Daniël de Kok 2020-07-03 18:13:29 +02:00 committed by Frederik Rietdijk
parent 1bc8a6eeb3
commit 4855aa62fa

View File

@ -1,36 +1,76 @@
{ buildPythonPackage
, stdenv
, fetchFromGitHub
, sacremoses
, requests
, sentencepiece
, boto3
, tqdm
, filelock
, regex
, requests
, numpy
, pytest
, sacremoses
, sentencepiece
, timeout-decorator
, tokenizers
, tqdm
, pytestCheckHook
}:
buildPythonPackage rec {
pname = "transformers";
version = "2.2.1";
version = "3.0.1";
src = fetchFromGitHub {
owner = "huggingface";
repo = pname;
rev = "v${version}";
sha256 = "1p8p3lhhiyk1xl9gpgq4vbchyz57v3w7hhvsj1r90zs3cckindl8";
sha256 = "1l8l82zi021sq5dnzlbjx3wx0n4yy7k96n3m2fr893y9lfkhhd8z";
};
propagatedBuildInputs = [ numpy sacremoses requests sentencepiece boto3 tqdm regex ];
propagatedBuildInputs = [
boto3
filelock
numpy
regex
requests
sacremoses
sentencepiece
tokenizers
tqdm
];
checkInputs = [ pytest ];
# pretrained tries to download from s3
checkPhase = ''
cd transformers # avoid importing local files
HOME=$TMPDIR pytest -k 'not pretrained_tokenizers'
checkInputs = [
pytestCheckHook
timeout-decorator
];
postPatch = ''
substituteInPlace setup.py \
--replace "tokenizers == 0.8.0-rc4" "tokenizers>=0.8,<0.9"
'';
preCheck = ''
export HOME="$TMPDIR"
cd tests
'';
# Disable tests that require network access.
disabledTests = [
"test_all_tokenizers"
"test_batch_encoding_is_fast"
"test_batch_encoding_pickle"
"test_config_from_model_shortcut"
"test_config_model_type_from_model_identifier"
"test_from_pretrained_use_fast_toggle"
"test_hf_api"
"test_outputs_can_be_shorter"
"test_outputs_not_longer_than_maxlen"
"test_pretokenized_tokenizers"
"test_tokenizer_equivalence_en_de"
"test_tokenizer_from_model_type"
"test_tokenizer_from_model_type"
"test_tokenizer_from_pretrained"
"test_tokenizer_identifier_with_correct_config"
];
meta = with stdenv.lib; {
homepage = "https://github.com/huggingface/transformers";
description = "State-of-the-art Natural Language Processing for TensorFlow 2.0 and PyTorch";