* Add tie_weights() to LM heads and set bias in set_output_embeddings()
The bias were not tied correctly in some LM heads, and this change should fix that.
* Moving test_save_and_load_low_cpu_mem_usage to ModelTesterMixin
* Adding _tie_weights() to MPNet and Vilt
* Skip test for low cpu mem usage for Deta/DeformableDetr since they cannot init on meta device
* Rename to test name to save_load to match the convention
* test that tied output embeddings aren't initialized on load
* don't initialize the output embeddings if we're going to tie them to the input embeddings
* skip bf16 test if not supported by device
* fix
* fix bis
* use is_torch_bf16_available_on_device
* use is_torch_fp16_available_on_device
* fix & use public llama
* use 1b model
* fix flacky test
---------
Co-authored-by: Your Name <you@example.com>
* Skip nn.Module.reset_parameters
* Actually skip
* Check quality
* Maybe change all inits
* Fix init issues: only modify public functions
* Add a small test for now
* Style
* test updates
* style
* nice tes
* style
* make it even faster
* one more second
* remove fx icompatible
* Update tests/test_modeling_common.py
Co-authored-by: Lysandre Debut <hi@lysand.re>
* Update tests/test_modeling_common.py
Co-authored-by: Lysandre Debut <hi@lysand.re>
* skip
* fix quality
* protect the import
---------
Co-authored-by: Lysandre Debut <hi@lysand.re>
* add sdpa
* wip
* cleaning
* add ref
* yet more cleaning
* and more :)
* wip llama
* working llama
* add output_attentions=True support
* bigcode sdpa support
* fixes
* gpt-bigcode support, require torch>=2.1.1
* add falcon support
* fix conflicts falcon
* style
* fix attention_mask definition
* remove output_attentions from attnmaskconverter
* support whisper without removing any Copied from statement
* fix mbart default to eager renaming
* fix typo in falcon
* fix is_causal in SDPA
* check is_flash_attn_2_available in the models init as well in case the model is not initialized through from_pretrained
* add warnings when falling back on the manual implementation
* precise doc
* wip replace _flash_attn_enabled by config.attn_implementation
* fix typo
* add tests
* style
* add a copy.deepcopy on the config in from_pretrained, as we do not want to modify it inplace
* obey to config.attn_implementation if a config is passed in from_pretrained
* fix is_torch_sdpa_available when torch is not installed
* remove dead code
* Update src/transformers/modeling_attn_mask_utils.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
* Update src/transformers/modeling_attn_mask_utils.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
* Update src/transformers/modeling_attn_mask_utils.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
* Update src/transformers/modeling_attn_mask_utils.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
* Update src/transformers/modeling_attn_mask_utils.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
* Update src/transformers/models/bart/modeling_bart.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
* remove duplicate pretraining_tp code
* add dropout in llama
* precise comment on attn_mask
* add fmt: off for _unmask_unattended docstring
* precise num_masks comment
* nuke pretraining_tp in LlamaSDPAAttention following Arthur's suggestion
* cleanup modeling_utils
* backward compatibility
* fix style as requested
* style
* improve documentation
* test pass
* style
* add _unmask_unattended tests
* skip meaningless tests for idefics
* hard_check SDPA requirements when specifically requested
* standardize the use if XXX_ATTENTION_CLASSES
* fix SDPA bug with mem-efficient backend on CUDA when using fp32
* fix test
* rely on SDPA is_causal parameter to handle the causal mask in some cases
* fix FALCON_ATTENTION_CLASSES
* remove _flash_attn_2_enabled occurences
* fix test
* add OPT to the list of supported flash models
* improve test
* properly test on different SDPA backends, on different dtypes & properly handle separately the pad tokens in the test
* remove remaining _flash_attn_2_enabled occurence
* Update src/transformers/modeling_utils.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
* Update src/transformers/modeling_utils.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
* Update src/transformers/modeling_utils.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
* Update src/transformers/modeling_attn_mask_utils.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
* Update docs/source/en/perf_infer_gpu_one.md
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
* remove use_attn_implementation
* fix docstring & slight bug
* make attn_implementation internal (_attn_implementation)
* typos
* fix tests
* deprecate use_flash_attention_2=True
* fix test
* add back llama that was removed by mistake
* fix tests
* remove _flash_attn_2_enabled occurences bis
* add check & test that passed attn_implementation is valid
* fix falcon torchscript export
* fix device of mask in tests
* add tip about torch.jit.trace and move bt doc below sdpa
* fix parameterized.expand order
* move tests from test_modeling_attn_mask_utils to test_modeling_utils as a relevant test class is already there
* update sdpaattention class with the new cache
* Update src/transformers/configuration_utils.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
* Update src/transformers/models/bark/modeling_bark.py
* address review comments
* WIP torch.jit.trace fix. left: test both eager & sdpa
* add test for torch.jit.trace for both eager/sdpa
* fix falcon with torch==2.0 that needs to use sdpa
* fix doc
* hopefully last fix
* fix key_value_length that has no default now in mask converter
* is it flacky?
* fix speculative decoding bug
* tests do pass
* fix following #27907
---------
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
* Draft version of new KV Caching
This should allow Attention Sinks (https://github.com/tomaarsen/attention_sinks)
/ StreamingLLM (https://arxiv.org/abs/2309.17453) to be easily implemented
in a third-party or in transformers directly
* Address numerous PR suggestions
1. Move layer_idx from cache to ...Attention. Removes confusing set_layer_idx magic.
2. Always convert past_key_values to Cache instance at the start of ...Attention, removes all other isinstance calls.
3. Remove __bool__ and __getitem__ magic as they're confusing.
4. past_key_values.update(key, value, idx) now returns key, value.
5. Add use_legacy_cache flag, defaults to None, i.e. Falsey. This breaks generate for now, until 1) the cache is used is generate() or 2) use_legacy_cache is defaulted to True in generate() until we change it in another PR.
6. Separate key_cache and value_cache.
Some work is still needed to see if the SinkCache can conveniently be implemented with just one update method.
* Implement the SinkCache through backward+forward rotations
* Integrate (Sink)Cache with Llama FA2
* Set use_legacy_cache=True as default, allows for test passes
* Move from/to_legacy_cache to ...Model class
* Undo unnecessary newline change
* Remove copy utility from deprecated OpenLlama
* Match import style
* manual rebase with main
* Cache class working with generate (#1)
* Draft version of new KV Caching
This should allow Attention Sinks (https://github.com/tomaarsen/attention_sinks)
/ StreamingLLM (https://arxiv.org/abs/2309.17453) to be easily implemented
in a third-party or in transformers directly
* Address numerous PR suggestions
1. Move layer_idx from cache to ...Attention. Removes confusing set_layer_idx magic.
2. Always convert past_key_values to Cache instance at the start of ...Attention, removes all other isinstance calls.
3. Remove __bool__ and __getitem__ magic as they're confusing.
4. past_key_values.update(key, value, idx) now returns key, value.
5. Add use_legacy_cache flag, defaults to None, i.e. Falsey. This breaks generate for now, until 1) the cache is used is generate() or 2) use_legacy_cache is defaulted to True in generate() until we change it in another PR.
6. Separate key_cache and value_cache.
Some work is still needed to see if the SinkCache can conveniently be implemented with just one update method.
* Integrate (Sink)Cache with Llama FA2
* Move from/to_legacy_cache to ...Model class
* Undo unnecessary newline change
* Match import style
* working generate
* Add tests; Simplify code; Apply changes to Mistral and Persimmon
* fix rebase mess
* a few more manual fixes
* last manual fix
* propagate changes to phi
* upgrade test
* add use_legacy_cache docstring; beef up tests
* reintroduce unwanted deletes
---------
Co-authored-by: Tom Aarsen <Cubiegamedev@gmail.com>
* move import
* add default to model_kwargs.get('use_legacy_cache')
* correct failing test
* Apply suggestions from code review
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
* apply PR suggestions
* fix failing test
* Apply suggestions from code review
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
Co-authored-by: Tom Aarsen <37621491+tomaarsen@users.noreply.github.com>
* PR comments
* tmp commit
* add docstrings
* more tests, more docstrings, add to docs
* derp
* tmp commit
* tmp dbg
* more dbg
* fix beam search bug
* cache can be a list of tuples in some models
* fix group beam search
* all but sinkcache integration tests
* fix sink cache and add hard integration test
* now also compatible with input_embeds input
* PR comments
* add Cache support to Phi+FA2
* make fixup
---------
Co-authored-by: Joao Gante <joao@huggingface.co>
Co-authored-by: Joao Gante <joaofranciscocardosogante@gmail.com>
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
* try to stylify using ruff
* might need to remove these changes?
* use ruf format andruff check
* use isinstance instead of type comparision
* use # fmt: skip
* use # fmt: skip
* nits
* soem styling changes
* update ci job
* nits isinstance
* more files update
* nits
* more nits
* small nits
* check and format
* revert wrong changes
* actually use formatter instead of checker
* nits
* well docbuilder is overwriting this commit
* revert notebook changes
* try to nuke docbuilder
* style
* fix feature exrtaction test
* remve `indent-width = 4`
* fixup
* more nits
* update the ruff version that we use
* style
* nuke docbuilder styling
* leve the print for detected changes
* nits
* Remove file I/O
Co-authored-by: charliermarsh
<charlie.r.marsh@gmail.com>
* style
* nits
* revert notebook changes
* Add # fmt skip when possible
* Add # fmt skip when possible
* Fix
* More ` # fmt: skip` usage
* More ` # fmt: skip` usage
* More ` # fmt: skip` usage
* NIts
* more fixes
* fix tapas
* Another way to skip
* Recommended way
* Fix two more fiels
* Remove asynch
Remove asynch
---------
Co-authored-by: charliermarsh <charlie.r.marsh@gmail.com>
* add whisper fa2
* correct
* change all
* correct
* correct
* fix more
* fix more
* fix more
* fix more
* fix more
* fix more
* Apply suggestions from code review
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>
* fix more
* fix more
* fix more
* fix more
* fix more
---------
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>
* Safetensors serialization by default
* First pass on the tests
* Second pass on the tests
* Third pass on the tests
* Fix TF weight loading from TF-format safetensors
* Specific encoder-decoder fixes for weight crossloading
* Add VisionEncoderDecoder fixes for TF too
* Change filename test for pt-to-tf
* One missing fix for TFVisionEncoderDecoder
* Fix the other crossload test
* Support for flax + updated tests
* Apply suggestions from code review
Co-authored-by: Sanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com>
* Sanchit's comments
* Sanchit's comments 2
* Nico's comments
* Fix tests
* cleanup
* Apply suggestions from code review
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>
---------
Co-authored-by: Matt <rocketknight1@gmail.com>
Co-authored-by: Sanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com>
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>
* stronger GC tests
* better tests and skip failing tests
* break down into 3 sub-tests
* break down into 3 sub-tests
* refactor a bit
* more refactor
* fix
* last nit
* credits contrib and suggestions
* credits contrib and suggestions
---------
Co-authored-by: Yih-Dar <2521628+ydshieh@users.noreply.github.com>
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>
* add FA-2 support for mistral
* fixup
* add sliding windows
* fixing few nits
* v1 slicing cache - logits do not match
* add comment
* fix bugs
* more mem efficient
* add warning once
* add warning once
* oops
* fixup
* more comments
* copy
* add safety checker
* fixup
* Update src/transformers/models/mistral/modeling_mistral.py
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
* copied from
* up
* raise when padding side is right
* fixup
* add doc + few minor changes
* fixup
---------
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
* add Bros boilerplate
* copy and pasted modeling_bros.py from official Bros repo
* update copyright of bros files
* copy tokenization_bros.py from official repo and update import path
* copy tokenization_bros_fast.py from official repo and update import path
* copy configuration_bros.py from official repo and update import path
* remove trailing period in copyright line
* copy and paste bros/__init__.py from official repo
* save formatting
* remove unused unnecessary pe_type argument - using only crel type
* resolve import issue
* remove unused model classes
* remove unnecessary tests
* remove unused classes
* fix original code's bug - layer_module's argument order
* clean up modeling auto
* add bbox to prepare_config_and_inputs
* set temporary value to hidden_size (32 is too low because of the of the
Bros' positional embedding)
* remove decoder test, update create_and_check* input arguemnts
* add missing variable to model tests
* do make fixup
* update bros.mdx
* add boilerate plate for no_head inference test
* update BROS_PRETRAINED_MODEL_ARCHIVE_LIST (add naver-clova-ocr prefix)
* add prepare_bros_batch_inputs function
* update modeling_common to add bbox inputs in Bros Model Test
* remove unnecessary model inference
* add test case
* add model_doc
* add test case for token_classification
* apply fixup
* update modeling code
* update BrosForTokenClassification loss calculation logic
* revert logits preprocessing logic to make sure logits have original shape
* - update class name
* - add BrosSpadeOutput
- update BrosConfig arguments
* add boilerate plate for no_head inference test
* add prepare_bros_batch_inputs function
* add test case
* add test case for token_classification
* update modeling code
* update BrosForTokenClassification loss calculation logic
* revert logits preprocessing logic to make sure logits have original shape
* apply masking on the fly
* add BrosSpadeForTokenLinking
* update class name
put docstring to the beginning of the file
* separate the logits calculation logic and loss calculation logic
* update logic for loss calculation so that logits shape doesn't change
when return
* update typo
* update prepare_config_and_inputs
* update dummy node initialization
* update last_hidden_states getting logic to consider when return_dict is False
* update box first token mask param
* bugfix: remove random attention mask generation
* update keys to ignore on load missing
* run make style and quality
* apply make style and quality of other codes
* update box_first_token_mask to bool type
* update index.md
* apply make style and quality
* apply make fix-copies
* pass check_repo
* update bros model doc
* docstring bugfix fix
* add checkpoint for doc, tokenizer for doc
* Update README.md
* Update docs/source/en/model_doc/bros.md
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>
* Update bros.md
* Update src/transformers/__init__.py
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>
* Update docs/source/en/model_doc/bros.md
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>
* Apply suggestions from code review
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>
* apply suggestions from code review
* apply suggestions from code review
* revert test_processor_markuplm.py
* Update test_processor_markuplm.py
* apply suggestions from code review
* apply suggestions from code review
* apply suggestions from code review
* update BrosSpadeELForTokenClassification head name to entity linker
* add doc string for config params
* update class, var names to more explicit and apply suggestions from code review
* remove unnecessary keys to ignore
* update relation extractor to be initialized with config
* add bros processor
* apply make style and quality
* update bros.md
* remove bros tokenizer, add bros processor that wraps bert tokenizer
* revert change
* apply make fix-copies
* update processor code, update itc -> initial token, stc -> subsequent token
* add type hint
* remove unnecessary condition branches in embedding forward
* fix auto tokenizer fail
* update docstring for each classes
* update bbox input dimension as standard 2 points and convert them to 4
points in forward pass
* update bros docs
* apply suggestions from code review : update Bros -> BROS in bros.md
* 1. box prefix var -> bbox
2. update variable names to be more explicit
* replace einsum with torch matmul
* apply style and quality
* remove unused argument
* remove unused arguments
* update docstrings
* apply suggestions from code review: add BrosBboxEmbeddings, replace
einsum with classical matrix operations
* revert einsum update
* update bros processor
* apply suggestions from code review
* add conversion script for bros
* Apply suggestions from code review
* fix readme
* apply fix-copies
---------
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>
* Fix GPTNeoX beam search when using parallelize
* Fix beam search idx device when using model parallel
* remove onnx related stuff
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
* fix: move test_beam_search_on_multi_gpu to GenerationTesterMixin
* fix: add right item to _no_split_modules of MegaPreTrainedModel
* fix: add num_beams within parallelized beam_search test
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>
---------
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>
* fix
* revert cahnges and update resizing of embedding layer
* use wraning
* fixup
* more styling nits
* fix all tests that overload the embedding tests
* 👀👀 remove breakpoint
* remove useless overload + overload correctly where needed
* resize lm head with new vocab size
* reverse not necessary changes
* style
* fix CIs!
* fix last CI tests, adapt bark and Marian
* fixup
* Preliminary work on some models
* Fix test load missing and make sure nonpersistent buffers are tested
* Always ignore nonpersistent buffers if in state_dict
* Treat models
* More models
* Treat remaining models
* Fix quality
* Fix tests
* Remove draft
* This test is not needed anymore
* Fix copies
* Fix last test
* Newly added models
* Fix last tests
* Address review comments
