GitLab

[Marian Fixes] prevent predicting pad_token_id before softmax, support language codes, name multilingual models (#4290)

* add first text for generation

* add generation pipeline to usage

* Created using Colaboratory

* correct docstring

* finish

* Add QA trainer example for TF

* Make data_dir optional

* Fix parameter logic

* Fix feature convert

* Update the READMEs to add the question-answering task

* Apply style

* Change 'sequence-classification' to 'text-classification' and prefix with 'eval' all the metric names

* Apply style

* Apply style

Fix for #3865. PretrainedTokenizer mapped " do not" into " don't" when .decode(...) is called. Removed the " do not" --> " don't" mapping from clean_up_tokenization(...). (#4024)

* Improvements to the wandb integration

* small reorg + no global necessary

* feat(trainer): log epoch and final metrics

* Simplify logging a bit

* Fixup

* Fix crash when just running eval
Co-authored-by: Chris Van Pelt <vanpelt@gmail.com>
Co-authored-by: Boris Dayma <boris.dayma@gmail.com>

* Allow BatchEncoding to be initialized empty.

This is required by recent changes introduced in TF 2.2.

* Attempt to unpin Tensorflow to 2.2 with the previous commit.

* catch gpu len 1 set to gpu0

* Add mpc to trainer

* Add MPC for TF

* fix TF automodel for MPC and add Albert

* Apply style

* Fix import

* Note to self: double check

* Make shape None, None for datasetgenerator output shapes

* Add from_pt bool which doesnt seem to work

* Original checkpoint dir

* Fix docstrings for automodel

* Update readme and apply style

* Colab should probably not be from users

* Colabs should probably not be from users

* Add colab

* Update README.md

* Update README.md

* Cleanup __intit__

* Cleanup flake8 trailing comma

* Update src/transformers/training_args_tf.py

* Update src/transformers/modeling_tf_auto.py
Co-authored-by: Viktor Alm <viktoralm@pop-os.localdomain>
Co-authored-by: Julien Chaumond <chaumond@gmail.com>

fixed missing torch module import in example usage code

* pin TF to 2.1

* Pin flake8 as well

* simplify cache vars and allow for TRANSFORMERS_CACHE env

As it currently stands, "TRANSFORMERS_CACHE" is not an accepted variable. It seems that the these variables were not updated when moving from version pytorch_transformers to transformers. In addition, the fallback procedure could be improved. and simplified. Pathlib seems redundant here.

* Update file_utils.py

* allow gpt2 to be exported to valid ONNX model

* cast size from int to float explictly

"Migrating from pytorch-transformers to transformers" is missing in the main document. It is available in the main `readme` thought. Just move it to the document.

* Add ALBERT to convert command of transformers-cli

* Document ALBERT tf to pytorch model conversion

* docs: fix link to token classification (NER) example

* examples: fix links to NER scripts

* Create README.md

* Update model_cards/lserinol/bert-turkish-question-answering/README.md
Co-authored-by: Julien Chaumond <chaumond@gmail.com>

* Fix the issue to properly run the accumulator with TF 2.2

* Apply style

* Fix training_args_tf for TF 2.2

* Fix the TF training args when only one GPU is available

* Remove the fixed version of TF in setup.py

* adapt convert script

* update convert script

* finish

* fix marian pretrained docs

- MarianSentencepieceTokenizer - > MarianTokenizer
- Start using unk token.
- add docs page
- add better generation params to MarianConfig
- more conversion utilities

* [TPU] Doc, fix xla_spawn.py, only preprocess dataset once

* Update examples/README.md

* [xla_spawn] Add `_mp_fn` to other Trainer scripts

* [TPU] Fix: eval dataloader was None