GitLab

* First draft of RWKV-4

* Add support for generate

* Style post-rebase

* Properly use state

* Write doc

* Fix doc

* More math

* Add model to README, dummies and clean config

* Fix init

* multiple fixes:

- fix common tests
- fix configuraion default values
- add CI test for checking state computation
- fix some CI tests

* correct tokenizer

* some tweaks

- fix config docstring
- fix failing tests

* fix CI tests

- add output_attention / output_hidden_states
- override test_initialization
- fix failing CIs

* fix conversion script

- fix sharded case
- add new arguments

* add slow tests + more fixes on conversion script

* add another test

* final fixes

* change single name variable

* add mock attention mask for pipeline to work

* correct eos token id

* fix nits

* add checkpoints

* Apply suggestions from code review
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* add `tie_word_embeddings` in docstring

* change tensor name

* fix final nits

* Trigger CI

---------
Co-authored-by: younesbelkada <younesbelkada@gmail.com>
Co-authored-by: Younes Belkada <49240599+younesbelkada@users.noreply.github.com>
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* starcoder has joined the chat

* indexing that works for all

* temperature controls speed

Generation: only check for eos_token if set

The check for unfinished_sequences.max(), which is to find sequences
that have ended early via eos_token_id, creates a synchronization point
even when there is no eos_token, which slows inference down.

This change moves the calculation to inside the condition checking for
eos_token, so that such slowdown may be removed by disabling this token.
Co-authored-by: John Doe <john.doe@example.com>

* working mvp

* remove breakpoint

* fix commit

* standardize outputs

* tmp commit

* tests almost ready

* tmp commit

* skip a few models

* Add streaming; Docs and examples

* document limitations

* PR commits

* Amy PR comments

* haha tokens go brrrr

* add conditional generation

* add comments

Edited one line in src/transormers/generation/utils.py. Changed dist.world_size() to dist.get_world_size() since world_size() doesn't exist in pytorch.dist.

* [deepspeed zero3] need generate(synced_gpus=True, ...)

* fix

* rework per Sylvain's suggestion

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

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

* time to say goodbye, torch 1.7 and 1.8

* clean up torch_int_div

* clean up is_torch_less_than_1_8-9

* update

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

Fix: unfinished_sequences with correct device 

The original code was causing errors when running torch.jit.trace due to the tensor options being incorrect. I fixed this by using torch.ones to create a tensor with the correct device and dtype. This should resolve the issue with running torch.jit.trace.

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

fix broken links

fix missing unfinished_sequences

* add tests with multiple eos_token_ids

* make math.prod instead of sum

* make fixup

* fix long and also use np.prod since math.prod does not exist <python 3.8

* make fixup

* add prod util

* use prod util instead of np.prod

* make fixup

* previous .long location

* use tensor ops

* remove prod

* remove prod

* update device

* make fixup

* fix none

* fix past renamed to past_key_value

* update more `past`that were ski^êd

* fixup

* remove changes made to rag

* refactor `_reorder_cache` to use `past_key_values`

* fix git `prepare_inputs_for_generation` to pass tests when false is needed in use_cache

input_ids_seq_length doesn't exist in the GenerationConfig, it exists as local variable in the function.

Setting exponential_decay_length_penalty therefore results in an error:
`AttributeError: 'GenerationConfig' object has no attribute 'input_ids_seq_length'`

This simple change fixes this issue, and the exponential_decay_length_penalty works as expected.

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

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

* Add hallucination penalty

* Make quality changes

* Inverse penalty

* Fix imports & quality

* Fix name spelling issue

* set encoder_repetition_penalty and fix quality

* Fix failing test

* Add to config_common_kwargs

* Fix modelling_rag error

* Update src/transformers/generation_logits_process.py
Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>

* Remove breakpoint

* Make style fixes

* Update encoder_repetition_penalty default value

* Merge latest main changes

* Make fixup changes

* Add EncoderRepetitionPenaltyLogitsProcessor to generation/__init__.py

* Fix repo-inconsistency

* Remove venv

* Remove tensorflow-macos & add tests

* Add documentation

* Fix quality issues

* move encoder_repetition_penalty to config

* Update src/transformers/configuration_utils.py
Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>

* Update src/transformers/generation/configuration_utils.py
Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>

* Remove encoder_repetition_penalty from tests

* Fix type error

* Fix format error
Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>

* Add epsilon- and eta-sampling.

Add epsilon- and eta-sampling, following the official code from https://github.com/john-hewitt/truncation-sampling and adapting to be more configurable, as required by Huggingface transformers.

* Add unit tests for epsilon- and eta-sampling.

* Black: fix code formatting.

* Fix docstring spacing.

* Clean up newlines.

* Fix implementation bugs and their associated tests.

* Remove epsilon- and eta-sampling parameters from PretrainedConfig.

* Clarify and clean up the documentation.

* Remove parameters for PretrainedConfig test.

* initial commit, refactoring the text generation api reference

* removed repetitive code examples

* Refactoring the text generation docs to reduce repetition

* make style

Add `min_new_tokens` argument in generate() (implementation based on `MinNewTokensLengthLogitsProcessor`) (#21044)

add a new parameter min_new_tokens for generate()

* start cleanup

* more updates

* more models are affected

* more updates

* update generation utils

* style

* revert change that removed reorder cachce

* update generation utils

* style

* style

* remove reorder cache

* Add StopIdStoppingCriteria

* add a working test for stop id criteria

* add to global scope

* add stop_ids to generate

* add pipeline test

* use tokenizer encode in test

* add test to generation utils

* reformat

* fixup

* make-fix-copies

* rename to stop_token_id

* use stop_tokens instead

* add to text to text generation

* make fixup

* make repo-consistency

* Add support for list of ints for eos_token_id inside generation/utils.py

* Instead of having if elses, cast the eos_token_id into a List[int]

* Add List[int] support for logits_process.py

* add List[int] for beam_search.py

* add List[int] for forced_eos_token_id

* revert stop token id stopping criteria changes

* make fixup

* fix tests

* add eos_token_id to generation/utils.py and added tests test_utils.py

* add eos_token_id type hints and fix for pad tokens

* add comments

* remove some prints and remove forced false test

* fix

* put back test_stop_sequence_stopping_criteria

* remove unused import and make fixup

* add a none check

* update docstring

* add more docstring for list ints

* make fixup