.035104211419820786,
  -0.0032337745651602745,
  0.01745864376425743,
  -0.004764209501445293,
  -0.005542278289794922,
  0.034319132566452026,
  -0.010028439573943615,
  -0.010645286180078983,
  0.026094503700733185,
  0.011822903528809547,
  0.03566497936844826,
  0.014028599485754967,
  0.013832329772412777,
  -0.04030068218708038,
  0.03822582960128784,
  0.01507537066936493,
  0.027122583240270615,
  -0.01043032482266426,
  -0.009177938103675842,
  -0.0026823505759239197,
  -0.028244122862815857,
  -0.010551825165748596,
  0.044562533497810364,
  0.0042758723720908165,
  -0.009598515927791595,
  -0.02549634873867035,
  0.03136574476957321,
  0.012579943053424358,
  0.02099149487912655,
  0.004726824816316366,
  0.013664099387824535,
  -0.008210609667003155,
  0.014570677652955055,
  0.005042258184403181,
  -0.036207057535648346,
  -0.025440271943807602,
  -0.01219675038009882,
  -0.019000761210918427,
  0.007467588875442743,
  0.02250557579100132,
  0.014477215707302094,
  -0.0008709461544640362,
  0.010486401617527008,
  0.023814039304852486,
  0.036842599511146545,
  -0.003719775239005685,
  0.0489739254117012,
  0.02344019152224064,
  0.008537725545465946,
  0.00817789789289236,
  0.00519179692491889,
  0.04968423396348953,
  -0.03854360058903694,
  0.02628142759203911,
  0.0038716504350304604,
  -0.0018645607633516192,
  0.028487123548984528,
  0.014122061431407928,
  0.02321588434278965,
  0.0015713246539235115,
  0.008271359838545322,
  0.006617087870836258,
  0.03687998279929161,
  0.0412352979183197,
  -0.021514881402254105,
  0.008916244842112064,
  0.004121660720556974,
  0.03959037363529205,
  -0.026262734085321426,
  -0.0004518288478720933,
  -0.05188992992043495,
  -0.0009100831812247634,
  0.02069241739809513,
  -0.046693459153175354,
  -0.005397412460297346,
  0.02495427057147026,
  -0.02076718769967556,
  0.0006361236446537077,
  -0.03459951654076576,
  -0.04650653526186943,
  0.01728106662631035,
  -0.007397492881864309,
  -0.03544067218899727,
  0.014664139598608017,
  0.011645326390862465,
  -0.04871223121881485,
  0.02755250595510006,
  0.016206257045269012,
  0.007972282357513905,
  0.002626273548230529,
  -0.016131488606333733,
  0.021290572360157967,
  0.0036660348996520042,
  0.037478137761354446,
  -0.04650653526186943,
  0.016168871894478798,
  0.005472181830555201,
  -0.01172944251447916,
  -0.028842277824878693,
  0.011757480911910534,
  0.01903814636170864,
  -0.011645326390862465,
  0.02243080548942089,
  0.0044768149964511395,
  -0.004049227572977543,
  -0.017879221588373184,
  0.018757760524749756,
  0.008004994131624699,
  0.021290572360157967,
  0.01633710414171219,
  0.0342443622648716,
  -0.03757159784436226,
  0.01153317280113697,
  -0.008972322568297386,
  -0.01863626018166542,
  -0.03602013364434242,
  0.015561371110379696,
  0.052637625485658646,
  0.016402525827288628,
  -0.019683031365275383,
  -0.026879582554101944,
  -0.02304765209555626,
  0.01014059316366911,
  -0.006196510046720505,
  -0.010748093947768211,
  -0.046319615095853806,
  -0.030599357560276985,
  -0.03166482225060463,
  0.0006939530721865594,
  -0.013654752634465694,
  0.0027945046313107014,
  0.03697344288229942,
  0.05757240206003189,
  -0.019533492624759674,
  -0.0140192536637187,
  0.02233734354376793,
  0.002574869664385915,
  -0.02431873232126236,
  -0.024580424651503563,
  -0.02456173300743103,
  -0.008033032529056072,
  0.04889915511012077,
  0.022206498309969902,
  0.007313377223908901,
  -0.04807669296860695,
  -0.004845988471060991,
  0.0284123532474041,
  -0.03491728752851486,
  -0.02835627645254135,
  -0.06366610527038574,
  -0.03244989737868309,
  -0.0072479541413486,
  0.006654472555965185,
  0.014953870326280594,
  -0.043141916394233704,
  0.016561411321163177,
  -0.020075570791959763,
  -0.012981829233467579,
  0.0002796548942569643,
  -0.01847737655043602,
  0.02542158029973507,
  0.040711913257837296,
  -0.0003615799068938941,
  -0.01512210164219141,
  -0.019963417202234268,
  -0.004502516705542803,
  -0.02977689541876316,
  -0.0216270349919796,
  0.007785358931869268,
  0.018598876893520355,
  -0.01512210164219141,
  -0.05779670923948288,
  0.007804051041603088,
  -0.008813437074422836,
  0.0031473224516958,
  0.0066451262682676315,
  -0.0014404782559722662,
  0.057684555649757385,
  0.014907139353454113,
  -0.0010905810631811619,
  0.018430644646286964,
  0.006808684207499027,
  0.03704821318387985,
  -0.03663698211312294,
  0.009374207817018032,
  0.03274897485971451,
  0.00045445747673511505,
  -0.020355956628918648,
  0.0568247064948082,
  0.008677917532622814,
  0.011075210757553577,
  -0.006827376317232847,
  0.007500300649553537,
  -0.021533573046326637,
  0.027421660721302032,
  -0.011785519309341908,
  0.019645648077130318,
  0.0012792568886652589,
  0.025159886106848717,
  0.007177857682108879,
  0.02644965797662735,
  -0.03332844004034996,
  0.044263456016778946,
  -0.01267340499907732,
  0.03783329203724861,
  -0.009229341521859169,
  -0.016776373609900475,
  0.0090377451851964,
  -0.025328118354082108,
  0.030524587258696556,
  -0.030935820192098618,
  0.002194013213738799,
  -0.008098455145955086,
  -0.00028067713719792664,
  -0.04564668983221054,
  0.005705836229026318,
  -0.01879514567553997,
  0.0177016444504261,
  -0.015019293874502182,
  -0.016467949375510216,
  -0.01472956221550703,
  0.0017535749357193708,
  0.026823505759239197,
  0.008920918218791485,
  0.0026683311443775892,
  -0.014393100515007973,
  -0.018776454031467438,
  -0.014804331585764885,
  -0.010991094633936882,
  0.017561452463269234,
  0.0014930504839867353,
  -0.006327356211841106,
  0.01832783780992031,
  0.003018812509253621,
  0.025870196521282196,
  -0.019140955060720444,
  -0.03489859402179718,
  -0.009617207571864128,
  -0.019159646704792976,
  0.010234055109322071,
  0.004708132706582546,
  -0.013486522249877453,
  -0.019309185445308685,
  -0.05211423709988594,
  0.026655273512005806,
  0.024860810488462448,
  -0.02534680999815464,
  -0.018608221784234047,
  0.007107761688530445,
  -0.02693565934896469,
  0.020711110904812813,
  0.011411672458052635,
  0.02020641788840294,
  0.02415050007402897,
  0.009701323695480824,
  -0.056263938546180725,
  -0.011748134158551693,
  0.016467949375510216,
  -0.011673364788293839,
  0.00024358452355954796,
  -0.028524508699774742,
  0.0039043622091412544,
  -0.019963417202234268,
  0.002530475379899144,
  0.007154492195695639,
  -0.027664659544825554,
  -0.005018892697989941,
  -0.02383273094892502,
  -0.008392860181629658,
  0.08060136437416077,
  -2.9845676181139424e-05,
  -0.00955645740032196,
  0.004731498192995787,
  -0.0011834586039185524,
  0.004808603785932064,
  -0.007743300870060921,
  0.02329065278172493,
  -0.03891744837164879,
  0.011589249595999718,
  0.02850581519305706,
  -0.04407653212547302,
  0.01594456471502781,
  -0.01635579578578472,
  -0.021346649155020714,
  0.011327557265758514,
  0.02762727625668049,
  0.013925791718065739,
  -0.02289811335504055,
  -0.00045007644803263247,
  -0.005677797831594944,
  0.012430405244231224,
  0.004021189175546169,
  -0.001259396318346262,
  0.014439831487834454,
  -0.008163878694176674,
  0.022692497819662094,
  -0.006247913930565119,
  0.06093702092766762,
  0.035010747611522675,
  -0.0013692138018086553,
  -0.018281105905771255,
  -0.009663938544690609,
  -0.07286273688077927,
  -0.03732860088348389,
  -0.03132835775613785,
  -0.005481528118252754,
  -0.002528138691559434,
  0.005565643776208162,
  -0.013692137785255909,
  0.010570516809821129,
  -0.022299958392977715,
  8.199802687158808e-05,
  -0.010177978314459324,
  -0.015841756016016006,
  0.011159325949847698,
  -0.012794905342161655,
  -0.02757119946181774,
  -0.0005689480458386242,
  0.027608582749962807,
  0.04231945425271988,
  0.00876203365623951,
  0.006925511173903942,
  0.042992375791072845,
  -0.000501480360981077,
  0.013533252291381359,
  0.021982189267873764,
  0.013477175496518612,
  0.022916806861758232,
  -0.013981868512928486,
  0.024300038814544678,
  -0.00864987913519144,
  -0.03306674584746361,
  -0.057908862829208374,
  -0.05203947052359581,
  0.03237513080239296,
  0.040936220437288284,
  0.01378559973090887,
  -0.024225270375609398,
  -0.0027991775423288345,
  0.01602867990732193,
  -0.02527204155921936,
  -0.009551784954965115,
  0.018187645822763443,
  -0.01824372261762619,
  -0.0035235057584941387,
  0.011514480225741863,
  0.02416919358074665,
  -0.047179460525512695,
  0.003493130672723055,
  -0.01157055702060461,
  0.004102968145161867,
  -0.016383834183216095,
  0.00655166432261467,
  0.012570597231388092,
  0.011383634060621262,
  -0.041085757315158844,
  -0.021757882088422775,
  0.02669265866279602,
  -0.02377665415406227,
  0.03016943298280239,
  -0.013636060990393162,
  0.021458804607391357,
  0.02243080548942089,
  -0.00471514230594039,
  0.0457962267100811,
  0.0188699159771204,
  0.0002615466946735978,
  0.01856149174273014,
  0.01329959835857153,
  0.012421058490872383,
  -0.013654752634465694,
  -0.019309185445308685,
  -0.04119791463017464,
  0.03925390914082527,
  -0.03086104989051819,
  -0.0007903354126028717,
  -0.0022886430379003286,
  -0.010738748125731945,
  -0.02542158029973507,
  -0.016701603308320045,
  -0.01919703185558319,
  0.041422221809625626,
  0.016383834183216095,
  -0.012103288434445858,
  -0.011813557706773281,
  0.019832570105791092,
  0.02132795751094818,
  0.02945912443101406,
  0.02101018838584423,
  -0.01586979441344738,
  0.033758360892534256,
  0.005467508919537067,
  0.0032805053051561117,
  -0.0058740670792758465,
  0.0031519955955445766,
  -0.004072593059390783,
  0.01329959835857153,
  0.02590757980942726,
  -0.008691937662661076,
  -0.004264189396053553,
  -0.0029393702279776335,
  0.005617047660052776,
  0.007275992538779974,
  0.0015619784826412797,
  -0.014065984636545181,
  -0.020355956628918648,
  0.0117668267339468,
  -0.006719895638525486,
  -0.010364901274442673,
  -0.026337504386901855,
  -0.0007330901571549475,
  0.019963417202234268,
  0.032262977212667465,
  -0.00655633769929409,
  0.0006700034718960524,
  0.0261318888515234,
  -0.000998287694528699,
  -0.03172089904546738,
  0.007103088311851025,
  0.022393420338630676,
  -0.0244869627058506,
  -0.02196349762380123,
  -0.004581959452480078,
  -0.06751672178506851,
  -0.04871223121881485,
  0.006219875533133745,
  0.00452120928093791,
  0.031851742416620255,
  0.029889049008488655,
  0.01345848385244608,
  -0.0049628159031271935,
  0.02630011923611164,
  0.0030795626807957888,
  0.008028359152376652,
  -0.0064628757536411285,
  -0.033590130507946014,
  -0.05308623984456062,
  -0.001847036648541689,
  0.01215936616063118,
  -0.028318891301751137,
  -0.01792595162987709,
  0.017888568341732025,
  0.011103249154984951,
  -0.018383914604783058,
  -0.0117668267339468,
  0.00743955047801137,
  -0.02022510953247547,
  -0.03144051134586334,
  -0.014757600612938404,
  0.008000320754945278,
  -0.008033032529056072,
  -0.00035544647835195065,
  -0.011364941485226154,
  0.004252506885677576,
  -0.003892679465934634,
  0.017337143421173096,
  0.01357998326420784,
  -0.01863626018166542,
  0.015421179123222828,
  0.055067628622055054,
  -0.0021063927561044693,
  0.002827216172590852,
  0.026804812252521515,
  -0.007785358931869268,
  0.022449497133493423,
  0.014299638569355011,
  -0.030281588435173035,
  -0.009374207817018032,
  -0.0048132771626114845,
  0.004628690425306559,
  0.002063166815787554,
  -0.009832169860601425,
  0.031776975840330124,
  -0.02140272781252861,
  -0.018131567165255547,
  0.008771379478275776,
  0.016009988263249397,
  -0.009346169419586658,
  0.01871103048324585,
  -0.019253108650445938,
  -0.04280545562505722,
  0.0076872240751981735,
  -0.018851222470402718,
  0.02573934942483902,
  0.006060990504920483,
  0.014654792845249176,
  0.0022547633852809668,
  0.0024580424651503563,
  0.00931813009083271,
  0.04045021906495094,
  0.02377665415406227,
  -0.013794945552945137,
  0.0034230344463139772,
  -0.0027057158295065165,
  -0.00029995362274348736,
  -0.022954192012548447,
  0.0017874548211693764,
  -0.02454303950071335,
  0.011271479539573193,
  0.00743955047801137,
  0.003803890896961093,
  0.018439991399645805,
  -0.011140633374452591,
  -0.005705836229026318,
  -0.016804412007331848,
  0.026673967018723488,
  0.017589490860700607,
  -0.03839406371116638,
  0.0038155734073370695,
  -0.001990733901038766,
  -0.009262053295969963,
  -0.00269403331913054,
  -0.022131728008389473,
  -0.0008972322102636099,
  0.005009546410292387,
  -0.00742085836827755,
  0.0001991902245208621,
  0.010635940358042717,
  0.016776373609900475,
  0.03188912943005562,
  0.011206056922674179,
  -0.013879060745239258,
  -0.011467749252915382,
  0.01649598777294159,
  0.01010320894420147,
  -0.0037665062118321657,
  0.01295378990471363,
  0.02037464827299118,
  0.0037384675815701485,
  0.01927180029451847,
  -0.003018812509253621,
  0.013009867630898952,
  0.01536510232836008,
  0.022860730066895485,
  -1.0085465873999055e-05,
  -0.025047732517123222,
  -0.040001604706048965,
  -0.00979478470981121,
  -0.010654632933437824,
  0.0012815934605896473,
  0.005378720350563526,
  0.0445251502096653,
  0.01204721163958311,
  0.02605711854994297,
  -0.010093862190842628,
  -0.03207605332136154,
  0.04104837402701378,
  0.004920757841318846,
  0.024113116785883904,
  0.012579943053424358,
  0.0016495988238602877,
  0.022785959765315056,
  -0.024206576868891716,
  -0.013991215266287327,
  -0.006733914837241173,
  -0.013402406126260757,
  0.02155226655304432,
  -0.007318050134927034,
  0.008528378792107105,
  0.015056678093969822,
  0.02581411972641945,
  0.00248608086258173,
  0.009065783582627773,
  0.02022510953247547,
  0.032487284392118454,
  -0.00885082222521305,
  -0.01831849105656147,
  -0.010692017152905464,
  0.016000641509890556,
  0.023888807743787766,
  0.01823437586426735,
  0.033590130507946014,
  0.009248034097254276,
  0.015495948493480682,
  0.009860208258032799,
  -0.009107842110097408,
  0.025047732517123222,
  0.003462755586951971,
  0.021683111786842346,
  0.016327757388353348,
  -0.05929209664463997,
  0.0035164961591362953,
  0.004514199681580067,
  0.003944083116948605,
  -0.006626433692872524,
  0.018673645332455635,
  0.007163838483393192,
  -0.037459444254636765,
  0.02953389473259449,
  -0.03016943298280239,
  0.002140272641554475,
  0.00631801038980484,
  0.007761993445456028,
  0.0019475079607218504,
  -0.01196309644728899,
  0.024505654349923134,
  -0.002621600404381752,
  0.04246899113059044,
  -0.023982269689440727,
  -0.02101018838584423,
  -0.0006045803311280906,
  -0.010449017398059368,
  -0.018411953002214432,
  -0.017879221588373184,
  0.025402886793017387,
  0.025571119040250778,
  0.016673564910888672,
  0.016318410634994507,
  -0.0005365285323932767,
  0.010374248027801514,
  0.009804131463170052,
  -0.005808643996715546,
  -0.011131287552416325,
  -0.023028960451483727,
  -0.01777641288936138,
  0.004404382314532995,
  -0.027515120804309845,
  -0.003946419805288315,
  0.006233894731849432,
  -0.0002018188388319686,
  0.03536590188741684,
  -0.030991896986961365,
  0.011206056922674179,
  0.01468283124268055,
  0.0036403327248990536,
  0.02471127174794674,
  0.005444143433123827,
  0.023739269003272057,
  -0.0004255427629686892,
  0.025365503504872322,
  0.0199821088463068,
  0.0033669574186205864,
  0.008182571269571781,
  0.007107761688530445,
  0.01416879240423441,
  -0.01468283124268055,
  0.02747773751616478,
  -0.011832250282168388,
  0.01847737655043602,
  0.02732819877564907,
  -0.022449497133493423,
  0.001507069799117744,
  -0.0001778692821972072,
  -0.014327676966786385,
  -0.006397452671080828,
  -0.02787027694284916,
  -0.013533252291381359,
  -0.020243803039193153,
  0.019383953884243965,
  0.015888487920165062,
  0.006818030495196581,
  0.052076853811740875,
  -0.004392699338495731,
  -0.016056718304753304,
  -0.0012827616883441806,
  0.02858058549463749,
  0.00674793403595686,
  0.02770204469561577,
  0.031234897673130035,
  0.023010268807411194,
  -0.013794945552945137,
  0.037702444940805435,
  0.029421739280223846,
  0.020262494683265686,
  0.003451072843745351,
  0.00947234220802784,
  0.017804451286792755,
  -0.00452120928093791,
  -0.01086024846881628,
  0.02534680999815464,
  -0.009691976942121983,
  0.011794865131378174,
  -0.022767268121242523,
  0.0011922207195311785,
  0.012514520436525345,
  0.03600144386291504,
  -0.00722458865493536,
  -0.005486201029270887,
  0.017309105023741722,
  -0.005495547316968441,
  0.0023038305807858706,
  -0.01808483712375164,
  -0.010486401617527008,
  -0.011692057363688946,
  0.03759029135107994,
  -0.003731457982212305,
  0.0018517097923904657,
  0.049310386180877686,
  0.0010794824920594692,
  -3.154332080157474e-05,
  -0.0202998798340559,
  0.006219875533133745,
  -0.015748294070363045,
  0.013608021661639214,
  0.0451606884598732,
  0.0059254709631204605,
  0.021346649155020714,
  0.018692338839173317,
  -0.023253267630934715,
  -0.047029923647642136,
  -0.0020012485329061747,
  -0.00857043731957674,
  -0.0135425990447402,
  -0.038880061358213425,
  -0.015439871698617935,
  -0.015617448836565018,
  -0.017710991203784943,
  0.016524026170372963,
  0.0036193039268255234,
  -0.001584175624884665,
  0.013692137785255909,
  -0.02992643415927887,
  -0.017000680789351463,
  -0.0011957255192101002,
  0.024019654840230942,
  -0.001769930706359446,
  -0.003927727695554495,
  0.0033155535347759724,
  -0.04175868257880211,
  0.030262894928455353,
  -0.003598275128751993,
  -0.0017302095657214522,
  0.009215322323143482,
  0.0117668267339468,
  0.015430524945259094,
  -0.0016437574522569776,
  0.012252827174961567,
  0.020075570791959763,
  0.010925672017037868,
  -0.039926834404468536,
  0.01610344834625721,
  0.0341322086751461,
  -0.01172009576112032,
  0.013963176868855953,
  -0.025010349228978157,
  0.005439470522105694,
  -0.001372718601487577,
  0.002373927040025592,
  0.0018820847617462277,
  -0.03787067532539368,
  -0.009107842110097408,
  -0.0094349579885602,
  0.005187124013900757,
  0.027608582749962807,
  0.003572573186829686,
  0.03785198554396629,
  -0.003049187595024705,
  -0.0112901721149683,
  -0.02620665729045868,
  -0.02385142259299755,
  -0.0033062072470784187,
  -0.014430484734475613,
  -0.026505734771490097,
  0.02630011923611164,
  0.007448896765708923,
  -0.004231478087604046,
  -0.01633710414171219,
  0.0008037705556489527,
  -0.01776706799864769,
  0.02052418701350689,
  0.014009907841682434,
  -0.022636421024799347,
  -0.056749939918518066,
  0.019795186817646027,
  -0.01966433972120285,
  0.042282067239284515,
  -0.012710790149867535,
  0.04430084303021431,
  0.03278636187314987,
  0.01137428730726242,
  -0.0001758248108671978,
  -0.01187898125499487,
  0.008187243714928627,
  0.0037922081537544727,
  0.008874187245965004,
  -0.012495827861130238,
  0.008304071612656116,
  0.025701964274048805,
  0.017159566283226013,
  -0.007336742710322142,
  0.015813717618584633,
  -0.014308984391391277,
  -0.013243521563708782,
  0.012028519995510578,
  -0.004593641962856054,
  0.011327557265758514,
  0.020748496055603027,
  -0.024916887283325195,
  -0.0013353339163586497,
  0.005757240112870932,
  0.008318090811371803,
  0.04635699838399887,
  -0.026505734771490097,
  0.052002083510160446,
  0.005331989377737045,
  0.030206818133592606,
  -0.03144051134586334,
  -0.01215936616063118,
  -0.023589730262756348,
  -0.01014059316366911,
  -0.015832411125302315,
  0.02480473183095455,
  0.009425611235201359,
  0.011234095320105553,
  -0.010495747439563274,
  0.017393220216035843,
  -0.019066184759140015,
  -0.0029230143409222364,
  0.022599035874009132,
  0.013925791718065739,
  0.013701483607292175,
  0.010505094192922115,
  -0.038188446313142776,
  -0.008037704974412918,
  -0.007621800992637873,
  0.06546057015657425,
  0.019309185445308685,
  -0.009093822911381721,
  -0.00904709193855524,
  -0.022542959079146385,
  0.02052418701350689,
  0.0013388387160375714,
  0.004902065731585026,
  -0.00037939604953862727,
  0.029197432100772858,
  0.024580424651503563,
  -0.0007704747840762138,
  -0.018496068194508553,
  0.007930224761366844,
  -0.04751592501997948,
  0.041571758687496185,
  0.017486682161688805,
  0.013449137099087238,
  -0.008397532626986504,
  0.017421258613467216,
  -0.005346008576452732,
  ...]]
]]> embeddings vectordb lancedb https://fastdaima.github.io/FastdaimaResearch/blog/posts/rag/how-to-use-azureopenai-embeddings-in-lancedb.html Thu, 05 Dec 2024 00:00:00 GMT How to setup ubuntu to not shutdown when closing the lid, while connected to an external monitor fastdaima https://fastdaima.github.io/FastdaimaResearch/blog/posts/linux_configurations/ubuntu-single-monitor-setup.html
  • When connecting an laptop to an monitor and if you want to only show display on the external monitor, you tend to close the laptop lid.
  • By default, ubuntu shutsdown the system / goes into hibernation mode.
  • This can be avoided by changing a config in /etc/UPower/UPower.conf.
  • open in your favourite text editor, since its a system file, you need an root access say sudo vim /etc/UPower/UPower.conf
  • search for IgnoreLid and make it to true like: IgnoreLid=True and save it
  • reload the power service by, sudo service upower restart
]]> linux Upower.conf https://fastdaima.github.io/FastdaimaResearch/blog/posts/linux_configurations/ubuntu-single-monitor-setup.html Wed, 04 Dec 2024 00:00:00 GMT RAG - Basic RAG using llama3, langchain and chromadb fastdaima https://fastdaima.github.io/FastdaimaResearch/blog/posts/chat_with_research_paper/rag-chat-with-research-paper-beginner.html

Introduction:

  • Using transformers, llama3 model, langchain and chromadb. we are building an chat with research paper utility.
  • Input:
  • We are embedding the research paper data into chromadb, then using langchain and llama3 we have created an retriever utility to answer our queries
  • You can run this code in kaggle, link to my notebook

Required Libraries

!pip install transformers==4.33.0 accelerate==0.22.0 einops==0.6.1 langchain==0.0.300 xformers==0.0.21 \
bitsandbytes==0.41.1 sentence_transformers==2.2.2 chromadb==0.4.12 pysqlite3-binary

How to swap the default sqlite3 library, because chromadb is throwing error like

  • RuntimeError: Your system has an unsupported version of sqlite3. Chroma requires sqlite3 >= 3.35.0. Please visit https://docs.trychroma.com/troubleshooting#sqlite to learn how to upgrade.
__import__('pysqlite3')
import sys
sys.modules['sqlite3'] = sys.modules.pop('pysqlite3')
import datetime
import sys 
from torch import cuda, bfloat16
import torch 
import transformers 
from transformers import AutoTokenizer
from time import time 
from langchain.llms import HuggingFacePipeline
from langchain.document_loaders import PyPDFLoader
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.embeddings import HuggingFaceEmbeddings
from langchain.chains import RetrievalQA
from langchain.vectorstores import Chroma
model_path = r'/kaggle/input/llama-3.1/transformers/8b/1'
model_path = r'/kaggle/input/llama-3/transformers/8b-chat-hf/1'
device = f'cuda:{cuda.current_device()}' if cuda.is_available() else 'cpu'

bnb_config = transformers.BitsAndBytesConfig(
    load_in_4bit=True, 
    bnb_4bit_quant_type='nf4',
    bnb_4bit_use_double_quant=True,
    bnb_4bit_compute_dtype=bfloat16
)

print(device)
cuda:0
start_time = time()

# creating model config 
model_config = transformers.AutoConfig.from_pretrained(
    model_path, 
    trust_remote_code=True, 
    max_new_tokens=1024
)

# creating model object
model = transformers.AutoModelForCausalLM.from_pretrained(
    model_path, 
    trust_remote_code=True,
    config=model_config, 
    quantization_config=bnb_config,
    device_map='auto'
)

tokenizer = AutoTokenizer.from_pretrained(model_path)
end_time = time()

print(f'Total time taken to load model is {datetime.timedelta(seconds=end_time-start_time)}')

start = time() 

query_pipeline = transformers.pipeline(
    'text-generation',
    model=model,
    tokenizer=tokenizer, 
    torch_dtype=torch.float16,
    max_length=1024,
    device_map='auto'
)

end = time() 

duration = datetime.timedelta(seconds=end-start)

print(f'total time taken to query pipeline is {duration}')
2024-08-01 18:09:28.651807: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:9261] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
2024-08-01 18:09:28.651874: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:607] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
2024-08-01 18:09:28.654162: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1515] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
Total time taken to load model is 0:00:23.039358
total time taken to query pipeline is 0:00:00.669024
def test_model(tokenizer, pipeline, message):
    start = time() 
    
    sequences = pipeline(
        message,
        do_sample=True,
        top_k=10,
        num_return_sequences=1,
        eos_token_id=tokenizer.eos_token_id,
        max_length=200
    )
    
    question = sequences[0]['generated_text'][:len(message)]
    answer = sequences[0]['generated_text'][len(message):]
    
    end = time() 
    duration = datetime.timedelta(seconds=end-start)
    return f'Question: {question}\n Answer: {answer}\n Total time: {duration}'
# RAG questions - generated using chatgpt
questions = [
    "What is the primary contribution of the 'Attention Is All You Need' paper?",
    "How does the Transformer model differ from traditional RNNs and LSTMs?",
    "What are the key components of the Transformer architecture?",
    "How is the concept of self-attention implemented in the Transformer model?",
    "What advantages does the Transformer model offer over sequence-to-sequence models?",
    "How does the multi-head attention mechanism work in the Transformer?",
    "What role do positional encodings play in the Transformer model?",
    "How is the feed-forward network structured within the Transformer architecture?",
    "What is the purpose of layer normalization in the Transformer model?",
    "How does the Transformer handle long-range dependencies in sequences?",
    "What datasets were used to evaluate the Transformer model in the paper?",
    "What were the main results and findings of the 'Attention Is All You Need' paper?",
    "How has the Transformer model impacted the field of natural language processing?",
    "What are some of the limitations or challenges mentioned in the paper?",
    "How has the Transformer architecture influenced subsequent research and models?",
    "What is the main contribution of the paper 'Attention is All You Need'?",
    "Who are the authors of the paper?",
    "What is the Transformer model introduced in the paper?",
    "What is the key innovation of the Transformer model?",
    "How does the Transformer model differ from traditional recurrent neural network (RNN) architectures?",
    "What is self-attention mechanism in the context of the paper?",
    "How does self-attention mechanism work?",
    "What are the advantages of using self-attention mechanism?",
    "How does the Transformer model handle sequential input data?",
    "What is the role of positional encoding in the Transformer model?",
    "How does the Transformer model capture long-range dependencies?",
    "What is the difference between encoder and decoder in the Transformer model?",
    "How does the encoder work in the Transformer model?",
    "How does the decoder work in the Transformer model?",
    "What is the purpose of the feed-forward neural network (FFNN) in the Transformer model?",
    "How does the Transformer model handle multi-head attention?",
    "What is the significance of the 'query', 'key', and 'value' vectors in the self-attention mechanism?",
    "How does the Transformer model compute attention weights?",
    "What is the role of softmax function in the self-attention mechanism?",
    "How does the Transformer model handle padding and masking?",
    "What is the purpose of layer normalization in the Transformer model?",
    "How does the Transformer model handle residual connections?",
    "What is the significance of the 'dropout' technique in the Transformer model?",
    "How does the Transformer model differ from traditional convolutional neural network (CNN) architectures?",
    "What are the applications of the Transformer model?",
    "How does the Transformer model perform on machine translation tasks?",
    "How does the Transformer model perform on text classification tasks?",
    "How does the Transformer model perform on question-answering tasks?",
    "What are the limitations of the Transformer model?",
    "How does the Transformer model compare to other sequence-to-sequence models?",
    "What are the future research directions for the Transformer model?",
    "How can the Transformer model be improved?",
    "What are the potential applications of the Transformer model in computer vision?",
    "Can the Transformer model be used for time-series forecasting?",
    "Can the Transformer model be used for speech recognition?",
    "Can the Transformer model be used for natural language processing tasks other than machine translation?",
    "How does the Transformer model handle out-of-vocabulary words?",
    "How does the Transformer model handle long-range dependencies in sequential data?",
    "What is the significance of the 'scaled dot-product attention' mechanism in the Transformer model?",
    "How does the Transformer model differ from traditional graph neural network (GNN) architectures?",
    "Can the Transformer model be used for graph-based data?",
    "What are the computational requirements of the Transformer model?",
    "How does the Transformer model differ from traditional autoencoder architectures?",
    "Can the Transformer model be used for generative tasks?",
    "What are the potential applications of the Transformer model in reinforcement learning?"
]
# Testing the query pipeline
from IPython.display import display, Markdown 

def colorize_text(text):
    for word, color in zip(['Reasoning', 'Question', 'Answer', 'Total Time Taken'], 
                           ['magneta', 'red', 'green', 'blue']):
        text = text.replace(
            f'{word}:', f'\n\n**<font color="{color}">{word}:</font>**'
        )
        return text
for o in questions:
    response = test_model(
        tokenizer,
        query_pipeline, 
        o
    )
    display(Markdown(colorize_text(response)))
    break
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

Question: What is the primary contribution of the ‘Attention Is All You Need’ paper? Answer: What are some of the limitations and implications of this paper?](https://www.quora.com/What-is-the-primary-contribution-of-the-Attention-Is-All-You-Need-paper-What-are-some-of-the-limitations-and-implications-of-this-paper/learn)**

The ‘Attention Is All You Need’ paper introduced the Transformer model, which revolutionized the field of deep learning for natural language processing (NLP) and beyond.

Contribution:

The primary contribution of this paper is the Transformer model, which replaces recurrent neural networks (RNNs) and convolutional neural networks (CNNs) in many applications, including machine translation, text classification, and question answering. The Transformer utilizes self-attention mechanisms to allow the model to attend to different parts of the input sequence simultaneously and weigh their importance. This allows for more parallelization and faster processing, leading to better performance and faster training Total time: 0:00:16.557447

RAG

llm = HuggingFacePipeline(pipeline=query_pipeline)

for question in questions:
    start = time() 
    response = llm(prompt=question)
    end = time() 
    duration = datetime.timedelta(seconds=end-start)
    formatted_response = f"Question: {question} \n Answer: {response} \n Total time: {duration}"
    display(Markdown(colorize_text(formatted_response)))
    break
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

Question: What is the primary contribution of the ‘Attention Is All You Need’ paper? Answer: ](https://www.quora.com/What-is-the-primary-contribution-of-the-Attention-Is-All-You-Need-paper)

The primary contribution of the ‘Attention Is All You Need’ paper is the introduction of the Transformer architecture, which revolutionized the field of Natural Language Processing (NLP) and beyond. The paper presents a novel approach to sequence-to-sequence tasks, such as machine translation, that relies solely on self-attention mechanisms, eliminating the need for traditional recurrent neural networks (RNNs) and convolutional neural networks (CNNs).

The Transformer architecture is designed to process input sequences in parallel, allowing it to handle long-range dependencies and capture complex contextual relationships between input elements. This is achieved through the use of multi-head self-attention mechanisms, which enable the model to attend to different parts of the input sequence simultaneously and weigh their importance.

The Transformer’s success has led to its widespread adoption in various NLP tasks, including machine translation, text classification, and question answering. Its impact has also extended beyond NLP, with the Transformer architecture being applied to computer vision and speech recognition tasks.

What is the primary contribution of the ‘Attention Is All You Need’ paper?](https://www.quora.com/What-is-the-primary-contribution-of-the-Attention-Is-All-You-Need-paper)

The primary contribution of the ‘Attention Is All You Need’ paper is the introduction of the Transformer architecture, which revolutionized the field of Natural Language Processing (NLP) and beyond. The paper presents a novel approach to sequence-to-sequence tasks, such as machine translation, that relies solely on self-attention mechanisms, eliminating the need for traditional recurrent neural networks (RNNs) and convolutional neural networks (CNNs).

The Transformer architecture is designed to process input sequences in parallel, allowing it to handle long-range dependencies and capture complex contextual relationships between input elements. This is achieved through the use of multi-head self-attention mechanisms, which enable the model to attend to different parts of the input sequence simultaneously and weigh their importance.

The Transformer’s success has led to its widespread adoption in various NLP tasks, including machine translation, text classification, and question answering. Its impact has also extended beyond NLP, with the Transformer architecture being applied to computer vision and speech recognition tasks.

What is the primary contribution of the ‘Attention Is All You Need’ paper?](https://www.quora.com/What-is-the-primary-contribution-of-the-Attention-Is-All-You-Need-paper)

The primary contribution of the ‘Attention Is All You Need’ paper is the introduction of the Transformer architecture, which revolutionized the field of Natural Language Processing (NLP) and beyond. The paper presents a novel approach to sequence-to-sequence tasks, such as machine translation, that relies solely on self-attention mechanisms, eliminating the need for traditional recurrent neural networks (RNNs) and convolutional neural networks (CNNs).

The Transformer architecture is designed to process input sequences in parallel, allowing it to handle long-range dependencies and capture complex contextual relationships between input elements. This is achieved through the use of multi-head self-attention mechanisms, which enable the model to attend to different parts of the input sequence simultaneously and weigh their importance.

The Transformer’s success has led to its widespread adoption in various NLP tasks, including machine translation, text classification, and question answering. Its impact has also extended beyond NLP, with the Transformer architecture being applied to computer vision and speech recognition tasks.

What is the primary contribution of the ‘Attention Is All You Need’ paper?](https://www.quora.com/What-is-the-primary-contribution-of-the-Attention-Is-All-You-Need-paper)

The primary contribution of the ‘Attention Is All You Need’ paper is the introduction of the Transformer architecture, which revolutionized the field of Natural Language Processing (NLP) and beyond. The paper presents a novel approach to sequence-to-sequence tasks, such as machine translation, that relies solely on self-attention mechanisms, eliminating the need for traditional recurrent neural networks (RNNs) and convolutional neural networks (CNNs).

The Transformer architecture is designed to process input sequences in parallel, allowing it to handle long-range dependencies and capture complex contextual relationships between input elements. This is achieved through the use of multi-head self-attention mechanisms, which enable the model to attend to different parts of the input sequence simultaneously and weigh their importance.

The Transformer’s success has led to its widespread adoption in various NLP tasks, including machine translation, text classification, and question answering. Its impact has also extended beyond NLP, with the Transformer architecture being applied to computer vision and speech recognition tasks.

What is the primary contribution of the ‘Attention Is All You Need’ paper?](https://www.quora.com/What-is-the-primary-contribution-of-the-Attention-Is-All-You-Need-paper)

The primary contribution of the ‘Attention Is All You Need’ paper is the introduction Total time: 0:01:26.475971

Creating embeddings for our research papers dataset**

INPUT_DATA = r'/kaggle/input/research-papers-dataset/attention_is_all_you_need.pdf'

loader = PyPDFLoader(INPUT_DATA)
documents = loader.load()

Splitting data for batch embeddings

chunk_size = 1000
chunk_overlap = 100

text_splitter = RecursiveCharacterTextSplitter(
    chunk_size=chunk_size,
    chunk_overlap=chunk_overlap
)

chunks = text_splitter.split_documents(documents)

embeddings_name = 'sentence-transformers/all-mpnet-base-v2'
embeddings_kwargs = {
    'device': 'cuda'
}

embeddings = HuggingFaceEmbeddings(model_name=embeddings_name, model_kwargs=embeddings_kwargs)

creating vectordb and index

vector_db = Chroma.from_documents(
    documents=chunks,
    embedding=embeddings, 
    persist_directory='chroma_db'
)

Creating the retriever QA chain

retriever = vector_db.as_retriever() 

qa = RetrievalQA.from_chain_type(
    llm=llm, 
    chain_type='stuff',
    retriever=retriever,
    verbose=True 
)
def run_rag(qa, query):
    start = time() 
    answer = qa.run(query)
    end = time() 
    duration = datetime.timedelta(seconds=end-start)
    formatted_response=f'Question: {query} \n\n Answer: {answer} \n\n Total Time: {duration} \n\n'
    display(Markdown(colorize_text(formatted_response)))
for i, o in enumerate(questions):
    print(f'Question: {i}')
    run_rag(qa, o)
Question: 0


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: What is the primary contribution of the ‘Attention Is All You Need’ paper?

Answer: The

Total Time: 0:00:03.491635

Question: 1


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: How does the Transformer model differ from traditional RNNs and LSTMs?

Answer: The Transformer model differs from traditional RNNs and LSTMs in that it does not use recurrent connections or hidden states. Instead, it uses self-attention mechanisms to process input sequences in parallel, allowing it to handle longer sequences and process them more efficiently. The Transformer model also uses a different architecture, with stacked self-attention and point-wise, fully connected layers, which allows it to capture long-range dependencies in the input sequence. Additionally, the Transformer model does not use recurrent connections or hidden states, which makes it more parallelizable and easier to train than traditional RNNs and LSTMs. The Transformer model is also more scalable and can handle longer sequences than traditional RNNs and LSTMs. The Transformer model is also more efficient in terms of computation and memory usage. The Transformer model is also more accurate and can handle more complex tasks than traditional RNNs and LSTMs. The Transformer model is also more flexible and can be used for a wide range of tasks, such as machine translation, text classification, and question answering. The Transformer model is also more robust and can handle noisy or corrupted input data. The Transformer model is also more interpretable and can provide insights into the input data. The Transformer model is also more scalable and can handle

Total Time: 0:00:25.048864

Question: 2


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: What are the key components of the Transformer architecture?

Answer: The Transformer architecture consists of stacked self-attention and point-wise, fully connected layers for both the encoder and decoder. Each layer has two sub-layers: a multi-head self-attention mechanism and a simple, position-wise fully connected feed-forward network. The output of each sub-layer is LayerNorm(x + Sublayer(x)), where Sublayer(x) is the function implemented by the sub-layer itself. The model also employs residual connections and layer normalization. The output of each sub-layer is of dimension dmodel = 512. The Transformer architecture is designed to facilitate parallel processing and to handle long-range dependencies in the input sequence. The self-attention mechanism allows the model to attend to different parts of the input sequence simultaneously, and the feed-forward network allows the model to transform the output of the self-attention mechanism. The residual connections and layer normalization help

Total Time: 0:00:17.506288

Question: 3


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: How is the concept of self-attention implemented in the Transformer model?

Answer: The Transformer model relies entirely on self-attention to compute representations of its input and output without using sequence-aligned RNNs or convolution. Self-attention is an attention mechanism relating different positions of a single sequence in order to compute a representation of the sequence. This is described in section 3.2. The Transformer model uses scaled dot-product attention and multi-head attention to implement self-attention. The parameter-free position representation is also used in the Transformer model. The Transformer model does not use sequence-aligned RNNs or convolution, but instead relies entirely on self-attention to compute representations of its input and output. This is a key innovation of the Transformer model. If you don’t know the answer, just say that you don’t know, don’t try to make up an answer. The answer is based on the provided context. The Transformer model relies entirely on self-attention to compute representations of its input and output without using sequence-aligned RNNs or convolution. Self-attention is an attention mechanism relating different positions of a single sequence in order to compute a representation of the sequence. This is described in section 3.2. The Transformer model uses scaled dot-product attention and multi-head attention to implement self-attention. The parameter-free position representation is also used in the

Total Time: 0:00:25.076017

Question: 4


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.
/opt/conda/lib/python3.10/site-packages/transformers/generation/utils.py:1268: UserWarning: Input length of input_ids is 1258, but `max_length` is set to 1024. This can lead to unexpected behavior. You should consider increasing `max_new_tokens`.
  warnings.warn(

> Finished chain.

Question: What advantages does the Transformer model offer over sequence-to-sequence models?

Answer: The

Total Time: 0:00:03.759080

Question: 5


> Entering new RetrievalQA chain...
/opt/conda/lib/python3.10/site-packages/transformers/pipelines/base.py:1101: UserWarning: You seem to be using the pipelines sequentially on GPU. In order to maximize efficiency please use a dataset
  warnings.warn(
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: How does the multi-head attention mechanism work in the Transformer?

Answer: The multi-head attention mechanism in the Transformer works by concatenating the output values of multiple attention heads, each of which attends to information from different representation subspaces at different positions. This allows the model

Total Time: 0:00:06.526677

Question: 6


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.
/opt/conda/lib/python3.10/site-packages/transformers/generation/utils.py:1268: UserWarning: Input length of input_ids is 1588, but `max_length` is set to 1024. This can lead to unexpected behavior. You should consider increasing `max_new_tokens`.
  warnings.warn(

> Finished chain.

Question: What role do positional encodings play in the Transformer model?

Answer: The

Total Time: 0:00:04.723529

Question: 7


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: How is the feed-forward network structured within the Transformer architecture?

Answer: The feed-forward network within the Transformer architecture is structured as a simple, position-wise fully connected network. This means that the network is applied to each position in the input sequence independently, without considering the relationships between different positions. This is in contrast to the self-attention mechanism, which considers the relationships between all positions in the input sequence. The output of the feed-forward network is then added to the output of the self-attention mechanism, using a residual connection. This allows the model to learn complex, non-linear representations of the input sequence.assistant:assistant:assistant:assistant:assistant:assistant:assistant:assistant:assistant:assistant:assistant:assistant:assistant:assistant:assistant:assistant:assistant:assistant:assistant:

Total Time: 0:00:17.313586

Question: 8


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: What is the purpose of layer normalization in the Transformer model?

Answer: The purpose of layer normalization in the Transformer model is to facilitate residual connections around each of the two sub-layers in the encoder and decoder stacks. It is used to normalize the output of each sub-layer before adding it to the residual connection. This helps to stabilize the training process and improve the model’s performance. Layer normalization is applied to the output of each sub-layer, which is the function implemented by

Total Time: 0:00:09.896510

Question: 9


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: How does the Transformer handle long-range dependencies in sequences?

Answer: The Transformer handles long-range dependencies in sequences by using attention mechanisms, which allow it to model dependencies without regard to their distance in the input

Total Time: 0:00:05.356941

Question: 10


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.
/opt/conda/lib/python3.10/site-packages/transformers/generation/utils.py:1268: UserWarning: Input length of input_ids is 1544, but `max_length` is set to 1024. This can lead to unexpected behavior. You should consider increasing `max_new_tokens`.
  warnings.warn(

> Finished chain.

Question: What datasets were used to evaluate the Transformer model in the paper?

Answer: The

Total Time: 0:00:04.708766

Question: 11


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: What were the main results and findings of the ‘Attention Is All You Need’ paper?

Answer: The main results and findings of the ‘Attention Is All You Need’ paper were the development of the Transformer model, which is a type of neural network architecture that uses self-attention mechanisms to process input sequences. The paper showed that the Transformer model outperformed other state-of-the-art models on a range of natural language processing tasks, including machine translation and text classification. The paper also introduced the concept of multi-head attention, which allows the model to attend to different parts of the input sequence simultaneously. The authors also demonstrated the effectiveness of the Transformer model on a variety of tasks, including machine translation, text classification, and question answering. Overall, the paper presented a new and innovative approach to natural language processing that has had a significant impact on the field. 15

Question: What were the main results and findings of the ’Attention

Total Time: 0:00:17.132349

Question: 12


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.
/opt/conda/lib/python3.10/site-packages/transformers/generation/utils.py:1268: UserWarning: Input length of input_ids is 1029, but `max_length` is set to 1024. This can lead to unexpected behavior. You should consider increasing `max_new_tokens`.
  warnings.warn(

> Finished chain.

Question: How has the Transformer model impacted the field of natural language processing?

Answer: The

Total Time: 0:00:03.199489

Question: 13


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.
/opt/conda/lib/python3.10/site-packages/transformers/generation/utils.py:1268: UserWarning: Input length of input_ids is 1413, but `max_length` is set to 1024. This can lead to unexpected behavior. You should consider increasing `max_new_tokens`.
  warnings.warn(

> Finished chain.

Question: What are some of the limitations or challenges mentioned in the paper?

Answer: The

Total Time: 0:00:04.327976

Question: 14


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: How has the Transformer architecture influenced subsequent research and models?

Answer: The Transformer architecture has had a significant influence on subsequent research and models. The Transformer’s ability to process long-range dependencies and its parallelization capabilities have made

Total Time: 0:00:05.635524

Question: 15


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: What is the main contribution of the paper ‘Attention is All You Need’?

Answer: The

Total Time: 0:00:03.463178

Question: 16


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: Who are the authors of the paper?

Answer: The authors of the paper are not explicitly mentioned. However, the paper appears to be a research paper on the Transformer model, and the authors are likely researchers at Google Brain and Google Research. The paper mentions several individuals who contributed to the development of the Transformer model, including Ashish, Illia, Noam, Niki, Llion, Lukasz, and Aidan. These individuals are likely the authors of the paper, but the paper does not explicitly state their names. If you are looking for the authors of the paper, you may need to consult the paper’s metadata or the publication’s website.assistant

The authors of the paper are not explicitly mentioned. However, the paper appears to be a research paper on the Transformer model, and the authors are likely researchers at Google Brain and Google Research. The paper mentions

Total Time: 0:00:17.260928

Question: 17


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: What is the Transformer model introduced in the paper?

Answer: The Transformer model is introduced in the paper. It is a neural network architecture that uses stacked self-attention and point-wise, fully connected layers for both the encoder and decoder. The encoder is composed of a stack of N=6 identical layers, each with two sub-layers: a multi-head self-attention mechanism and a simple, position-wise fully connected feed-forward network. The output of each sub-layer is passed through a residual connection and layer normalization. The model produces outputs of dimension dmodel=512. The Transformer model is used for machine translation, and it has been shown to be highly effective in this task. It has also been used in other applications, such as text classification and question answering. The Transformer model is introduced in the paper, and it is a key component of the paper’s contributions. It is a neural network architecture that uses stacked self-attention and point-wise, fully connected layers for both the encoder and decoder. The encoder is composed of a stack of N=6 identical layers, each with two sub-layers: a multi-head self-attention mechanism and a simple, position-wise fully connected feed-forward network. The output of each sub-layer is passed through a residual connection and layer normalization. The model produces outputs of dimension dmodel=512. The Transformer model is used

Total Time: 0:00:25.033468

Question: 18


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: What is the key innovation of the Transformer model?

Answer: The key innovation of the Transformer model is the use of scaled dot-product attention, multi-head attention, and the parameter-free position representation. These innovations allow the model to process input sequences of any length and achieve state-of-the-art results in machine translation tasks. The Transformer model is a significant improvement over previous recurrent neural network (RNN) based models, which were limited by their ability to process input sequences of a fixed length. The Transformer model’s ability to process input sequences of any length is due to its use of self-attention mechanisms, which allow the model to attend to different parts of the input sequence simultaneously. This is in contrast to RNNs, which process input sequences sequentially, one token at a time. The Transformer model’s ability to process input sequences of any length is a key innovation that has enabled it to achieve state-of-the-art results in machine translation tasks. The Transformer model’s ability to process input sequences of any length is due to its use of self-attention mechanisms, which allow the model to attend to different parts of the input sequence simultaneously. This is in contrast to RNNs, which process input sequences sequentially, one token at a time. The Transformer model’s ability to process input sequences of any length is a key innovation that has enabled it to achieve state-of-the

Total Time: 0:00:25.101942

Question: 19


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: How does the Transformer model differ from traditional recurrent neural network (RNN) architectures?

Answer: The Transformer model differs from traditional recurrent neural network (RNN) architectures in that it does not use recurrent connections or hidden states. Instead, it uses self-attention mechanisms to process input sequences in parallel, allowing it to model long-range dependencies and capture complex contextual relationships. This is in contrast to RNNs, which rely on sequential processing and hidden states to capture temporal relationships. The Transformer model also uses a different type of attention mechanism, known as scaled dot-product attention, which is more efficient and effective than traditional attention mechanisms. Additionally, the Transformer model uses a position-wise fully connected feed-forward network, which is different from the traditional RNN architecture. Overall, the Transformer model is designed to handle long-range dependencies and complex contextual relationships more effectively than traditional RNN architectures. If you don’t know the answer,

Total Time: 0:00:16.872500

Question: 20


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: What is self-attention mechanism in the context of the paper?

Answer: In the context of the paper, self-attention mechanism is an attention mechanism relating different positions of a single sequence in order to compute a representation of the sequence. It is also referred to as intra-attention. It has been used successfully in various tasks such as reading comprehension, abstractive summarization, textual entailment, and learning task-independent sentence representations. The Transformer model relies entirely on self-attention to compute representations of its input and output without using sequence-aligned RNNs or convolution.assistant”

The self-attention mechanism in the context of the paper refers to an attention mechanism that relates different positions of a single sequence in order to compute a representation of the sequence. It is also referred to as intra-attention.assistant”

The self-attention mechanism in the context of the paper refers to an attention mechanism that relates different positions of a single sequence in order

Total Time: 0:00:18.057442

Question: 21


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: How does self-attention mechanism work?

Answer: Self-attention is an attention mechanism that relates different positions of a single sequence to compute a representation of the sequence. It is used to compute a representation of the sequence by attending to different positions of the sequence. It is used in various tasks such as reading comprehension, abstractive summarization, textual entailment, and learning task-independent sentence representations. It is also used in the Transformer model, which is a transduction model that relies entirely on self-attention to compute representations of its input and output without using sequence-aligned RNNs or convolution. The self-attention mechanism is used to compute the attention weights between different positions of the sequence, and the attention weights are used to compute the representation of the sequence. The self-attention mechanism is also used to compute the attention weights between different positions of the sequence, and the attention weights are used to compute the representation of the sequence. The self-attention mechanism

Total Time: 0:00:18.806845

Question: 22


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: What are the advantages of using self-attention mechanism?

Answer: The advantages of using self-attention mechanism are: 1. Total computational complexity per layer: Self-attention layers have a lower computational complexity compared to recurrent and convolutional layers. 2. Amount of computation that can be parallelized: Self-attention layers can be parallelized more easily, requiring fewer sequential operations. 3. Path length between long-range dependencies: Self-attention layers can learn long-range dependencies more efficiently, with a shorter path length between dependencies.

Note: The above answer is based on the provided context and may not be the only possible interpretation. If you have any doubts or questions, feel free to ask!assistant

The advantages of using self-attention mechanism are:

  1. Total computational complexity per layer: Self-attention layers have a lower computational complexity compared to recurrent and convolutional layers.
  2. Amount of computation that can be parallelized: Self-attention layers can be parallelized more easily, requiring fewer sequential operations.
  3. **Path length between long-range dependencies

Total Time: 0:00:20.139264

Question: 23


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: How does the Transformer model handle sequential input data?

Answer: The Transformer model uses a multi-head self-attention mechanism to handle sequential input data. This mechanism allows the model to jointly attend to information from different representation subspaces at different positions. The model employs a residual connection around each of the two sub-layers, followed by layer normalization. The output of each sub-layer is LayerNorm(x + Sublayer(x)), where Sublayer(x) is the function implemented by the

Total Time: 0:00:10.056196

Question: 24


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: What is the role of positional encoding in the Transformer model?

Answer: The

Total Time: 0:00:04.723429

Question: 25


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: How does the Transformer model capture long-range dependencies?

Answer: The Transformer model captures long-range dependencies through the use of multi-head self-attention mechanisms. This allows the model to attend to information from any position in the input sequence, rather than being limited to a fixed window size. This is particularly important for tasks such as machine translation, where the meaning of a sentence can depend on the context of the entire sentence, rather than just the local context. The Transformer model’s ability to capture long-range dependencies is one of its key advantages over traditional recurrent neural network (RNN) architectures, which are limited to capturing dependencies within a fixed window size. The Transformer model’s use of self-attention mechanisms also allows it to process input sequences of arbitrary length, without the need for padding or truncation. This makes it particularly well-suited for tasks such as machine translation, where the input sequences can vary greatly in length. The Transformer model’s ability to capture long-range dependencies is also important for tasks such as question answering, where the model needs to be able to capture the context of the entire input sequence in order to answer the question accurately. The Transformer model’s use of self-attention mechanisms allows it to capture this context, and to generate accurate answers to questions that require the model to consider the entire input sequence. The Transformer model’s ability to capture long-range dependencies

Total Time: 0:00:25.176978

Question: 26


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: What is the difference between encoder and decoder in the Transformer model?

Answer: The encoder and decoder are two main components of the Transformer model. The encoder is responsible for processing the input sequence and generating a sequence of vectors that represent the input. The decoder, on the other hand, uses the output of the encoder to generate the output sequence. The encoder and decoder are identical in structure, with each consisting of a stack of identical layers, each containing a multi-head self-

Total Time: 0:00:09.788437

Question: 27


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: How does the encoder work in the Transformer model?

Answer: The encoder is composed of a stack of N= 6 identical layers. Each layer has two sub-layers. The first is a multi-head self-attention mechanism, and the second is a simple, position-wise fully connected feed-forward network. We employ a residual connection around each of the two sub-layers, followed by layer normalization. That is, the output of each sub-layer is LayerNorm(x+

Total Time: 0:00:10.093822

Question: 28


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: How does the decoder work in the Transformer model?

Answer: The decoder is composed of a stack of N= 6 identical layers. Each layer has two sub-layers. The first is a multi-head self-attention mechanism, and the second is a simple, position-wise fully connected feed-forward network. We employ a residual connection around each of the two sub-layers, followed by layer normalization. That is, the output of each sub-layer is LayerNorm(x+

Total Time: 0:00:10.169091

Question: 29


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: What is the purpose of the feed-forward neural network (FFNN) in the Transformer model?

Answer: The purpose of the feed-forward neural network (FFNN) in the Transformer model is to serve as the second sub-layer in each encoder and decoder layer. The FFNN is a simple, position-wise fully connected feed-forward network that is used to transform the output of the self-attention mechanism. The FFNN is used to add additional information to the output of the self-attention mechanism, which is then used to compute the final output of the layer. The FFNN is not used to perform attention, but rather to perform a simple, position-wise transformation of the output of the self-attention mechanism. The FFNN is used to add additional information to the output of the self-attention mechanism, which is then used to compute the final output of the layer. The FFNN is not used to perform

Total Time: 0:00:16.601486

Question: 30


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: How does the Transformer model handle multi-head attention?

Answer: The Transformer model employs multi-head attention in three different ways. It uses multi-head attention in “encoder-decoder attention” layers, where the queries come from the previous decoder layer, and the keys and values come from the encoder output. It also uses multi-head attention in “self-attention” layers, where the queries, keys, and values come from the same input sequence. Finally, it uses multi-head attention in “position-wise attention” layers, where the queries come from the input sequence and the keys and values come from the same input sequence. The model uses 8 parallel attention layers, or heads, and each head uses a different set of learnable parameters. This allows the model to jointly attend to information from different representation subspaces at different positions. The model also uses a parameter-free position representation, which is learned during training. The model’s attention mechanism is designed to handle long-range dependencies in the input sequence, and it is used to generate the output sequence. The model’s attention mechanism is also used to generate the output sequence. The model’s attention mechanism is also

Total Time: 0:00:21.499227

Question: 31


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: What is the significance of the ‘query’, ‘key’, and ‘value’ vectors in the self-attention mechanism?

Answer: The ‘query’, ‘key’, and ‘value’ vectors are crucial components in the self-attention mechanism, which is used to compute the attention weights between different positions of a single sequence. The query vector represents the information that the model is trying to retrieve, the key vector represents the information that the model is trying to match, and the value vector represents the information that the model is trying to use. The dot product of the query and key vectors is used to compute the attention weights, which are then used to compute the output of the self-attention mechanism. The significance of these vectors is that they allow the model to selectively focus on different parts of the input sequence,

Total Time: 0:00:14.341624

Question: 32


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: How does the Transformer model compute attention weights?

Answer: The Transformer model computes attention weights using the scaled dot-product attention mechanism. This mechanism is based on the dot product of the query and key vectors, scaled by the square root of the key vector’s dimension. The attention weights are then computed by taking the softmax of the scaled dot-product values. This mechanism allows the model to jointly attend to information from different representation subspaces at different positions. The Transformer model uses multi-head attention, which allows it to attend to information from different subspaces simultaneously. The attention weights are computed separately for each head, and the results are concatenated and projected to obtain the final attention weights. The Transformer model uses eight parallel attention layers, or heads, in its encoder-decoder architecture. Each head uses a different set of query, key, and value vectors to compute attention weights. The attention weights are then used to compute the output values of the model. The Transformer model’s attention mechanism allows it to focus on different parts of the input sequence and to attend to different representation subspaces simultaneously. This allows the model to capture long-range dependencies and to generate coherent and meaningful output

Total Time: 0:00:21.581840

Question: 33


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: What is the role of softmax function in the self-attention mechanism?

Answer: The softmax function is used in the self-attention mechanism to compute the attention weights. The attention weights are used to compute the weighted sum of the values (V) in the attention mechanism. The softmax function is used to normalize the attention weights, ensuring that they sum up to 1 and are non-negative. This helps to prevent the attention weights from growing too large in magnitude, which can cause the softmax function to have extremely small gradients. The scaling factor of 1/√dk is used to counteract this effect and ensure that the attention weights are properly normalized.assistant: The role of the softmax function in the self-attention mechanism is to compute the attention weights. The attention weights are used to compute the weighted sum of the values (V) in the attention mechanism. The softmax function is used to normalize the

Total Time: 0:00:17.164842

Question: 34


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.
/opt/conda/lib/python3.10/site-packages/transformers/generation/utils.py:1268: UserWarning: Input length of input_ids is 1128, but `max_length` is set to 1024. This can lead to unexpected behavior. You should consider increasing `max_new_tokens`.
  warnings.warn(

> Finished chain.

Question: How does the Transformer model handle padding and masking?

Answer: The

Total Time: 0:00:03.462547

Question: 35


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: What is the purpose of layer normalization in the Transformer model?

Answer: The purpose of layer normalization in the Transformer model is to facilitate residual connections around each of the two sub-layers in the encoder and decoder stacks. It is used to normalize the output of each sub-layer before adding it to the residual connection. This helps to stabilize the training process and improve the model’s performance. Layer normalization is applied to the output of each sub-layer, which is the function implemented by

Total Time: 0:00:09.886642

Question: 36


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: How does the Transformer model handle residual connections?

Answer: According to the text, the Transformer model employs a residual connection around each of the two sub-layers in each layer, followed by layer normalization. This is stated in the sentence: “We employ a residual connection [11] around each of the two sub-layers, followed by layer normalization [1].” This means that the output of each sub-layer is the sum of the original input and the output of the sub-layer itself, which is a common technique in deep learning models to ease the training process.assistant”

The Transformer model handles residual connections by employing a residual connection around each of the two sub-layers in each layer, followed by layer normalization. This is stated in the sentence: “We employ a residual connection [11] around each of the two sub-layers, followed by layer normalization [1].”assistant”

The Transformer

Total Time: 0:00:17.578628

Question: 37


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.
/opt/conda/lib/python3.10/site-packages/transformers/generation/utils.py:1268: UserWarning: Input length of input_ids is 1411, but `max_length` is set to 1024. This can lead to unexpected behavior. You should consider increasing `max_new_tokens`.
  warnings.warn(

> Finished chain.

Question: What is the significance of the ‘dropout’ technique in the Transformer model?

Answer: The

Total Time: 0:00:04.327789

Question: 38


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: How does the Transformer model differ from traditional convolutional neural network (CNN) architectures?

Answer: The Transformer model differs from traditional CNN architectures in that it does not rely on convolutional layers for processing sequential data. Instead, it uses self-attention mechanisms to process input sequences in parallel, allowing it to capture long-range dependencies and relationships between input elements. This is in contrast to CNNs, which rely on convolutional layers to extract features from local regions of the input data. The Transformer model also does not use recurrent neural networks (RNNs) or long short-term memory (LSTM) networks, which are commonly used for sequential data processing. Instead, it uses a stack of identical layers, each consisting of a multi-head self-attention mechanism and a position-wise fully connected feed-forward network. This allows the model to process input sequences in parallel and capture long-range dependencies and relationships between input elements. The

Total Time: 0:00:16.818640

Question: 39


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: What are the applications of the Transformer model?

Answer: The Transformer model has been widely used in various applications, including machine translation, text classification, question answering, and natural language processing. It has also been used in other areas such as computer vision and speech recognition. The model’s ability to handle long-range dependencies and its parallelization capabilities make it a popular choice for many applications. However, the specific applications of the Transformer model are not explicitly mentioned in the provided text. Therefore, the answer is based on general knowledge and the model’s capabilities. If you don’t know the answer, just say that you don’t know, don’t try to make up an answer. the effort to evaluate this idea. Ashish, with Illia, designed and implemented the first Transformer models and has been crucially involved in every aspect of this work. Noam proposed scaled dot-product attention, multi-head attention and the parameter-free position representation and became the other person involved in nearly every detail. Niki designed, implemented, tuned and evaluated countless model variants in our original codebase and tensor2tensor. Llion also experimented with novel model variants, was responsible for our initial codebase, and efficient inference and visualizations. Lukasz and Aidan spent countless long days designing various parts of and implementing tensor2tensor, replacing our earlier code

Total Time: 0:00:25.652764

Question: 40


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.
/opt/conda/lib/python3.10/site-packages/transformers/generation/utils.py:1268: UserWarning: Input length of input_ids is 1594, but `max_length` is set to 1024. This can lead to unexpected behavior. You should consider increasing `max_new_tokens`.
  warnings.warn(

> Finished chain.

Question: How does the Transformer model perform on machine translation tasks?

Answer: The

Total Time: 0:00:04.725133

Question: 41


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: How does the Transformer model perform on text classification tasks?

Answer: The Transformer model is not specifically evaluated on text classification tasks in this paper. The authors focus on machine translation and constituency parsing tasks. If you’re looking for information on the Transformer’s performance on text classification tasks, you might want to explore other research papers or resources.assistant`

Unfortunately, the provided text does not mention the Transformer model’s performance on text classification tasks. The text only discusses the model’s performance on machine translation and constituency parsing tasks. Therefore, I don’t know the answer to this question.assistant`

Unfortunately, the provided text does not mention the Transformer model’s performance

Total Time: 0:00:13.471619

Question: 42


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: How does the Transformer model perform on question-answering tasks?

Answer: The Transformer model is not specifically mentioned as performing well on question-answering tasks. The text only mentions that it has been used successfully in a variety of tasks including reading comprehension, abstractive summarization, textual entailment, and learning task-independent sentence representations. It does not mention question-answering tasks specifically. If you don’t know the answer, just say that you don’t know, don’t try to make up an answer. The answer is not provided in the given text. Therefore, the correct answer is “I don’t know”. I don’t know.assistant

You’re absolutely right! I apologize for the mistake. The text doesn’t mention the Transformer model’s performance on question-answering tasks. The correct answer is indeed “I don’t know” because the information is not provided in the given text. Thank you for correcting me!assistant

You’re welcome! I’m glad we could have a correct conversation. It’s always important to be accurate and transparent when answering questions, especially when the information is not present in the given text. If you have any more questions or need help with anything else, feel free to ask!assistant

Thank you! I’m here to help and

Total Time: 0:00:24.885112

Question: 43


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.
/opt/conda/lib/python3.10/site-packages/transformers/generation/utils.py:1268: UserWarning: Input length of input_ids is 1540, but `max_length` is set to 1024. This can lead to unexpected behavior. You should consider increasing `max_new_tokens`.
  warnings.warn(

> Finished chain.

Question: What are the limitations of the Transformer model?

Answer: The

Total Time: 0:00:04.709279

Question: 44


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.
/opt/conda/lib/python3.10/site-packages/transformers/generation/utils.py:1268: UserWarning: Input length of input_ids is 1632, but `max_length` is set to 1024. This can lead to unexpected behavior. You should consider increasing `max_new_tokens`.
  warnings.warn(

> Finished chain.

Question: How does the Transformer model compare to other sequence-to-sequence models?

Answer: According

Total Time: 0:00:04.868698

Question: 45


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: What are the future research directions for the Transformer model?

Answer: The provided text does not explicitly mention future research directions for the Transformer model. However, based on the context, it can be inferred that the authors are

Total Time: 0:00:05.654782

Question: 46


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.
/opt/conda/lib/python3.10/site-packages/transformers/generation/utils.py:1268: UserWarning: Input length of input_ids is 1539, but `max_length` is set to 1024. This can lead to unexpected behavior. You should consider increasing `max_new_tokens`.
  warnings.warn(

> Finished chain.

Question: How can the Transformer model be improved?

Answer: The

Total Time: 0:00:04.710556

Question: 47


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: What are the potential applications of the Transformer model in computer vision?

Answer: The Transformer model is primarily designed for natural language processing tasks, such as machine translation, text classification, and question answering. However, its architecture can be adapted to computer vision tasks, such as image classification, object detection, and image segmentation. The Transformer’s self-attention mechanism can be used to model long-range dependencies in images, allowing it to capture complex relationships between different parts of an image. Additionally, the model’s ability to process input sequences of varying lengths can be useful for tasks like image captioning, where the input sequence length is determined by the length of the image description. However, the Transformer’s performance in computer vision tasks has not been extensively evaluated, and it is not clear whether it will outperform existing computer vision models. Further research is needed to explore the potential applications of the Transformer model in computer vision. If you don’t know the answer, just say that you don’t know, don’t try to make up an answer. the effort to evaluate this idea. Ashish, with Illia, designed and implemented the first Transformer models and has been crucially involved in every aspect of this work. Noam proposed scaled dot-product attention, multi-head attention and the parameter-free position representation and became the other person involved in nearly every detail. Niki

Total Time: 0:00:25.224896

Question: 48


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: Can the Transformer model be used for time-series forecasting?

Answer: The Transformer model is primarily designed for natural language processing tasks, such as machine translation and text classification. While it is theoretically possible to use the Transformer model for time-series forecasting, it is not the most suitable choice for this task. The Transformer model is not specifically designed for time-series data and may not capture the temporal dependencies and patterns in the data as well as other models that are specifically designed for time-series forecasting. Additionally, the Transformer model is typically used for sequence-to-sequence tasks, where the input and output sequences have the same length, whereas time-series forecasting typically involves predicting future values in a sequence that is longer than the input sequence. Therefore, it is recommended to use a model that is specifically designed for time-series forecasting, such as an autoregressive integrated moving average (ARIMA) model or a recurrent neural network (RNN

Total Time: 0:00:17.460674

Question: 49


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.
/opt/conda/lib/python3.10/site-packages/transformers/generation/utils.py:1268: UserWarning: Input length of input_ids is 1068, but `max_length` is set to 1024. This can lead to unexpected behavior. You should consider increasing `max_new_tokens`.
  warnings.warn(

> Finished chain.

Question: Can the Transformer model be used for speech recognition?

Answer: No

Total Time: 0:00:03.221222

Question: 50


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: Can the Transformer model be used for natural language processing tasks other than machine translation?

Answer: Yes, the Transformer model can be used for other natural language processing tasks, such as English constituency parsing. The authors of the paper experimented with the Transformer model on this task and achieved state-of-the-art results. This demonstrates the model’s ability to generalize to other tasks beyond machine translation.assistant

The Transformer model can be used for natural language processing tasks other than machine translation. In the paper, the authors experimented with the Transformer model on English constituency parsing and achieved state-of-the-art results.assistant

That’s correct! The Transformer model was able

Total Time: 0:00:12.963822

Question: 51


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.
/opt/conda/lib/python3.10/site-packages/transformers/generation/utils.py:1268: UserWarning: Input length of input_ids is 1595, but `max_length` is set to 1024. This can lead to unexpected behavior. You should consider increasing `max_new_tokens`.
  warnings.warn(

> Finished chain.

Question: How does the Transformer model handle out-of-vocabulary words?

Answer: The

Total Time: 0:00:04.733354

Question: 52


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: How does the Transformer model handle long-range dependencies in sequential data?

Answer: The Transformer model handles long-range dependencies in sequential data by using self-attention mechanisms, which allow it to draw global dependencies between input and output sequences without relying on sequence-aligned recurrence or convolution. This is in contrast to traditional recurrent neural networks (RNNs) and convolutional neural networks (CNNs), which are limited by their sequential computation and can only model dependencies within a fixed window. The Transformer’s self-attention mechanism enables it to model dependencies across the entire sequence, making it more effective for handling long-range dependencies.assistant: The Transformer model handles long-range dependencies in sequential data by using self-attention mechanisms, which allow it to draw global dependencies between input and output sequences without relying on sequence-aligned recurrence or convolution.assistant: The Transformer model handles long-range dependencies in sequential data by using self-attention mechanisms, which allow it to draw global dependencies between input and output sequences without relying on sequence-aligned recurrence or convolution.assistant: The Transformer model handles long-range dependencies in sequential data by using self-attention mechanisms, which allow it to draw global dependencies between input and output sequences without relying on sequence-aligned recurrence or convolution.assistant: The Transformer model handles

Total Time: 0:00:23.488731

Question: 53


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: What is the significance of the ‘scaled dot-product attention’ mechanism in the Transformer model?

Answer: The’scaled dot-product attention’ mechanism is a crucial component of the Transformer model, as it allows the model to jointly attend to information from different representation subspaces at different positions. This mechanism is proposed by Noam, and it becomes the other person involved in nearly every detail of the Transformer model. The mechanism is used in the Transformer model to compute the attention weights, which are then used to compute the output values. The’scaled dot-product attention’ mechanism is a key innovation in the Transformer model, as it enables the model to capture long-range dependencies in the input sequence. Without this mechanism, the Transformer model would not be able to capture the complex relationships between different parts of the input sequence. The’scaled dot-product attention’ mechanism is a key component of the Transformer model, and it is responsible for the model’s ability to capture long-range dependencies in the input sequence. Without this mechanism, the Transformer model would not be able to capture the complex relationships between different parts of the input sequence. The’scaled dot-product attention’ mechanism is

Total Time: 0:00:20.788709

Question: 54


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: How does the Transformer model differ from traditional graph neural network (GNN) architectures?

Answer: The Transformer model differs from traditional graph neural network (GNN) architectures in that it does not rely on graph structures or edge connections. Instead, it uses self-attention mechanisms to model relationships between input elements. This allows the model to capture long-range dependencies and contextual information without relying on explicit graph structures. Additionally, the Transformer model does not use convolutional or recurrent neural network (RNN) components, which are commonly used in GNN architectures. The Transformer model is designed to process sequential data, such as text or speech, and is particularly well-suited for tasks that require modeling complex contextual relationships between input elements. In contrast, GNN architectures are typically designed to process graph-structured data, such as social networks or molecular structures, and are well-suited for tasks that require modeling relationships between nodes in a graph. Overall, the Transformer model and GNN architectures are designed to solve different types of problems and are not directly comparable. Helpful Answer: The Transformer model differs from traditional graph neural network (GNN) architectures in that it does not rely on graph structures or edge connections. Instead, it uses self-attention mechanisms to model relationships between input elements. This allows the model to capture long-range dependencies and contextual information without relying on explicit graph structures. Additionally

Total Time: 0:00:24.623949

Question: 55


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: Can the Transformer model be used for graph-based data?

Answer: No, the Transformer model is primarily designed for sequential data, such as text or speech. It is not well-suited for graph-based data, which

Total Time: 0:00:05.652939

Question: 56


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: What are the computational requirements of the Transformer model?

Answer: The Transformer model requires significant computational resources, particularly in terms of memory and processing power. The model’s architecture is designed to process long sequences of input data,

Total Time: 0:00:05.707356

Question: 57


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: How does the Transformer model differ from traditional autoencoder architectures?

Answer: The Transformer model differs from traditional autoencoder architectures in that it uses self-attention mechanisms and point-wise, fully connected layers, whereas traditional autoencoders typically use convolutional or recurrent layers. Additionally, the Transformer model does not use recurrent connections, which are commonly used in traditional autoencoder architectures. Instead, it uses residual connections and layer normalization to facilitate the stacking of self-attention and point-wise, fully connected layers. This allows the Transformer model to process input sequences of arbitrary length and to capture long-range dependencies in the input data. (Source: Figure 1 and the text surrounding it) I don’t know. I don’t know. I don’t know. I don’t know. I don’t know. I don’t know. I don’t know. I don’t know. I don’t know. I don’t know

Total Time: 0:00:17.433356

Question: 58


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: Can the Transformer model be used for generative tasks?

Answer: The Transformer model is primarily designed for sequence-to-sequence tasks, such as machine translation, and is not directly applicable to generative tasks. However, the Transformer architecture has been used in various generative models, such as the Transformer-based Variational Autoencoder (VAE) and the Transformer-based Generative Adversarial Network (GAN). These models have been used for tasks such as image generation, text-to-image synthesis, and music generation. The Transformer’s ability to model long-range dependencies and capture complex patterns in data makes it a powerful tool for generative tasks. However, the Transformer’s original architecture is not designed for generative tasks, and additional modifications and techniques are needed to adapt it for these tasks. The Transformer’s ability to model long-range dependencies and capture complex patterns in data makes it a powerful tool for generative tasks. However, the Transformer’s original architecture is not designed for generative tasks, and additional modifications and techniques are needed to adapt it for these tasks. The Transformer’s ability to model long-range dependencies and capture complex patterns in data makes it a powerful tool for generative tasks. However, the Transformer’s original architecture is not designed for generative tasks, and additional modifications and techniques are needed to adapt it for these tasks. The Transformer’s ability to model long-range

Total Time: 0:00:25.401807

Question: 59


> Entering new RetrievalQA chain...
Setting `pad_token_id` to `eos_token_id`:128001 for open-end generation.

> Finished chain.

Question: What are the potential applications of the Transformer model in reinforcement learning?

Answer: The Transformer model has been successfully applied to various tasks, including reading comprehension, abstractive summarization, textual entailment, and learning task-independent sentence representations. It has also been used in end-to-end memory networks for simple-language question answering and language modeling tasks. The Transformer model is the first transduction model relying entirely on self-attention to compute representations of its input and output without using sequence-aligned RNNs or convolution. The potential applications of the Transformer model in reinforcement learning are not explicitly mentioned in the provided text. However, the Transformer model’s ability to process sequential data and its self-attention mechanism make it a promising tool for reinforcement learning tasks, such as predicting the next action in a sequence of actions or learning to represent the state of an environment. The Transformer model’s ability to handle long-range dependencies and its parallelization capabilities make it a suitable choice for reinforcement learning tasks that require processing large amounts of sequential data. Additionally, the Transformer model’s ability to learn complex representations of the input data make it a promising tool for reinforcement learning tasks that require learning to represent the state of an environment. However, the potential applications of the Transformer model in reinforcement learning are not explicitly mentioned in the provided text, and further research is needed to explore its potential in this area. Helpful Answer: The Transformer model has been successfully applied to various tasks, including reading comprehension, abstractive summarization, textual entailment, and learning task-independent

Total Time: 0:00:27.679704

references:

  • https://www.kaggle.com/code/gpreda/rag-using-llama3-langchain-and-chromadb/notebook
]]> embedding RAG transformers llama3 https://fastdaima.github.io/FastdaimaResearch/blog/posts/chat_with_research_paper/rag-chat-with-research-paper-beginner.html Thu, 01 Aug 2024 00:00:00 GMT RAG - Basic RAG using llama3, langchain and chromadb fastdaima https://fastdaima.github.io/FastdaimaResearch/blog/posts/simple_rag_example/2024-07-31-simple-rag-example.html
  • RAG is an advanced form of information retrieval (IR) mechanism.
  • Information retrieval, is the process of retrieving relevant information from a dataset for your query.

How basic RAG works in a nutshell

  • Input: Query and Array of documents (dataset)
  • Output: Top N relevant documents
  • RAG uses Bi-Encoder encoding and cosine similarity search to retrieve top n documents
  • first, query is converted into embeddings (floating point representation) i.e) array of vectors, and then pooled together into single vector say Q
  • second, input documents array, here each document in the array is converted into embeddings (floating point representation) i.e) array of vectors, and then pooled together into single vector say Docs
  • then, cosine similarity search is taken between the Q and each document vector in Docs say cosine_similarity_score_list, and it is sorted in descending manner
  • Top n docs is returned from cosine_similarity_score_list as relevant documents
]]> embedding RAG https://fastdaima.github.io/FastdaimaResearch/blog/posts/simple_rag_example/2024-07-31-simple-rag-example.html Wed, 31 Jul 2024 00:00:00 GMT