riaz.somc/transformers
1
0
Fork 0
mirror of https://github.com/huggingface/transformers.git synced 2025-07-18 20:18:24 +06:00
Code Issues Actions 23 Packages Projects Releases Wiki Activity
5,522 Commits 875 Branches 227 Tags 2 GiB
d8ca57d2ce
Commit Graph

1246 Commits

Author SHA1 Message Date
Julien Chaumond
42f08e596f [examples] rename run_lm_finetuning to run_language_modeling 2020-02-07 09:15:28 -05:00
Julien Chaumond
4f7bdb0958 [examples] Fix broken markdown 2020-02-07 09:15:28 -05:00
Peter Izsak
6fc3d34abd Fix multi-gpu evaluation in run_glue.py 2020-02-06 16:38:55 -05:00
Julien Chaumond
ada24def22 [run_lm_finetuning] Tweak fix for non-long tensor, close #2728 
see 1ebfeb7946 and #2728

Co-Authored-By: Lysandre Debut <lysandre.debut@reseau.eseo.fr>
 2020-02-05 12:49:18 -05:00
Yuval Pinter
d1ab1fab1b
pass langs parameter to certain XLM models (#2734) 
* pass langs parameter to certain XLM models

Adding an argument that specifies the language the SQuAD dataset is in so language-sensitive XLMs (e.g. `xlm-mlm-tlm-xnli15-1024`) don't default to language `0`.
Allows resolution of issue #1799 .

* fixing from `make style`

* fixing style (again)
 2020-02-04 17:12:42 -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
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
Antonio Carlos Falcão Petri
2ba147ecff Fix typo in examples/utils_ner.py 
"%s-%d".format() -> "{}-{}".format()
 2020-02-01 11:10:57 -05:00
Lysandre
d18d47be67 run_generation style 2020-01-31 12:05:48 -05:00
Lysandre
7365f01d43 do_sample should be set to True in run_generation.py 2020-01-31 11:49:32 -05:00
Jared Nielsen
71a382319f Correct documentation 2020-01-30 18:41:24 -05:00
Hang Le
f0a4fc6cd6 Add Flaubert 2020-01-30 10:04:18 -05:00
Jared Nielsen
adb8c93134 Remove lines causing a KeyError 2020-01-29 14:01:16 -05:00
Lysandre
335dd5e68a Default save steps 50 to 500 in all scripts 2020-01-28 09:42:11 -05:00
Julien Chaumond
6b4c3ee234 [run_lm_finetuning] GPT2 tokenizer doesn't have a pad_token 
ping @lysandrejik
 2020-01-27 20:14:02 -05:00
VictorSanh
1ce3fb5cc7 update correct eval metrics (distilbert & co) 2020-01-24 11:45:22 -05:00
Julien Chaumond
1a8e87be4e Line-by-line text dataset (including padding) 2020-01-21 16:57:38 -05:00
Julien Chaumond
b94cf7faac change order 2020-01-21 16:57:38 -05:00
Julien Chaumond
2eaa8b6e56 Easier to not support this, as it could be confusing 
cc @lysandrejik
 2020-01-21 16:57:38 -05:00
Julien Chaumond
801aaa5508 make style 2020-01-21 16:57:38 -05:00
Julien Chaumond
56d4ba8ddb [run_lm_finetuning] Train from scratch 2020-01-21 16:57:38 -05:00
jiyeon_baek
6d5049a24d Fix typo in examples/run_squad.py 
Rul -> Run
 2020-01-17 11:22:51 -05:00
Lysandre
6e2c28a14a Run SQuAD warning when the doc stride may be too high 2020-01-16 13:59:26 -05:00
thomwolf
258ed2eaa8 adding details in readme 2020-01-16 13:21:30 +01:00
thomwolf
50ee59578d update formating - make flake8 happy 2020-01-16 13:21:30 +01:00
thomwolf
1c9333584a formating 2020-01-16 13:21:30 +01:00
thomwolf
e25b6fe354 updating readme 2020-01-16 13:21:30 +01:00
thomwolf
27c7b99015 adding details in readme - moving file 2020-01-16 13:21:30 +01:00
Nafise Sadat Moosavi
99d4515572 HANS evaluation 2020-01-16 13:21:30 +01:00
Julien Chaumond
83a41d39b3 💄 super 2020-01-15 18:33:50 -05:00
Julien Chaumond
715fa638a7 Merge branch 'master' into from_scratch_training 2020-01-14 18:58:21 +00:00
Julien Chaumond
b803b067bf Config to Model mapping 2020-01-13 20:05:20 +00:00
IWillPull
a3085020ed Added repetition penalty to PPLM example (#2436) 
* Added repetition penalty

* Default PPLM repetition_penalty to neutral

* Minor modifications to comply with reviewer's suggestions. (j -> token_idx)

* Formatted code with `make style`
 2020-01-10 23:00:07 -05:00
VictorSanh
e83d9f1c1d cleaning - change ' to " (black requirements) 2020-01-10 19:34:25 -05:00
VictorSanh
ebba9e929d minor spring cleaning - missing configs + processing 2020-01-10 19:14:58 -05:00
Victor SANH
331065e62d missing import 2020-01-10 11:42:53 +01:00
Victor SANH
414e9e7122 indents test 2020-01-10 11:42:53 +01:00
Victor SANH
3cdb38a7c0 indents 2020-01-10 11:42:53 +01:00
Victor SANH
ebd45980a0 Align with run_squad + fix some errors 2020-01-10 11:42:53 +01:00
Victor SANH
45634f87f8 fix Sampler in distributed training - evaluation 2020-01-10 11:42:53 +01:00
Victor SANH
af1ee9e648 Move torch.nn.utils.clip_grad_norm_ 2020-01-10 11:42:53 +01:00
Lysandre
164c794eb3 New SQuAD API for distillation script 2020-01-10 11:42:53 +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
Oren Amsalem
43114b89ba spelling correction (#2434) 2020-01-07 17:25:25 +01:00
Lysandre Debut
27c1b656cc Fix error with global step in run_lm_finetuning.py 2020-01-07 16:16:12 +01:00
Simone Primarosa
176d3b3079 Add support for Albert and XLMRoberta for the Glue example (#2403) 
* Add support for Albert and XLMRoberta for the Glue example
 2020-01-07 14:55:55 +01:00
alberduris
81d6841b4b GPU text generation: mMoved the encoded_prompt to correct device 2020-01-06 15:11:12 +01:00
alberduris
dd4df80f0b Moved the encoded_prompts to correct device 2020-01-06 15:11:12 +01:00
karajan1001
f01b3e6680 fix #2399 an ImportError in official example (#2400) 
* fix #2399 an ImportError in official example

* style

Co-authored-by: Julien Chaumond <chaumond@gmail.com>
 2020-01-05 12:50:20 -05:00
Julien Chaumond
629b22adcf [run_lm_finetuning] mask_tokens: document types 2020-01-01 12:55:10 -05:00
Thomas Wolf
0412f3d929
Merge pull request #2291 from aaugustin/fix-flake8-F841 
Fix F841 flake8 warning
 2019-12-25 22:37:42 +01:00
Aymeric Augustin
a8d34e534e Remove [--editable] in install instructions. 
Use -e only in docs targeted at contributors.

If a user copy-pastes  command line with [--editable], they will hit
an error. If they don't know the --editable option, we're giving them
a choice to make before they can move forwards, but this isn't a choice
they need to make right now.
 2019-12-24 08:46:08 +01:00
Aymeric Augustin
81422c4e6d Remove unused variables in examples. 2019-12-23 22:29:02 +01:00
Aymeric Augustin
c3783399db Remove redundant requirements with transformers. 2019-12-23 19:17:27 +01:00
Aymeric Augustin
9fc8dcb2a0 Standardize import. 
Every other file uses this pattern.
 2019-12-23 18:45:42 +01:00
Aymeric Augustin
1c62e87b34 Use built-in open(). 
On Python 3, `open is io.open`.
 2019-12-22 18:38:56 +01:00
Aymeric Augustin
d6eaf4e6d2 Update comments mentioning Python 2. 2019-12-22 18:38:56 +01:00
Aymeric Augustin
75a23d24af Remove import fallbacks. 2019-12-22 18:38:56 +01:00
Aymeric Augustin
798b3b3899 Remove sys.version_info[0] == 2 or 3. 2019-12-22 18:38:42 +01:00
Aymeric Augustin
6b2200fc88 Remove u-prefixes. 2019-12-22 17:47:54 +01:00
Aymeric Augustin
c824d15aa1 Remove __future__ imports. 2019-12-22 17:47:54 +01:00
Aymeric Augustin
7e98e211f0 Remove unittest.main() in test modules. 
This construct isn't used anymore these days.

Running python tests/test_foo.py puts the tests/ directory on
PYTHONPATH, which isn't representative of how we run tests.

Use python -m unittest tests/test_foo.py instead.
 2019-12-22 14:42:03 +01:00
Aymeric Augustin
ced0a94204 Switch test files to the standard test_*.py scheme. 2019-12-22 14:15:13 +01:00
Aymeric Augustin
c11b3e2926 Sort imports for optional third-party libraries. 
These libraries aren't always installed in the virtual environment where
isort is running. Declaring them properly avoids mixing these
third-party imports with local imports.
 2019-12-22 11:19:13 +01:00
Aymeric Augustin
939148b050 Fix F401 flake8 warning (x28). 
Do manually what autoflake couldn't manage.
 2019-12-22 10:59:08 +01:00
Aymeric Augustin
783a616999 Fix F401 flake8 warning (x88 / 116). 
This change is mostly autogenerated with:

    $ python -m autoflake --in-place --recursive --remove-all-unused-imports --ignore-init-module-imports examples templates transformers utils hubconf.py setup.py

I made minor changes in the generated diff.
 2019-12-22 10:59:08 +01:00
Aymeric Augustin
80327a13ea Fix F401 flake8 warning (x152 / 268). 
This change is mostly autogenerated with:

    $ python -m autoflake --in-place --recursive examples templates transformers utils hubconf.py setup.py

I made minor changes in the generated diff.
 2019-12-22 10:59:08 +01:00
Aymeric Augustin
fa2ccbc081 Fix E266 flake8 warning (x90). 2019-12-22 10:59:08 +01:00
Aymeric Augustin
2ab78325f0 Fix F821 flake8 warning (x47). 
Ignore warnings related to Python 2, because it's going away soon.
 2019-12-22 10:59:07 +01:00
Aymeric Augustin
631be27078 Fix E722 flake8 warnings (x26). 2019-12-22 10:59:07 +01:00
Aymeric Augustin
b0f7db73cd Fix E741 flake8 warning (x14). 2019-12-22 10:59:07 +01:00
Aymeric Augustin
fd2f17a7a1 Fix E714 flake8 warning (x8). 2019-12-22 10:59:07 +01:00
Aymeric Augustin
5eab3cf6bc Fix W605 flake8 warning (x5). 2019-12-22 10:59:07 +01:00
Aymeric Augustin
7dce8dc7ac Fix E731 flake8 warning (x3). 2019-12-22 10:59:07 +01:00
Aymeric Augustin
357db7098c Fix E712 flake8 warning (x1). 2019-12-22 10:59:07 +01:00
Aymeric Augustin
f9c5317db2 Fix E265 flake8 warning (x1). 2019-12-22 10:59:07 +01:00
Aymeric Augustin
28e608a2c2 Remove trailing whitespace from all Python files. 
Fixes flake8 warning W291 (x224).
 2019-12-22 10:59:07 +01:00
Aymeric Augustin
158e82e061 Sort imports with isort. 
This is the result of:

    $ isort --recursive examples templates transformers utils hubconf.py setup.py
 2019-12-22 10:57:46 +01:00
Aymeric Augustin
fa84ae26d6 Reformat source code with black. 
This is the result of:

    $ black --line-length 119 examples templates transformers utils hubconf.py setup.py

There's a lot of fairly long lines in the project. As a consequence, I'm
picking the longest widely accepted line length, 119 characters.

This is also Thomas' preference, because it allows for explicit variable
names, to make the code easier to understand.
 2019-12-21 17:52:29 +01:00
Thomas Wolf
73f6e9817c
Merge pull request #2115 from suvrat96/add_mmbt_model 
[WIP] Add MMBT Model to Transformers Repo
 2019-12-21 15:26:08 +01:00
thomwolf
344126fe58 move example to mm-imdb folder 2019-12-21 15:06:52 +01:00
Thomas Wolf
5b7fb6a4a1
Merge pull request #2134 from bkkaggle/saving-and-resuming 
closes #1960 Add saving and resuming functionality for remaining examples
 2019-12-21 15:03:53 +01:00
Thomas Wolf
6f68d559ab
Merge pull request #2130 from huggingface/ignored-index-coherence 
[BREAKING CHANGE] Setting all ignored index to the PyTorch standard
 2019-12-21 14:55:40 +01:00
thomwolf
1ab25c49d3 Merge branch 'master' into pr/2115 2019-12-21 14:54:30 +01:00
thomwolf
b03872aae0 fix merge 2019-12-21 14:49:54 +01:00
Thomas Wolf
518ba748e0
Merge branch 'master' into saving-and-resuming 2019-12-21 14:41:39 +01:00
Thomas Wolf
18601c3b6e
Merge pull request #2173 from erenup/master 
run_squad with roberta
 2019-12-21 14:33:16 +01:00
Thomas Wolf
eeb70cdd77
Merge branch 'master' into saving-and-resuming 2019-12-21 14:29:59 +01:00
Thomas Wolf
ed9b84816e
Merge pull request #1840 from huggingface/generation_sampler 
[WIP] Sampling sequence generator for transformers
 2019-12-21 14:27:35 +01:00
thomwolf
cfa0380515 Merge branch 'master' into generation_sampler 2019-12-21 14:12:52 +01:00
thomwolf
300ec3003c fixing run_generation example - using torch.no_grad 2019-12-21 14:02:19 +01:00
thomwolf
1c37746892 fixing run_generation 2019-12-21 13:52:49 +01:00
thomwolf
8a2be93b4e fix merge 2019-12-21 13:31:28 +01:00
Thomas Wolf
562f864038
Merge branch 'master' into fix-xlnet-squad2.0 2019-12-21 12:48:10 +01:00
Thomas Wolf
59941c5d1f
Merge pull request #2189 from stefan-it/xlmr 
Add support for XLM-RoBERTa
 2019-12-20 13:26:38 +01:00
Julien Chaumond
a5a06a851e [doc] Param name consistency 2019-12-19 16:24:20 -05:00
Aidan Kierans
1718fb9e74 Minor/basic text fixes (#2229) 
* Small clarification

Matches line 431 to line 435 for additional clarity and consistency.

* Fixed minor typo

The letter "s" was previously omitted from the word "docstrings".
 2019-12-19 16:23:18 -05:00
Francesco
62c1fc3c1e Removed duplicate XLMConfig, XLMForQuestionAnswering and XLMTokenizer from import statement of run_squad.py script 2019-12-19 09:50:56 -05:00
Ejar
284572efc0 Updated typo on the link 
Updated documentation due to typo
 2019-12-19 09:36:43 -05:00
Stefan Schweter
a26ce4dee1 examples: add XLM-RoBERTa to glue script 2019-12-19 02:23:01 +01:00
thomwolf
3d2096f516 further cleanup 2019-12-18 11:50:54 +01:00
thomwolf
83bc5235cf Merge branch 'master' into pr/2189 2019-12-17 11:47:32 +01:00
Thomas Wolf
f061606277
Merge pull request #2164 from huggingface/cleanup-configs 
[SMALL BREAKING CHANGE] Cleaning up configuration classes - Adding Model Cards
 2019-12-17 09:10:16 +01:00
Lysandre
18a879f475 fix #2180 2019-12-16 16:44:29 -05:00
Lysandre
d803409215 Fix run squad evaluate during training 2019-12-16 16:31:38 -05:00
Stefan Schweter
71b4750517 examples: add support for XLM-RoBERTa to run_ner script 2019-12-16 16:37:27 +01:00
thomwolf
dc667ce1a7 double check cc @LysandreJik 2019-12-14 09:56:27 +01:00
thomwolf
7140363e09 update bertabs 2019-12-14 09:44:53 +01:00
Thomas Wolf
a52d56c8d9
Merge branch 'master' into cleanup-configs 2019-12-14 09:43:07 +01:00
erenup
c7780700f5 Merge branch 'refs/heads/squad_roberta' 
# Conflicts:
#	transformers/data/processors/squad.py
 2019-12-14 08:53:59 +08:00
erenup
8e9526b4b5 add multiple processing 2019-12-14 08:43:58 +08:00
Lysandre
c8ed1c82c8 [SQUAD] Load checkpoint when evaluating without training 2019-12-13 12:13:48 -05:00
Pierric Cistac
5a5c4349e8
Fix summarization to_cpu doc 2019-12-13 10:02:33 -05:00
thomwolf
47f0e3cfb7 cleaning up configuration classes 2019-12-13 14:33:24 +01:00
erenup
9b312f9d41 initial version for roberta squad 2019-12-13 14:51:40 +08:00
LysandreJik
7296f1010b Cleanup squad and add allow train_file and predict_file usage 2019-12-12 13:01:04 -05:00
LysandreJik
3fd71c4431 Update example scripts 2019-12-12 12:08:54 -05:00
Alan deLevie
fbf5455a86 Fix typo in examples/run_glue.py args declaration. 
deay -> decay
 2019-12-12 11:16:19 -05:00
Bilal Khan
6aa919469d Update run_xnli to save optimizer and scheduler states, then resume training from a checkpoint 2019-12-10 19:31:22 -06:00
Bilal Khan
89896fe04f Update run_ner to save optimizer and scheduler states, then resume training from a checkpoint 2019-12-10 19:31:22 -06:00
Bilal Khan
fdc05cd68f Update run_squad to save optimizer and scheduler states, then resume training from a checkpoint 2019-12-10 19:31:22 -06:00
Bilal Khan
854ec5784e Update run_glue to save optimizer and scheduler states, then resume training from a checkpoint 2019-12-10 19:30:36 -06:00
LysandreJik
b72f9d340e Correct index in script 2019-12-10 18:33:17 -05:00
LysandreJik
6a73382706 Complete warning + cleanup 2019-12-10 14:33:24 -05:00
Lysandre
dc4e9e5cb3 DataParallel for SQuAD + fix XLM 2019-12-10 19:21:20 +00:00
Rémi Louf
07bc8efbc3 add greedy decoding and sampling 2019-12-10 17:27:50 +01:00
Rémi Louf
4b82c485de remove misplaced summarization documentation 2019-12-10 09:13:33 -05:00
Thomas Wolf
e57d00ee10
Merge pull request #1984 from huggingface/squad-refactor 
[WIP] Squad refactor
 2019-12-10 11:07:26 +01:00
Suvrat Bhooshan
df3961121f Add MMBT Model to Transformers Repo 2019-12-09 18:36:48 -08:00
Julien Chaumond
1d18930462 Harmonize no_cuda flag with other scripts 2019-12-09 20:37:55 -05:00
Rémi Louf
f7eba09007 clean for release 2019-12-09 20:37:55 -05:00
Rémi Louf
2a64107e44 improve device usage 2019-12-09 20:37:55 -05:00
Rémi Louf
c0707a85d2 add README 2019-12-09 20:37:55 -05:00
Rémi Louf
ade3cdf5ad integrate ROUGE 2019-12-09 20:37:55 -05:00
Rémi Louf
076602bdc4 prevent BERT weights from being downloaded twice 2019-12-09 20:37:55 -05:00
Rémi Louf
a1994a71ee simplified model and configuration 2019-12-09 20:37:55 -05:00
Rémi Louf
3a9a9f7861 default output dir to documents dir 2019-12-09 20:37:55 -05:00
Rémi Louf
693606a75c update the docs 2019-12-09 20:37:55 -05:00
Rémi Louf
2403a66598 give transformers API to BertAbs 2019-12-09 20:37:55 -05:00
Rémi Louf
ba089c780b share pretrained embeddings 2019-12-09 20:37:55 -05:00
Rémi Louf
9660ba1cbd Add beam search 2019-12-09 20:37:55 -05:00
Rémi Louf
1c71ecc880 load the pretrained weights for encoder-decoder 
We currently save the pretrained_weights of the encoder and decoder in
two separate directories `encoder` and `decoder`. However, for the
`from_pretrained` function to operate with automodels we need to
specify the type of model in the path to the weights.

The path to the encoder/decoder weights is handled by the
`PreTrainedEncoderDecoder` class in the `save_pretrained` function. Sice
there is no easy way to infer the type of model that was initialized for
the encoder and decoder we add a parameter `model_type` to the function.
This is not an ideal solution as it is error prone, and the model type
should be carried by the Model classes somehow.

This is a temporary fix that should be changed before merging.
 2019-12-09 20:37:55 -05:00
Rémi Louf
07f4cd73f6 update function to add special tokens 
Since I started my PR the `add_special_token_single_sequence` function
has been deprecated for another; I replaced it with the new function.
 2019-12-09 20:37:55 -05:00
Bilal Khan
79526f82f5 Remove unnecessary epoch variable 2019-12-09 16:24:35 -05:00
Bilal Khan
9626e0458c Add functionality to continue training from last saved global_step 2019-12-09 16:24:35 -05:00
Bilal Khan
2d73591a18 Stop saving current epoch 2019-12-09 16:24:35 -05:00
Bilal Khan
0eb973b0d9 Use saved optimizer and scheduler states if available 2019-12-09 16:24:35 -05:00
Bilal Khan
a03fcf570d Save tokenizer after each epoch to be able to resume training from a checkpoint 2019-12-09 16:24:35 -05:00
Bilal Khan
f71b1bb05a Save optimizer state, scheduler state and current epoch 2019-12-09 16:24:35 -05:00
LysandreJik
2a4ef098d6 Add ALBERT and XLM to SQuAD script 2019-12-09 10:46:47 -05:00
Lysandre Debut
00c4e39581
Merge branch 'master' into squad-refactor 2019-12-09 10:41:15 -05:00
Thomas Wolf
5482822a2b
Merge pull request #2046 from jplu/tf2-ner-example 
Add NER TF2 example.
 2019-12-06 12:12:22 +01:00
LysandreJik
e9217da5ff Cleanup 
Improve global visibility on the run_squad script, remove unused files and fixes related to XLNet.
 2019-12-05 16:01:51 -05:00
LysandreJik
9ecd83dace Patch evaluation for impossible values + cleanup 2019-12-05 14:44:57 -05:00
VictorSanh
35ff345fc9 update requirements 2019-12-05 12:07:04 -05:00
VictorSanh
552c44a9b1 release distilm-bert 2019-12-05 10:14:58 -05:00
Rosanne Liu
ee53de7aac Pr for pplm (#2060) 
* license

* changes

* ok

* Update paper link and commands to run

* pointer to uber repo
 2019-12-05 09:20:07 -05:00
Julien Plu
9200a759d7 Add few tests on the TF optimization file with some info in the documentation. Complete the README. 2019-12-05 12:56:43 +01:00
thomwolf
75a97af6bc fix #1450 - add doc 2019-12-05 11:26:55 +01:00
LysandreJik
f7e4a7cdfa Cleanup 2019-12-04 16:24:15 -05:00
LysandreJik
cca75e7884 Kill the demon spawn 2019-12-04 15:42:29 -05:00
LysandreJik
9ddc3f1a12 Naming update + XLNet/XLM evaluation 2019-12-04 10:37:00 -05:00
thomwolf
5bfcd0485e fix #1991 2019-12-04 14:53:11 +01:00
Julien Plu
ecb923da9c Create a NER example similar to the Pytorch one. It takes the same options, and can be run the same way. 2019-12-04 09:43:15 +01:00
LysandreJik
de276de1c1 Working evaluation 2019-12-03 17:15:51 -05:00
Julien Chaumond
7edb51f3a5 [pplm] split classif head into its own file 2019-12-03 22:07:25 +00:00
VictorSanh
48cbf267c9 Use full dataset for eval (SequentialSampler in Distributed setting) 2019-12-03 11:01:37 -05:00
Julien Chaumond
f434bfc623 [pplm] Update S3 links 
Co-Authored-By: Piero Molino <w4nderlust@gmail.com>
 2019-12-03 10:53:02 -05:00
Ethan Perez
96e83506d1 Always use SequentialSampler during evaluation 
When evaluating, shouldn't we always use the SequentialSampler instead of DistributedSampler? Evaluation only runs on 1 GPU no matter what, so if you use the DistributedSampler with N GPUs, I think you'll only evaluate on 1/N of the evaluation set. That's at least what I'm finding when I run an older/modified version of this repo.
 2019-12-03 10:15:39 -05:00
Julien Chaumond
3b48806f75 [pplm] README: add setup + tweaks 2019-12-03 10:14:02 -05:00
Julien Chaumond
0cb2c90890 readme 
Co-Authored-By: Rosanne Liu <mimosavvy@gmail.com>
 2019-12-03 10:14:02 -05:00
Julien Chaumond
1efb2ae7fc [pplm] move scripts under examples/pplm/ 2019-12-03 10:14:02 -05:00
Piero Molino
a59fdd1627 generate_text_pplm now works with batch_size > 1 2019-12-03 10:14:02 -05:00
w4nderlust
893d0d64fe Changed order of some parameters to be more consistent. Identical results. 2019-12-03 10:14:02 -05:00
w4nderlust
f42816e7fc Added additional check for url and path in discriminator model params 2019-12-03 10:14:02 -05:00
w4nderlust
f10b925015 Imrpovements: model_path renamed pretrained_model, tokenizer loaded from pretrained_model, pretrained_model set to discriminator's when discrim is specified, sample = False by default but cli parameter introduced. To obtain identical samples call the cli with --sample 2019-12-03 10:14:02 -05:00
w4nderlust
75904dae66 Removed global variable device 2019-12-03 10:14:02 -05:00
piero
7fd54b55a3 Added support for generic discriminators 2019-12-03 10:14:02 -05:00
piero
b0eaff36e6 Added a +1 to epoch when saving weights 2019-12-03 10:14:02 -05:00
piero
611961ade7 Added tqdm to preprocessing 2019-12-03 10:14:02 -05:00
piero
afc7dcd94d Now run_pplm works on cpu. Identical output as before (when using gpu). 2019-12-03 10:14:02 -05:00
piero
61399e5afe Cleaned perturb_past. Identical output as before. 2019-12-03 10:14:02 -05:00
piero
ffc2935405 Fix for making unditioned generation work. Identical output as before. 2019-12-03 10:14:02 -05:00
piero
9f693a0c48 Cleaned generate_text_pplm. Identical output as before. 2019-12-03 10:14:02 -05:00
piero
61a12f790d Renamed SmallConst to SMALL_CONST and introduced BIG_CONST. Identical output as before. 2019-12-03 10:14:02 -05:00
piero
ef47b2c03a Removed commented code. Identical output as before. 2019-12-03 10:14:02 -05:00
piero
7ea12db3f5 Removed commented code. Identical output as before. 2019-12-03 10:14:02 -05:00
piero
08c6e456a3 Cleaned full_text_generation. Identical output as before. 2019-12-03 10:14:02 -05:00
piero
6c9c131780 More cleanup for run_model. Identical output as before. 2019-12-03 10:14:02 -05:00
piero
7ffe47c888 Improved device specification 2019-12-03 10:14:02 -05:00
piero
4f2164e40e First cleanup step, changing function names and passing parameters all the way through without using args. Identical output as before. 2019-12-03 10:14:02 -05:00
piero
821de121e8 Minor changes 2019-12-03 10:14:02 -05:00
w4nderlust
7469d03b1c Fixed minor bug when running training on cuda 2019-12-03 10:14:02 -05:00
piero
0b51fba20b Added script for training a discriminator for pplm to use 2019-12-03 10:14:02 -05:00
Piero Molino
34a83faabe Let's make PPLM great again 2019-12-03 10:14:02 -05:00
Julien Chaumond
d5faa74cd6 tokenizer white space: revert to previous behavior 2019-12-03 10:14:02 -05:00
Julien Chaumond
0b77d66a6d rm extraneous import 2019-12-03 10:14:02 -05:00
Rosanne Liu
83b1e6ac9e fix the loss backward issue 
(cherry picked from commit 566468cc984c6ec7e10dfc62b5b4191781a99cd2)
 2019-12-03 10:14:02 -05:00
Julien Chaumond
572c24cfa2 PPLM (squashed) 
Co-authored-by: piero <piero@uber.com>
Co-authored-by: Rosanne Liu <mimosavvy@gmail.com>
 2019-12-03 10:14:02 -05:00
Thomas Wolf
f19a78a634
Merge pull request #1903 from valohai/master 
Valohai integration
 2019-12-03 16:13:01 +01:00
maxvidal
b0ee7c7df3 Added Camembert to available models 2019-11-29 14:17:02 -05:00
Juha Kiili
41aa0e8003 Refactor logs and fix loss bug 2019-11-29 15:33:25 +02:00
Lysandre
bd41e8292a Cleanup & Evaluation now works 2019-11-28 16:03:56 -05:00
Stefan Schweter
8c276b9c92
Merge branch 'master' into distilbert-german 2019-11-27 18:11:49 +01:00
VictorSanh
d5478b939d add distilbert + update run_xnli wrt run_glue 2019-11-27 11:07:22 -05:00
VictorSanh
73fe2e7385 remove fstrings 2019-11-27 11:07:22 -05:00
VictorSanh
3e7656f7ac update readme 2019-11-27 11:07:22 -05:00
VictorSanh
abd397e954 uniformize w/ the cache_dir update 2019-11-27 11:07:22 -05:00
VictorSanh
d5910b312f move xnli processor (and utils) to transformers/data/processors 2019-11-27 11:07:22 -05:00
VictorSanh
289cf4d2b7 change default for XNLI: dev --> test 2019-11-27 11:07:22 -05:00
VictorSanh
84a0b522cf mbert reproducibility results 2019-11-27 11:07:22 -05:00
VictorSanh
c4336ecbbd xnli - output_mode consistency 2019-11-27 11:07:22 -05:00
VictorSanh
d52e98ff9a add xnli examples/README.md 2019-11-27 11:07:22 -05:00
VictorSanh
71f71ddb3e run_xnli + utils_xnli 2019-11-27 11:07:22 -05:00
Julien Chaumond
b5d884d25c Uniformize #1952 2019-11-27 11:05:55 -05:00
Lysandre
4374eaea78 ALBERT for SQuAD 2019-11-26 13:08:12 -05:00
Lysandre
c110c41fdb Run GLUE and remove LAMB 2019-11-26 13:08:12 -05:00
manansanghi
5d3b8daad2 Minor bug fixes on run_ner.py 2019-11-25 16:48:03 -05:00
İbrahim Ethem Demirci
aa92a184d2 resize model when special tokenizer present 2019-11-25 15:06:32 -05:00
Lysandre
7485caefb0 fix #1894 2019-11-25 09:33:39 -05:00
Julien Chaumond
176cd1ce1b [doc] homogenize instructions slightly 2019-11-23 11:18:54 -05:00
Lysandre
c3ba645237 Works for XLNet 2019-11-22 16:27:37 -05:00
Lysandre
72e506b22e wip 2019-11-22 16:26:00 -05:00
Rémi Louf
26db31e0c0 update the documentation 2019-11-21 14:41:19 -05:00
Juha Kiili
2cf3447e0a Glue: log in Valohai-compatible JSON format too 2019-11-21 12:35:25 +02:00
Thomas Wolf
0cdfcca24b
Merge pull request #1860 from stefan-it/camembert-for-token-classification 
[WIP] Add support for CamembertForTokenClassification
 2019-11-21 10:56:07 +01:00
Jin Young Sohn
e70cdf083d Cleanup TPU bits from run_glue.py 
TPU runner is currently implemented in:
https://github.com/pytorch-tpu/transformers/blob/tpu/examples/run_glue_tpu.py.

We plan to upstream this directly into `huggingface/transformers`
(either `master` or `tpu`) branch once it's been more thoroughly tested.
 2019-11-20 17:54:34 -05:00
Lysandre
454455c695 fix #1879 2019-11-20 09:42:48 -05:00
Stefan Schweter
e7cf2ccd15 distillation: add German distilbert model 2019-11-19 19:55:19 +01:00
Kazutoshi Shinoda
f3386d9383 typo "deay" -> "decay" 2019-11-18 11:50:06 -05:00
Stefan Schweter
56c84863a1 camembert: add support for CamemBERT in run_ner example 2019-11-18 17:06:57 +01:00
Julien Chaumond
26858f27cb [camembert] Upload to s3 + rename script 2019-11-16 00:11:07 -05:00
Louis MARTIN
3e20c2e871 Update demo_camembert.py with new classes 2019-11-16 00:11:07 -05:00
Louis MARTIN
f12e4d8da7 Move demo_camembert.py to examples/contrib 2019-11-16 00:11:07 -05:00
Louis MARTIN
6e72fd094c Add demo_camembert.py 2019-11-16 00:11:07 -05:00
Xu Hongshen
ca99a2d500 Update example readme 2019-11-15 14:55:26 +08:00
Xu Hongshen
7da3ef24cd add is_impossible tensor to model inputs during fine-tuning xlnet on squad2.0 2019-11-15 14:18:53 +08:00
Thomas Wolf
74ce8de7d8
Merge pull request #1792 from stefan-it/distilbert-for-token-classification 
DistilBERT for token classification
 2019-11-14 22:47:53 +01:00
Thomas Wolf
05db5bc1af
added small comparison between BERT, RoBERTa and DistilBERT 2019-11-14 22:40:22 +01:00
Thomas Wolf
9629e2c676
Merge pull request #1804 from ronakice/master 
fix multi-gpu eval in torch examples
 2019-11-14 22:24:05 +01:00
Thomas Wolf
df99f8c5a1
Merge pull request #1832 from huggingface/memory-leak-schedulers 
replace LambdaLR scheduler wrappers by function
 2019-11-14 22:10:31 +01:00
Rémi Louf
2276bf69b7 update the examples, docs and template 2019-11-14 20:38:02 +01:00
Lysandre
d7929899da Specify checkpoint in saved file for run_lm_finetuning.py 2019-11-14 10:49:00 -05:00
ronakice
2e31176557 fix multi-gpu eval 2019-11-12 05:55:11 -05:00
Stefan Schweter
2b07b9e5ee examples: add DistilBert support for NER fine-tuning 2019-11-11 16:19:34 +01:00
Adrian Bauer
7a9aae1044 Fix run_bertology.py 
Make imports and args.overwrite_cache match run_glue.py
 2019-11-08 16:28:40 -05:00
Julien Chaumond
f88c104d8f [run_tf_glue] Add comment for context 2019-11-05 19:56:43 -05:00
Julien Chaumond
30968d70af misc doc 2019-11-05 19:06:12 -05:00
Thomas Wolf
e99071f105
Merge pull request #1734 from orena1/patch-1 
add progress bar to convert_examples_to_features
 2019-11-05 11:34:20 +01:00
Thomas Wolf
ba973342e3
Merge pull request #1553 from WilliamTambellini/timeSquadInference 
Add speed log to examples/run_squad.py
 2019-11-05 11:13:12 +01:00
Thomas Wolf
237fad339c
Merge pull request #1709 from oneraghavan/master 
Fixing mode in evaluate during training
 2019-11-05 10:55:33 +01:00
Oren Amsalem
d7906165a3
add progress bar for convert_examples_to_features 
It takes considerate amount of time (~10 min) to parse the examples to features, it is good to have a progress-bar to track this
 2019-11-05 10:34:27 +02:00
thomwolf
89d6272898 Fix #1623 2019-11-04 16:21:12 +01:00
Thomas Wolf
9a3b173cd3
Merge branch 'master' into master 2019-11-04 11:41:26 +01:00
thomwolf
ad90868627 Update example readme 2019-11-04 11:27:22 +01:00
Raghavan
e5b1048bae
Fixing mode in evaluate during training 2019-11-03 16:14:46 +05:30
Lysandre
1a2b40cb53 run_tf_glue MRPC evaluation only for MRPC 2019-10-31 18:00:51 -04:00
Timothy Liu
be36cf92fb Added mixed precision support to benchmarks.py 2019-10-31 17:24:37 -04:00
Julien Chaumond
f96ce1c241 [run_generation] Fix generation with batch_size>1 2019-10-31 18:27:11 +00:00
Julien Chaumond
3c1b6f594e
Merge branch 'master' into fix_top_k_top_p_filtering 2019-10-31 13:53:51 -04:00
Victor SANH
fa735208c9
update readme - fix example command distil* 2019-10-30 14:27:28 -04:00
Thomas Wolf
c7058d8224
Merge pull request #1608 from focox/master 
Error raised by "tmp_eval_loss += tmp_eval_loss.item()" when using multi-gpu
 2019-10-30 17:14:07 +01:00
Thomas Wolf
04c69db399
Merge pull request #1628 from huggingface/tfglue 
run_tf_glue works with all tasks
 2019-10-30 17:04:03 +01:00
Thomas Wolf
3df4367244
Merge pull request #1601 from huggingface/clean-roberta 
Clean roberta model & all tokenizers now add special tokens by default (breaking change)
 2019-10-30 17:00:40 +01:00
Thomas Wolf
36174696cc
Merge branch 'master' into clean-roberta 2019-10-30 16:51:06 +01:00
Thomas Wolf
228cdd6a6e
Merge branch 'master' into conditional-generation 2019-10-30 16:40:35 +01:00
Rémi Louf
070507df1f format utils for summarization 2019-10-30 11:24:12 +01:00
Rémi Louf
da10de8466 fix bug with padding mask + add corresponding test 2019-10-30 11:19:58 +01:00
Rémi Louf
3b0d2fa30e rename seq2seq to encoder_decoder 2019-10-30 10:54:46 +01:00
Rémi Louf
9c1bdb5b61 revert renaming of lm_labels to ltr_lm_labels 2019-10-30 10:43:13 +01:00
Rémi Louf
098a89f312 update docstrings; rename lm_labels to more explicit ltr_lm_labels 2019-10-29 20:08:03 +01:00
Rémi Louf
dfce409691 resolve PR comments 2019-10-29 17:10:20 +01:00
altsoph
079bfb32fb Evaluation fixed. 2019-10-28 10:18:58 -04:00
altsoph
438f2730a0 Evaluation code fixed. 2019-10-28 10:18:58 -04:00
Rémi Louf
4c3ac4a7d8 here's one big commit 2019-10-28 10:49:50 +01:00
Rémi Louf
932543f77e fix test of truncation function 2019-10-28 10:49:49 +01:00
Rémi Louf
a67413ccc8 extend works in-place 2019-10-28 10:49:49 +01:00
Rémi Louf
b915ba9dfe pad sequence with 0, mask with -1 2019-10-28 10:49:49 +01:00
Lysandre
bab6ad01aa run_tf_glue works with all tasks 2019-10-24 21:41:45 +00:00
Matt Maybeno
ae1d03fc51 Add roberta to doc 2019-10-24 14:32:48 -04:00
Matt Maybeno
4e5f88b74f Add Roberta to run_ner.py 2019-10-24 14:32:48 -04:00
VictorSanh
5b6cafb11b [release] fix table weirdness 2019-10-23 10:35:16 -04:00
VictorSanh
8ad5c591cd [RELEASE] DistilRoBERTa 2019-10-23 10:29:47 -04:00
focox@qq.com
bd847ce7d7 fixed the bug raised by "tmp_eval_loss += tmp_eval_loss.item()" when parallelly using multi-gpu. 2019-10-23 20:27:13 +08:00
Julien Chaumond
ef1b8b2ae5 [CTRL] warn if generation prompt does not start with a control code 
see also https://github.com/salesforce/ctrl/pull/50
 2019-10-22 21:30:32 +00:00
Lysandre
7d709e55ed Remove 2019-10-22 14:12:33 -04:00
Lysandre
1cfd974868 Option to benchmark only one of the two libraries 2019-10-22 13:32:23 -04:00
Pasquale Minervini
abd7110e21 gradient norm clipping should be done right before calling the optimiser - fixing run_glue and run_ner as well 2019-10-21 19:56:52 +01:00
Pasquale Minervini
3775550c4b gradient norm clipping should be done right before calling the optimiser 2019-10-20 22:33:56 +01:00
LysandreJik
7dd29ed2f1 Benchmarks example script 2019-10-18 10:53:04 -04:00
William Tambellini
0919389d9a Add speed log to examples/run_squad.py 
Add a speed estimate log (time per example)
for evaluation to examples/run_squad.py
 2019-10-17 14:41:04 -07:00
leo-du
ecd15667f3 fix repetition penalty 2019-10-17 14:47:14 -04:00
thomwolf
8cd56e3036 fix data processing in script 2019-10-17 16:33:26 +02:00
Rémi Louf
578d23e061 add training pipeline (formatting temporary) 2019-10-17 14:02:27 +02:00
Rémi Louf
47a06d88a0 use two different tokenizers for storyand summary 2019-10-17 13:04:26 +02:00
Rémi Louf
bfb9b540d4 add Model2Model to __init__ 2019-10-17 12:59:51 +02:00
Rémi Louf
c1bc709c35 correct the truncation and padding of dataset 2019-10-17 10:41:53 +02:00
Rémi Louf
e4e0ee14bd add separator between data import and train 2019-10-16 20:05:32 +02:00
Rémi Louf
0d81fc853e specify in readme that both datasets are required 2019-10-15 15:26:33 +02:00
Rémi Louf
1aec940587 test the full story processing 2019-10-15 15:18:07 +02:00
Rémi Louf
22e1af6859 truncation function is fully tested 2019-10-15 14:43:50 +02:00
Rémi Louf
260ac7d9a8 wip commit, switching computers 2019-10-15 12:24:35 +02:00
thomwolf
be916cb3fb Merge branch 'master' of https://github.com/huggingface/transformers 2019-10-15 10:37:13 +02:00
thomwolf
5875aaf762 install tensorboard 2019-10-15 10:36:46 +02:00
Thomas Wolf
40f14ff545
Merge pull request #1513 from slayton58/amp_fp16_einsum 
Force einsum to run in fp16
 2019-10-15 10:25:00 +02:00
Thomas Wolf
d147671c6c
Merge pull request #1508 from tlkh/master 
Added performance enhancements (XLA, AMP) to examples
 2019-10-15 09:57:18 +02:00
thomwolf
2c1d5564ad add readme information 2019-10-15 09:56:52 +02:00
thomwolf
c55badcee0 Add NER finetuning details by @stefan-it in example readme 2019-10-15 09:33:52 +02:00
Julien Chaumond
788e632622 [ner] Honor args.overwrite_cache 2019-10-15 09:17:31 +02:00
thomwolf
0f9ebb0b43 add seqeval as requirement for examples 2019-10-15 09:17:31 +02:00
thomwolf
66adb71734 update to transformers 2019-10-15 09:17:31 +02:00
Marianne Stecklina
5ff9cd158a Add option to predict on test set 2019-10-15 09:17:31 +02:00
Marianne Stecklina
7f5367e0b1 Add cli argument for configuring labels 2019-10-15 09:17:31 +02:00
Marianne Stecklina
e1d4179b64 Make file reading more robust 2019-10-15 09:17:31 +02:00
Marianne Stecklina
383ef96747 Implement fine-tuning BERT on CoNLL-2003 named entity recognition task 2019-10-15 09:17:31 +02:00
Marianne Stecklina
5adb39e757 Add option to predict on test set 2019-10-15 09:14:53 +02:00
Marianne Stecklina
99b189df6d Add cli argument for configuring labels 2019-10-15 09:14:53 +02:00
Marianne Stecklina
3e9420add1 Make file reading more robust 2019-10-15 09:14:53 +02:00
Marianne Stecklina
cde42c4354 Implement fine-tuning BERT on CoNLL-2003 named entity recognition task 2019-10-15 09:14:53 +02:00
hlums
74c5035808 Fix token order in xlnet preprocessing. 2019-10-14 21:27:11 +00:00
Rémi Louf
fe25eefc15 add instructions to fetch the dataset 2019-10-14 20:45:39 +02:00
Rémi Louf
412793275d delegate the padding with special tokens to the tokenizer 2019-10-14 20:45:16 +02:00
Rémi Louf
447fffb21f process the raw CNN/Daily Mail dataset 
the data provided by Li Dong et al. were already tokenized, which means
that they are not compatible with  all the models in the library. We
thus process the raw data directly and tokenize them using the models'
tokenizers.
 2019-10-14 18:12:20 +02:00
Simon Layton
4e6a55751a Force einsum to fp16 2019-10-14 11:12:41 -04:00
Rémi Louf
67d10960ae load and prepare CNN/Daily Mail data 
We write a function to load an preprocess the CNN/Daily Mail dataset as
provided by Li Dong et al. The issue is that this dataset has already
been tokenized by the authors, so we actually need to find the original,
plain-text dataset if we want to apply it to all models.
 2019-10-14 14:11:20 +02:00
Timothy Liu
376e65a674 Added automatic mixed precision and XLA options to run_tf_glue.py 2019-10-13 13:19:06 +00:00
Timothy Liu
86f23a1944 Minor enhancements to run_tf_glue.py 2019-10-13 10:21:35 +00:00
VictorSanh
d844db4005 Add citation bibtex 2019-10-11 16:55:42 -04:00
Rémi Louf
b3261e7ace read parameters from CLI, load model & tokenizer 2019-10-11 18:40:38 +02:00
Rémi Louf
d889e0b71b add base for seq2seq finetuning 2019-10-11 17:36:12 +02:00
Thomas Wolf
4428aefc63
Merge pull request #1488 from huggingface/pytorch-tpu 
GLUE on TPU
 2019-10-11 16:33:00 +02:00
Luran He
f382a8decd convert int to str before adding to a str 2019-10-10 19:20:39 -04:00
Lysandre
639f4b7190 Don't save/load when on TPU 2019-10-10 19:17:25 +00:00
Lysandre
d4e7934ac3 GLUE on TPU 2019-10-10 19:03:06 +00:00
Rémi Louf
1e68c28670 add test for initialization of Bert2Rnd 2019-10-10 18:07:11 +02:00
Thomas Wolf
6596e3d566
Merge pull request #1454 from bkkaggle/pytorch-built-in-tensorboard 
Change tensorboard imports to use built-in tensorboard if available
 2019-10-10 11:56:55 +02:00
thomwolf
177a721205 move back to simple space spliting 2019-10-10 11:45:47 +02:00
thomwolf
a5997dd81a better error messages 2019-10-10 11:31:01 +02:00
Lysandre Debut
2431fea98a
Merge pull request #1383 from keskarnitish/master 
Adding CTRL
 2019-10-09 11:31:05 -04:00
thomwolf
d9e60f4f0d Merge branch 'master' into pr/1383 2019-10-09 17:25:08 +02:00
Lysandre Debut
e84470ef81
Merge pull request #1384 from huggingface/encoding-qol 
Quality of life enhancements in encoding + patch MLM masking
 2019-10-09 11:18:24 -04:00
jinoobaek-qz
69629c4f0f Improve naming and only do regex when necessary 2019-10-09 08:48:40 -04:00
jinoobaek-qz
bf34a252b8 Golden path 2019-10-09 08:48:40 -04:00
jinoobaek-qz
528d3f327b Improve readability and improve make less assumptions about checkpoint format 2019-10-09 08:48:40 -04:00
jinoobaek-qz
56301bd9e8 Extract method 2019-10-09 08:48:40 -04:00
jinoobaek-qz
d6c5469712 Delete older checkpoint after saving new checkpoint 2019-10-09 08:48:40 -04:00
jinoobaek-qz
54a31f50fb Add save_total_limit 2019-10-09 08:48:40 -04:00
Thomas Wolf
439fac723a
Merge pull request #1409 from brian41005/master 
Evaluation result.txt path changing #1286
 2019-10-09 03:14:34 +02:00
Bilal Khan
5ce8d29abe Change tensorboard imports to use built-in tensorboard if available 2019-10-08 16:29:43 -05:00
VictorSanh
7ce83b4931 update weights for distilgpt2 2019-10-07 12:30:27 -04:00
LysandreJik
f3e0218fbb Correct device assignment in run_generation 2019-10-05 21:05:16 -04:00
thomwolf
78ef1a9930 fixes 2019-10-04 17:59:44 -04:00
thomwolf
6c1d0bc066 update encode_plus - add truncation strategies 2019-10-04 17:38:38 -04:00
VictorSanh
0820bb0555 unecessary carriage return 2019-10-04 17:23:15 -04:00
VictorSanh
f5891c3821 run_squad --> run_squad_w_distillation 2019-10-04 17:23:15 -04:00
VictorSanh
764a7923ec add distillation+finetuning option in run_squad 2019-10-04 17:23:15 -04:00
thomwolf
92c0f2fb90 Merge remote-tracking branch 'origin/julien_multiple-choice' into encoding-qol 2019-10-04 15:48:06 -04:00
Julien Chaumond
9e136ff57c Honor args.overwrite_cache (h/t @erenup) 2019-10-04 15:00:56 -04:00
keskarnitish
dbed1c5d94 Adding CTRL (squashed commit) 
adding conversion script

adding first draft of modeling & tokenization

adding placeholder for test files

bunch of changes

registering the tokenizer/model/etc

tests

change link; something is very VERY wrong here

weird end-of-word thingy going on

i think the tokenization works now ; wrote the unit tests

overall structure works;load w next

the monster is alive!

works after some cleanup as well

adding emacs autosave to gitignore

currently only supporting the 48 layer one; seems to infer fine on my macbook

cleanup

fixing some documentation

fixing some documentation

tests passing?

now works on CUDA also

adding greedy?

adding greedy sampling

works well
 2019-10-03 22:29:03 -07:00
Lysandre Debut
d3f24dfad7
Merge branch 'master' into master 2019-10-03 22:43:09 +00:00
LysandreJik
ecc4f1bdfa XLM use_lang_embedding flag in run_generation 2019-10-03 17:42:16 -04:00
LysandreJik
c2c2ca0fdb Added XLM to run_generation, with prompt language selection. 2019-10-03 17:18:48 -04:00
LysandreJik
aebd83230f Update naming + remove f string in run_lm_finetuning example 2019-10-03 11:31:36 -04:00
LysandreJik
5ed50a93fb LM finetuning won't mask special tokens anymore 2019-10-03 11:31:36 -04:00
Brian Ma
7af0777910 Update run_glue.py 
add DistilBert model shortcut into ALL_MODELS
 2019-10-03 15:31:11 +00:00
VictorSanh
5f07d8f11a prepare release 2019-10-03 10:27:11 -04:00
VictorSanh
35071007cb incoming release 🔥 update links to arxiv preprint 2019-10-03 10:27:11 -04:00
VictorSanh
2a91f6071f upddate README - TODO updadte link to paper 2019-10-03 10:27:11 -04:00
VictorSanh
c51e533a5f update train.py 2019-10-03 10:27:11 -04:00
VictorSanh
a76c3f9cb0 update requirements 2019-10-03 10:27:11 -04:00
VictorSanh
bb9c5ead54 update distiller 2019-10-03 10:27:11 -04:00
VictorSanh
a12ab0a8db update binarized_data 2019-10-03 10:27:11 -04:00
VictorSanh
4d6dfbd376 update extract 2019-10-03 10:27:11 -04:00
VictorSanh
23edebc079 update extract_distilbert 2019-10-03 10:27:11 -04:00
VictorSanh
cbfcfce205 update token_counts 2019-10-03 10:27:11 -04:00
VictorSanh
19e4ebbe3f grouped_batch_sampler 2019-10-03 10:27:11 -04:00
VictorSanh
594202a934 lm_seqs_dataset 2019-10-03 10:27:11 -04:00
VictorSanh
38084507c4 add distillation_configs 2019-10-03 10:27:11 -04:00
Brian Ma
2195c0d5f9 Evaluation result.txt path changing #1286 2019-10-03 12:49:12 +08:00
Thomas Wolf
963529e29b
Merge pull request #1288 from echan00/master 
Typo with LM Fine tuning script
 2019-10-01 18:46:07 -04:00
thomwolf
f7978f70ec use format instead of f-strings 2019-10-01 18:45:38 -04:00
Julien Chaumond
b350662955 overflowing_tokens do not really make sense here, let's just return a number 
Co-Authored-By: Lysandre Debut <lysandre.debut@reseau.eseo.fr>
 2019-09-30 16:37:09 -04:00
Julien Chaumond
f5bcde0b2f [multiple-choice] Simplify and use tokenizer.encode_plus 2019-09-30 16:04:55 -04:00
Denny
9478590630
Update run_lm_finetuning.py 
The previous method, just as phrased, did not exist in the class.
 2019-09-27 15:18:42 -03:00
Thomas Wolf
d83d295763
Merge pull request #1337 from mgrankin/fastdataset 
faster dataset building
 2019-09-27 10:35:12 +02:00
thomwolf
da2e47ad15 clean up a little run_tf_glue 2019-09-27 09:41:15 +02:00
thomwolf
528c288fa9 clean up run_tf_glue 2019-09-27 09:40:29 +02:00
VictorSanh
702f589848 fix input in run_glue for distilbert 2019-09-27 00:20:14 -04:00
mgrankin
f71a4577b8 faster dataset building 2019-09-26 16:53:13 +03:00
thomwolf
481d9c4fb5 Merge branch 'master' into tf2 2019-09-26 12:02:54 +02:00
thomwolf
31c23bd5ee [BIG] pytorch-transformers => transformers 2019-09-26 10:15:53 +02:00
thomwolf
5705333441 add initialization for everybody 2019-09-26 10:06:20 +02:00
thomwolf
7c9f8f93f9 fix tests 2019-09-26 01:59:53 +02:00
thomwolf
d6dde438ea add batch dimension in encode 2019-09-26 01:45:55 +02:00
thomwolf
4a21c4d88d add warning if neither pt nor tf are found 2019-09-26 01:30:06 +02:00
thomwolf
3b7fb48c3b fix loading from tf/pt 2019-09-25 17:46:16 +02:00
thomwolf
a049c8043b push fix to training 2019-09-25 17:33:16 +02:00
mataney
a9f24a16bc [FIX] fix run_generation.py to work with batch_size > 1 2019-09-25 15:53:29 +03:00
thomwolf
5def3302f4 update run_glue 2019-09-25 12:38:08 +02:00
thomwolf
f71758f7a4 update internal glue processors 2019-09-25 12:00:50 +02:00
thomwolf
b5ec526f85 updated data processor and metrics 2019-09-24 17:10:50 +02:00
LysandreJik
f09e5ecef0 [Proposal] GLUE processors included in library 2019-09-24 09:47:34 -04:00
LysandreJik
c832f43a4d output_token_type -> token_type_ids 2019-09-24 07:21:38 -04:00
LysandreJik
3927d7756c Updated the GLUE pre-processing method 2019-09-24 07:15:11 -04:00
LysandreJik
9d44236f70 Updated DistilBERT 2019-09-24 07:03:24 -04:00
Lorenzo Ampil
4b543c3007 Add option to use a 'stop token' which will be used to truncate the output text to everything till right before the 'stop token' 2019-09-22 21:38:38 +08:00
VictorSanh
9f995b99d4 minor fixes 2019-09-19 21:36:06 +00:00
VictorSanh
3fe5c8e8a8 update bert-base-uncased rslts 2019-09-19 19:34:22 +00:00
VictorSanh
354944e607 [distillation] big update w/ new weights 2019-09-19 19:25:21 +00:00
LysandreJik
60414f31a9 GLUE updated with new methods 2019-09-19 10:55:06 +02:00
LysandreJik
bf503158c5 Sentence -> Sequence. Removed output_mask from the special token addition methods. 2019-09-19 10:55:06 +02:00
LysandreJik
de8e14b6c0 Added DistilBERT to run_squad script 2019-09-19 10:55:06 +02:00
LysandreJik
88368c2a16 Added DistilBERT to run_lm_finetuning 2019-09-19 10:55:06 +02:00
LysandreJik
75635072e1 Updated GLUE script to add DistilBERT. Cleaned up unused args in the utils file. 2019-09-19 10:55:06 +02:00
LysandreJik
59057abe52 typo 2019-09-19 10:55:06 +02:00
LysandreJik
bac332fec0 Updated the GLUE data processor. Corrections to RoBERTa and XLNet. 2019-09-19 10:55:06 +02:00
Erik Chan
f0340eccf9
Typo 
Typo
 2019-09-18 13:42:11 -07:00
erenup
8960988f35 fixed to find best dev acc 2019-09-19 01:10:05 +08:00
erenup
46ffc28329 Merge branch 'master' into run_multiple_choice_merge 
# Please enter a commit message to explain why this merge is necessary,
# especially if it merges an updated upstream into a topic branch.
#
# Lines starting with '#' will be ignored, and an empty message aborts
# the commit.
 2019-09-18 21:43:46 +08:00
erenup
15143fbad6 move run_multiple_choice.py and utils_multiple_choice.py to examples 2019-09-18 21:18:46 +08:00
erenup
3cd6289758 Merge remote-tracking branch 'huggingface/master' into run_multiple_choice_merge 
# Conflicts:
#	examples/contrib/run_swag.py
 2019-09-18 21:16:59 +08:00
erenup
36362cf086 move schedule.step after optimizer.step 2019-09-18 21:13:40 +08:00
thomwolf
e768f2322a update run_openai_gpt to fix #1264 2019-09-18 10:07:47 +02:00
thomwolf
8334993915 clean up examples - updated to new keyword inputs - #1246 2019-09-18 10:01:27 +02:00
erenup
5882c442e5 add example usage 2019-09-16 22:38:08 +08:00
erenup
982f181aa7 Merge remote-tracking branch 'origin/master' into run_multiple_choice_add_doc 2019-09-16 19:12:00 +08:00
erenup
84b9d1c423 Merge remote-tracking branch 'huggingface/master' 
# Conflicts:
#	pytorch_transformers/__init__.py
 2019-09-16 19:06:12 +08:00
erenup
603b470a3d add warnning info 2019-09-16 18:53:37 +08:00
erenup
4812a5a767 add doc string 2019-09-16 11:50:18 +08:00
VictorSanh
32e1332acf [distil] fix once for all general logger for scripts 2019-09-11 14:19:07 +00:00
VictorSanh
364920e216 fix small bug/typo 2019-09-10 21:45:01 +00:00
Thomas Wolf
23c23f5399
Merge pull request #1229 from SKRohit/master 
changes in evaluate function in run_lm_finetuning.py
 2019-09-10 22:16:45 +02:00
searchivarius
eab980fd68 Fix to prevent crashing on assert len(tokens_b)>=1 2019-09-09 19:58:08 -04:00
VictorSanh
a95ced6260 [Distillation] save last chkpt as pytorch_model.bin 2019-09-09 19:53:35 +00:00
Rohit Kumar Singh
e5df36397b
changes in return statement of evaluate function 
changed `results` to `result` and removed `results` dict defined previously
 2019-09-09 19:55:57 +05:30
LysandreJik
3f91338be9 Patched a few outdated parameters 2019-09-06 17:48:06 -04:00
LysandreJik
f47f9a5874 Updated outdated examples 2019-09-06 17:10:33 -04:00
LysandreJik
5e151f5e77 Table of contents 2019-09-06 12:08:36 -04:00
LysandreJik
593c070435 Better examples 2019-09-06 12:00:12 -04:00
VictorSanh
dddd6b9927 Update DistilBERT training code 2019-09-05 18:26:14 +00:00
Stefan Schweter
a1c34bd286 distillation: fix ModuleNotFoundError error in token counts script 2019-08-31 12:21:38 +02:00
Thomas Wolf
51e980ce36
Merge pull request #1155 from anhnt170489/apex_fp16 
Update apex fp16 implementation
 2019-08-30 23:29:11 +02:00
VictorSanh
282c276e09 typos + file name coherence in distillation README 2019-08-30 12:02:29 -04:00
VictorSanh
803c1cc4ea fix relative import bug cf Issue #1140 2019-08-30 12:01:27 -04:00
Thomas Wolf
0a2fecdf90
Merge branch 'master' into master 2019-08-30 16:30:08 +02:00
Rabeeh KARIMI
39eb31e11e remove reloading tokenizer in the training, adding it to the evaluation part 2019-08-30 15:44:41 +02:00
Rabeeh KARIMI
350bb6bffa updated tokenizer loading for addressing reproducibility issues 2019-08-30 15:34:28 +02:00
Thomas Wolf
01ad55f8cf
Merge pull request #1026 from rabeehk/master 
loads the tokenizer for each checkpoint, to solve the reproducability…
 2019-08-30 14:15:36 +02:00