Commit Graph
100 Commits
Author SHA1 Message Date
Lysandre 5f96ebc0be Style 2020-02-04 18:05:35 -05:00
Lysandre 950c6a4f09 Flaubert PyTorch tests 2020-02-04 18:05:35 -05:00
Lysandre d28b81dc29 RoBERTa Pytorch tests 2020-02-04 18:05:35 -05:00
Lysandre 3bf5417258 Revert erroneous fix 2020-02-04 16:31:07 -05:00
Lysandre 1ebfeb7946 Cast to long when masking tokens 2020-02-04 15:56:16 -05:00
Lysandre 9c67196b83 Update quickstart 2020-02-04 11:11:37 -05:00
Lysandre 90ab15cb7a Remove redundant hidden states 2020-02-04 10:59:32 -05:00
Lysandre 6c1b23554f Sample instead of greedy decoding by default in generate 2020-02-03 17:23:53 -05:00
Lysandre 239dd23f64 [Follow up 213]
Masked indices should have -1 and not -100. Updating documentation + scripts that were forgotten
2020-02-03 16:08:05 -05:00
Lysandre d426b58b9e Patch: v2.4.1 2020-01-31 14:55:33 -05:00
Lysandre 1e82cd8457 Flaubert auto tokenizer + tests
cc @julien-c
2020-01-31 14:16:52 -05:00
Lysandre d18d47be67 run_generation style 2020-01-31 12:05:48 -05:00
Lysandre ff6f1492e8 FlauBERT load in AutoModel
The FlauBERT configuration file inherits from XLMConfig, and is recognized as such when loading from AutoModels as the XLMConfig is checked before the FlaubertConfig.

Changing the order solves this problem, but a test should be added.
2020-01-31 12:05:15 -05:00
Lysandre 7365f01d43 do_sample should be set to True in run_generation.py 2020-01-31 11:49:32 -05:00
Lysandre 0aa40e9569 v2.4.0 documentation 2020-01-31 09:55:34 -05:00
Lysandre 8036ceb7c5 Update commands for pypi test 2020-01-31 09:48:15 -05:00
Lysandre 6664ea943d Release: v2.4.0 2020-01-31 09:40:32 -05:00
Lysandre 01a14ebd8d Add FlauBERT to automodels 2020-01-30 18:40:22 -05:00
Lysandre df27648bd9 Rename test_examples to test_doc_samples 2020-01-30 10:07:22 -05:00
Lysandre 93dccf527b Pretrained models 2020-01-30 10:04:18 -05:00
Lysandre 90787fed81 Style 2020-01-30 10:04:18 -05:00
Lysandre 73306d028b FlauBERT documentation 2020-01-30 10:04:18 -05:00
Lysandre ce2f4227ab Fix failing FlauBERT test 2020-01-30 10:04:18 -05:00
Lysandre e63a81dd25 Style 2020-01-29 16:29:20 -05:00
Lysandre 217349016a Copy object instead of passing the reference 2020-01-29 16:15:39 -05:00
Lysandre c69b082601 Update documentation 2020-01-29 12:06:13 -05:00
Lysandre b5625f131d Style 2020-01-29 11:47:49 -05:00
Lysandre 44a5b4bbe7 Update documentation 2020-01-29 11:47:49 -05:00
Lysandre 335dd5e68a Default save steps 50 to 500 in all scripts 2020-01-28 09:42:11 -05:00
Lysandre ea2600bd5f Absolute definitive HeisenDistilBug solve
cc @julien-c @thomwolf
2020-01-27 21:58:36 -05:00
Lysandre 9ca21c838b Style 2020-01-27 14:49:12 -05:00
Lysandre 875c4ae48f Definitive HeisenDistilBug fix
cc @julien-c @@thomwolf
2020-01-27 12:09:58 -05:00
Lysandre f09f42d4d3 Input Embeddings should be assigned
cc @julien-c
2020-01-27 11:46:00 -05:00
Lysandre babd41e7fa Code quality 2020-01-24 17:06:55 -05:00
Lysandre 974d083c7b Accurate model for configuration 2020-01-24 16:46:03 -05:00
Lysandre 983fef469c AutoModels doc 2020-01-24 16:37:30 -05:00
Lysandre 009fcb0ec1 Configuration utils 2020-01-24 16:37:30 -05:00
Lysandre 908230d261 Pickle CamemBERT tokenizer 2020-01-24 10:08:59 -05:00
Lysandre 24d5ad1dcc Run the examples in slow 2020-01-23 09:38:45 -05:00
Lysandre 9ddf60b694 Tips + whitespaces 2020-01-23 09:38:45 -05:00
Lysandre 0e9899f451 Fixes 2020-01-23 09:38:45 -05:00
Lysandre 48ac24020d TF CTRL 2020-01-23 09:38:45 -05:00
Lysandre 7511f3dd89 PyTorch CTRL + Style 2020-01-23 09:38:45 -05:00
Lysandre 980211a63a XLM-RoBERTa 2020-01-23 09:38:45 -05:00
Lysandre 6bc966793a TF DistilBERT 2020-01-23 09:38:45 -05:00
Lysandre db1a7f27a1 PyTorch DistilBERT 2020-01-23 09:38:45 -05:00
Lysandre b28020f590 TF RoBERTa 2020-01-23 09:38:45 -05:00
Lysandre 3e1bc27e1b Pytorch RoBERTa 2020-01-23 09:38:45 -05:00
Lysandre f44ff574d3 Camembert 2020-01-23 09:38:45 -05:00
Lysandre 264eb23912 TF XLM 2020-01-23 09:38:45 -05:00
Lysandre ccebcae75f PyTorch XLM 2020-01-23 09:38:45 -05:00
Lysandre 92b3cb786d TF XLNet 2020-01-23 09:38:45 -05:00
Lysandre cd656fb21a PyTorch XLNet 2020-01-23 09:38:45 -05:00
Lysandre 83fa8d9fb5 TF Transformer-XL 2020-01-23 09:38:45 -05:00
Lysandre 98edad418e PyTorch Transformer-XL 2020-01-23 09:38:45 -05:00
Lysandre 96d21ad06b TF OpenAI GPT 2020-01-23 09:38:45 -05:00
Lysandre 850795c487 Pytorch GPT 2020-01-23 09:38:45 -05:00
Lysandre 1487b840d3 TF GPT2 2020-01-23 09:38:45 -05:00
Lysandre bd0d3fd76e GPT-2 PyTorch models + better tips for BERT 2020-01-23 09:38:45 -05:00
Lysandre dbeb7fb4e6 BERT TensorFlow 2020-01-23 09:38:45 -05:00
Lysandre cd77c750c5 BERT PyTorch models 2020-01-23 09:38:45 -05:00
Lysandre 3922a2497e TF ALBERT + TF Utilities + Fix warnings 2020-01-23 09:38:45 -05:00
Lysandre 00df3d4de0 ALBERT Modeling + required changes to utilities 2020-01-23 09:38:45 -05:00
Lysandre f81b6c95f2 Flake8 violation 2020-01-23 09:38:45 -05:00
Lysandre 632675ea88 Can test examples spread over multiple blocks 2020-01-23 09:38:45 -05:00
Lysandre eaa6b9afc6 Require Torch when testing examples 2020-01-23 09:38:45 -05:00
Lysandre 9bab9b83d2 Glossary 2020-01-23 09:38:45 -05:00
Lysandre 64abd3e0aa Multi-line examples can be tested + ALBERT patch for CircleCI
All tests should now work fine.
2020-01-23 09:38:45 -05:00
Lysandre 837577256b Automatic testing of examples
The CircleCI test should fail.
2020-01-23 09:38:45 -05:00
Lysandre c7f79815e7 Cleanup unused variables 2020-01-21 11:40:24 -05:00
Lysandre 15579e2d55 [SQuAD v2] Code quality 2020-01-21 11:36:46 -05:00
Lysandre 088fa7b759 Correct segment ID for XLNet single sequence 2020-01-21 11:33:45 -05:00
Lysandre 073219b43f Manage impossible examples SQuAD v2 2020-01-21 11:24:43 -05:00
Lysandre ca6ce3040d Fix style 2020-01-20 10:56:23 -05:00
Lysandre 6e2c28a14a Run SQuAD warning when the doc stride may be too high 2020-01-16 13:59:26 -05:00
Lysandre dfe012ad9d Fix misleading RoBERTa token type ids 2020-01-14 17:47:28 -05:00
Lysandre c024ab98df Improve padding side documentation 2020-01-14 17:44:23 -05:00
Lysandre 9aeb0b9b8a Improve padding side documentation 2020-01-14 17:43:00 -05:00
Lysandre 100e3b6f21 Bias should be resized with the weights
Created a link between the linear layer bias and the model attribute bias. This does not change anything for the user nor for the conversion scripts, but allows the `resize_token_embeddings` method to resize the bias as well as the weights of the decoder.

Added a test.
2020-01-14 13:43:45 -05:00
Lysandre 6c32d8bb95 Size > Dimensionality + Remove final TODOs 2020-01-14 14:09:09 +01:00
Lysandre 760164d63b RoBERTa example 2020-01-14 14:09:09 +01:00
Lysandre 387217bd3e Added example usage 2020-01-14 14:09:09 +01:00
Lysandre 7d1bb7f256 Add missing XLNet and XLM models 2020-01-14 14:09:09 +01:00
Lysandre a1cb100460 Wrap up configurations 2020-01-14 14:09:09 +01:00
Lysandre c11b6fd393 Update links in all configurations 2020-01-14 14:09:09 +01:00
Lysandre Debut 632682726f Updated Configurations 2020-01-14 14:09:09 +01:00
Lysandre 164c794eb3 New SQuAD API for distillation script 2020-01-10 11:42:53 +01:00
Lysandre 801f2ac8c7 Add PRETRAINED_INIT_CONFIGURATION to DistilBERT tokenizer 2020-01-10 11:42:21 +01:00
Lysandre 16ce15ed4b DistilBERT token type ids removed from inputs in run_squad 2020-01-08 13:18:30 +01:00
Lysandre Debut f24232cd1b Fix error with global step in run_squad.py 2020-01-08 11:39:00 +01:00
Lysandre Debut 27c1b656cc Fix error with global step in run_lm_finetuning.py 2020-01-07 16:16:12 +01:00
Lysandre 24df44d9c7 Black version python 3.5 2020-01-07 15:53:42 +01:00
Lysandre Debut 73be60c47b Quotes 2020-01-07 15:34:23 +01:00
Lysandre 6806f8204e fix #2410 2020-01-07 15:20:45 +01:00
Lysandre Debut 1efc208ff3 Complete DataProcessor class 2020-01-06 15:02:25 +01:00
Lysandre 74755c89b9 Example snippet for BertForQuestionAnswering 2020-01-06 14:41:53 +01:00
Lysandre a436574bfd Release: v2.3.0 2019-12-20 16:22:20 -05:00
Lysandre cb6d54bfda Numpy compatibility for sentence piece
convert to int earlier
2019-12-20 15:06:28 -05:00
Lysandre 65c75fc587 Clean special tokens test 2019-12-20 11:34:16 -05:00
Lysandre fb393ad994 Added test for all special tokens 2019-12-20 11:29:58 -05:00