GitLab

* Fix bad examples

* Add black formatting to style_doc

* Use first nonempty line

* Put it at the right place

* Don't add spaces to empty lines

* Better templates

* Deal with triple quotes in docstrings

* Result of style_doc

* Enable mdx treatment and fix code examples in MDXs

* Result of doc styler on doc source files

* Last fixes

* Break copy from

* New doc styler

* Fix issue with args at the start

* Code sample fixes

* Style code examples in MDX

* Fix more patterns

* Typo

* Typo

* More patterns

* Do without black for now

* Get more info in error

* Docstring style

* Re-enable check

* Quality

* Fix add_end_docstring decorator

* Fix docstring

* Convert file_utils docstrings to Markdown

* Test on BERT

* Return block indent

* Temporarily disable doc styler

* Remove from quality checks as well

* Remove doc styler mess

* Remove check from circleCI

* Fix typo

* Convert file_utils docstrings to Markdown

* Test on BERT

* Return block indent

* Temporarily disable doc styler

* Remove from quality checks as well

* Remove doc styler mess

* Remove check from circleCI

* Fix typo

* Let's go on all other model files

* Add templates too

* Styling and quality

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

Fix examples: 'CausalLMOutputWithCrossAttentions' object has no attribute 'last_hidden_state' (#14678)
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

* Add a post init method to all models

* Fix tests

* Fix last tests

* Fix templates

* Add comment

* Forgot to save

* minor modification to the wav2vec2 modeling file to support tensor-parallelism with DeepSpeed on this HuggingFace model

* refine the comments

* synch changes

* fix comments

* refine comments

* fix format

* raising exceptions instead of using assertions for few models

* fixed formatting issues

* fixing copy inconsistencies

* Generalize problem_type to all classification models

* Missing import

* Deberta BC and fix tests

* Fix template

* Missing imports

* Revert change to reformer test

* Fix style

* unispeech

* add copy from

* remove hubert copy from

* finish for today

* add unispeech-sat

* adapt more

* up

* up

* up

* up

* add modeling

* add tests

* up

* up

* finish

* up

* Apply suggestions from code review

* up

* up

* Apply suggestions from code review
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* up

* up
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* BartEnocder add set_input_embeddings

To unify the interface, add set_input_embeddings to BartEncoder.

* BartEnocder add get_input_embeddings

* up

* finish

* up

* up

* finish

In BartForConditionalGeneration.forward, if labels are provided,
   decoder_input_ids are set to the labels shifted to the right.
   This is problematic: if decoder_inputs_embeds is also set,
   the call to self.model, which eventually gets to BartDecoder.forward,
   will raise an error.
   The fix is quite simple, similar to what is there already in
   BartModel.forward. Mainly, we should not
   compute decoder_input_ids if decoder_inputs_embeds is provided.
Co-authored-by: Silviu Vlad Oprea <silviuvo@amazon.co.uk>

We use `torch.unique` here only to check whether every elements have
the same value.
Therefore, we can use `torch.unique_consecutive` here.

This function eliminates all but the first element from every consecutive
group of equivalent elements.
Like, if we apply this function to `[1, 2, 2, 1]`, it will result in
`[1, 2, 1]`.

As you could see, this is enough for checking whether every elements
have the same value.

Since `torch.unique_consecutive` do less thing, it is much more faster.
On my computer, it is 25x faster on GPU and 15x faster on CPU.

* Make gradient_checkpointing a training argument

* Update src/transformers/modeling_utils.py
Co-authored-by: Stas Bekman <stas00@users.noreply.github.com>

* Update src/transformers/configuration_utils.py
Co-authored-by: Stas Bekman <stas00@users.noreply.github.com>

* Fix tests

* Style

* document Gradient Checkpointing as a performance feature

* Small rename

* PoC for not using the config

* Adapt BC to new PoC

* Forgot to save

* Rollout changes to all other models

* Fix typo
Co-authored-by: Stas Bekman <stas00@users.noreply.github.com>
Co-authored-by: Stas Bekman <stas@stason.org>

* consistent nn. and nn.functional

* fix glitch

* fix glitch #2

* Fixing bug that appears when using distilation (and potentially other uses).
During backward pass Pytorch complains with:
RuntimeError: one of the variables needed for gradient computation has been modified by an inplace operation
This happens because the QA model code modifies the start_positions and end_positions input tensors, using clamp_ function: as a consequence the teacher and the student both modifies the inputs, and backward pass fails.

* Fixing all models QA clamp_ bug.

* Fixed Typo in modeling_bart.py - Issue #11895

* Fixed Typo in modeling_bart.py

* modify qa-trainer

* fix flax model

* Add missing head masking for generate() function

* Add head_mask, decoder_head_mask and cross_attn_head_mask
into prepare_inputs_for_generation for generate() function
for multiple encoder-decoder models.

* Add test_genereate_with_head_masking

* [WIP] Update the new test and handle special cases

* make style

* Omit ProphetNet test so far

* make fix-copies

* fix docs for decoder_input_ids

* revert the changes for bart and mbart

* Fix cross-attention head mask for Torch BART models

* Fix head masking for cross-attention module for the following
models: BART, Blenderbot, Blenderbot_small, M2M_100, Marian, MBart,
Pegasus

* Enable test_headmasking for M2M_100 model

* Fix cross_head_mask for FSMT, LED and T5

* This commit fixes `head_mask` for cross-attention modules
in the following models: FSMT, LED, T5

* It also contains some smaller changes in doc so that
it is be perfectly clear the shape of `cross_head_mask`
is the same as of `decoder_head_mask`

* Update template

* Fix template for BartForCausalLM

* Fix cross_head_mask for Speech2Text models

* Fix cross_head_mask in templates

* Fix args order in BartForCausalLM template

* Fix doc in BART templates

* Make more explicit naming

* `cross_head_mask` -> `cross_attn_head_mask`

* `cross_layer_head_mask` -> `cross_attn_layer_head_mask`

* Fix doc

* make style quality

* Fix speech2text docstring

* The 'warn' method is deprecated

* fix test

* Remove version warning in pretrained BART models

* Put it at the base model

* Refactor checkpoint name in ALBERT and ALBERT_tf

* Refactor checkpoint name in BART and BART_tf

* Refactor checkpoint name in BERT generation

* Refactor checkpoint name in Blenderbot_tf

* Refactor checkpoint name in Blenderbot_small_tf

* Refactor checkpoint name in ConvBERT AND CONVBERT_TF

* Refactor checkpoint name in CTRL AND CTRL_TF

* Refactor checkpoint name in DistilBERT AND DistilBERT_TF

* Refactor checkpoint name in DistilBERT redo

* Refactor checkpoint name in Electra and Electra_tf

* Refactor checkpoint name in FlauBERT and FlauBERT_tf

* Refactor checkpoint name in FSMT

* Refactor checkpoint name in GPT2 and GPT2_tf

* Refactor checkpoint name in IBERT

* Refactor checkpoint name in LED and LED_tf

* Refactor checkpoint name in Longformer and Longformer_tf

* Refactor checkpoint name in Lxmert and Lxmert_tf

* Refactor checkpoint name in Marian_tf

* Refactor checkpoint name in MBART and MBART_tf

* Refactor checkpoint name in MobileBERT and MobileBERT_tf

* Refactor checkpoint name in mpnet and mpnet_tf

* Refactor checkpoint name in openai and openai_tf

* Refactor checkpoint name in pegasus_tf

* Refactor checkpoint name in reformer

* Refactor checkpoint name in Roberta and Roberta_tf

* Refactor checkpoint name in SqueezeBert

* Refactor checkpoint name in Transformer_xl and Transformer_xl_tf

* Refactor checkpoint name in XLM and XLM_tf

* Refactor checkpoint name in XLNET and XLNET_tf

* Refactor checkpoint name in BERT_tf

* run make tests, style, quality, fixup

* fix speed degradation bug t5

* fix for all models

* fix code quality

* Assumption of padding_idx <2 might not stand

* Use offset instead of 2

* Fix with black

* Change behavior to warning instead for backward compatibility.

* Fix with black

* Remove warning

* Make padding_idx non-required

* padding_idx fix for blenderbot

* padding_idx fix for blenderbot_small

* padding_idx fix for led

* padding_idx fix for mbart

* Remove extra whitespaces

* padding_idx fix for template

* Fix padding_idx passed to nn.Embedding mistake

* Fixed padding_idx passed to positional embedding in template

* Remove padding_idx from pytorch learned positional embeddings

* Remove accidentally added quotes

* Remove padding_idx from tf learned positional embeddings

* Remove zeroing of weights in __init__
Co-authored-by: Wang Ming Rui <mingrui.wang@C02CJTUYMD6M.local>

* add forced logits processors

* delete adjust_logits method

* add forced_eos_token_id argument in config

* add tests for forced logits processors

* update gen utils tests

* add forced option to tf generate

* remove adjust_logits method from tf models

* update adjust_logits for marian

* delete _force_token_id_to_be_generated method

* style

* import warnings

* pass max_length to _get_logits_processor

* set forced_eos_token_id to None

* set forced attributes in conf utils

* typo

* fix rag generate

* add forced_eos_token_id in rag config

* remove force_bos_token_to_be_generated from BartConfig

* remove _force_token_ids_generation from FSMT

* nit

* fix negative constant

* apply suggestions from code review

* add prepare_decoder_input_ids_from_labels in s2s models

* support lbl smoothing and enc/emb freezing

* fix freezing

* use pad_token_id from config

* remove embed freezing and add warning

* prepare decoder_input_ids inside DataCollatorForSeq2Seq

* initiliaze bart4causalLM

* create BartDecoderWrapper, setters/getters

* delete spaces

* forward and additional methods

* update cache function, loss function, remove ngram* params in data class.

* add bartcausallm, bartdecoder testing

* correct bart for causal lm

* remove at

* add mbart as well

* up

* fix typo

* up

* correct

* add pegasusforcausallm

* add blenderbotforcausallm

* add blenderbotsmallforcausallm

* add marianforcausallm

* add test for MarianForCausalLM

* add Pegasus test

* add BlenderbotSmall test

* add blenderbot test

* fix a fail

* fix an import fail

* a fix

* fix

* Update modeling_pegasus.py

* fix models

* fix inputs_embeds setting getter

* adapt tests

* correct repo utils check

* finish test improvement

* fix tf models as well

* make style

* make fix-copies

* fix copies

* run all tests

* last changes

* fix all tests
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>

* Fix model templates and use less than 119 chars

* Missing new line

* Update past_key_values in gpt2 (#9391)

* Update generation_utils, and rename some items

* Update modeling_gpt2 to avoid an error in gradient_checkpointing

* Remove 'reorder_cache' from util and add variations to XLNet, TransfoXL, GPT-2

* Change the location of '_reorder_cache' in modeling files

* Add '_reorder_cache' in modeling_ctrl

* Fix a bug of my last commit in CTRL

* Add '_reorder_cache' to GPT2DoubleHeadsModel

* Manage 'use_cache' in config of test_modeling_gpt2

* Clean up the doc string

* Update src/transformers/models/gpt2/modeling_gpt2.py
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* Fix the doc string (GPT-2, CTRL)

* improve gradient_checkpointing_behavior
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>

* Add head_mask/decoder_head_mask for BART

This branch implement head_mask and decoder_head_mask
for BART-based models. Full list below:
- BART
- MBart
- Blenderbot
- BlenderbotSmall
- Marian
- Pegasus

Everything is accompanied with updated testing.

* Fix test_headmasking for BART models

* Fix text_headmasking for BART-like models
which has only 2 layers in each modules.
The condition
```
self.assertNotEqual(attentions[1][..., 0, :, :].flatten().sum().item(), 0.0)
```
is, therefore, invalid for encoder-decoder models considering
the `head_mask`
```
head_mask = torch.ones(
    self.model_tester.num_hidden_layers,
    self.model_tester.num_attention_heads,
    device=torch_device,
)
head_mask[0, 0] = 0
head_mask[-1, :-1] = 0
```
specified in the `test_headmasking` test/function.

* Adjust test_modeling_common.py to reflect T5 input args

* Update tests/test_modeling_common.py
Co-authored-by: Lysandre Debut <lysandre@huggingface.co>

* Apply suggestions from code review
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* make style

* make fix-copies
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Co-authored-by: Lysandre Debut <lysandre@huggingface.co>
Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>