mirror of
https://github.com/huggingface/transformers.git
synced 2025-07-04 21:30:07 +06:00
80377eb018
6 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
fxmarty
|
80377eb018
|
F.scaled_dot_product_attention support (#26572)
* 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> |
||
|
Susnato Dhar
|
f84d85ba67
|
[FA-2] Add Flash Attention to Phi (#27661)
* add FA and modify doc file * test_flash_attn_2_generate_padding_right test overwritten * comment * modify persimmon modeling file * added speedup graph * more changes |
||
|
Alex McKinney
|
75336c1794
|
Add Llama Flax Implementation (#24587)
* Copies `modeling_flax_gpt_neo.py` to start * MLP Block. WIP Attention and Block * Adds Flax implementation of `LlamaMLP` Validated with in-file test. Some slight numeric differences, but assuming it isn't an issue * Adds `FlaxLlamaRMSNorm` layer `flax.linen` includes `RMSNorm` layer but not necessarily in all versions. Hence, we add in-file. * Adds FlaxLlamaAttention Copied from GPT-J as it has efficient caching implementation as well as rotary embeddings. Notice numerically different, but not by a huge amount. Needs investigating * Adds `FlaxLlamaDecoderLayer` numerically inaccurate, debugging.. * debugging rotary mismatch gptj uses interleaved whilst llama uses contiguous i think they match now but still final result is wrong. maybe drop back to just debugging attention layer? * fixes bug with decoder layer still somewhat numerically inaccurate, but close enough for now * adds markers for what to implement next the structure here diverges a lot from the PT version. not a big fan of it, but just get something working for now * implements `FlaxLlamaBlockCollection`] tolerance must be higher than expected, kinda disconcerting * Adds `FlaxLlamaModule` equivalent PyTorch model is `LlamaModel` yay! a language model🤗 * adds `FlaxLlamaForCausalLMModule` equivalent to `LlamaForCausalLM` still missing returning dict or tuple, will add later * start porting pretrained wrappers realised it probably needs return dict as a prereq * cleanup, quality, style * readds `return_dict` and model output named tuples * (tentatively) pretrained wrappers work 🔥 * fixes numerical mismatch in `FlaxLlamaRMSNorm` seems `jax.lax.rsqrt` does not match `torch.sqrt`. manually computing `1 / jax.numpy.sqrt` results in matching values. * [WIP] debugging numerics * numerical match I think issue was accidental change of backend. forcing CPU fixes test. We expect some mismatch on GPU. * adds in model and integration tests for Flax Llama summary of failing: - mul invalid combination of dimensions - one numerical mismatch - bf16 conversion (maybe my local backend issue) - params are not FrozenDict * adds missing TYPE_CHECKING import and `make fixup` * adds back missing docstrings needs review on quality of docstrings, not sure what is required. Furthermore, need to check if `CHECKPOINT_FOR_DOC` is valid. See TODO * commenting out equivalence test as can just use common * debugging * Fixes bug where mask and pos_ids were swapped in pretrained models This results in all tests passing now 🔥 * cleanup of modeling file * cleanup of test file * Resolving simpler review comments * addresses more minor review comments * fixing introduced pytest errors from review * wip additional slow tests * wip tests need to grab a GPU machine to get real logits for comparison otherwise, slow tests should be okay * `make quality`, `make style` * adds slow integration tests - checking logits - checking hidden states - checking generation outputs * `make fix-copies` * fix mangled function following `make fix-copies` * adds missing type checking imports * fixes missing parameter checkpoint warning * more finegrained 'Copied from' tags avoids issue of overwriting `LLAMA_INPUTS_DOCSTRING` * swaps import guards ??? how did these get swapped initially? * removing `inv_freq` again as pytorch version has now removed * attempting to get CI to pass * adds doc entries for llama flax models * fixes typo in __init__.py imports * adds back special equivalence tests these come from the gpt neo flax tests. there is special behaviour for these models that needs to override the common version * overrides tests with dummy to see if CI passes need to fill in these tests later * adds my contribution to docs * `make style; make quality` * replaces random masking with fixed to work with flax version * `make quality; make style` * Update src/transformers/models/llama/modeling_flax_llama.py Co-authored-by: Sanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com> * Update src/transformers/models/llama/modeling_flax_llama.py Co-authored-by: Sanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com> * Update src/transformers/models/llama/modeling_flax_llama.py Co-authored-by: Sanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com> * Update src/transformers/models/llama/modeling_flax_llama.py Co-authored-by: Sanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com> * Update src/transformers/models/llama/modeling_flax_llama.py Co-authored-by: Sanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com> * Update src/transformers/models/llama/modeling_flax_llama.py Co-authored-by: Sanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com> * updates `x`->`tensor` in `rotate_half` * addresses smaller review comments * Update docs/source/en/model_doc/llama.md Co-authored-by: Sanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com> * adds integration test class * adds `dtype` to rotary embedding to cast outputs * adds type to flax llama rotary layer * `make style` * `make fix-copies` * Apply suggestions from code review Co-authored-by: Sanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com> * applies suggestions from review * Update modeling_flax_llama.py * `make fix-copies` * Update tests/models/llama/test_modeling_llama.py Co-authored-by: Sanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com> * Update src/transformers/models/llama/modeling_flax_llama.py Co-authored-by: Sanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com> * fixes shape mismatch in FlaxLlamaMLP * applies some suggestions from reviews * casts attn output logits to f32 regardless of dtype * adds attn bias using `LlamaConfig.attention_bias` * adds Copied From comments to Flax Llama test * mistral and persimmon test change -copy from llama * updates docs index * removes Copied from in tests it was preventing `make fix-copies` from succeeding * quality and style * ignores FlaxLlama input docstring * adds revision to `_CHECKPOINT_FOR_DOC` * repo consistency and quality * removes unused import * removes copied from from Phi test now diverges from llama tests following FlaxLlama changes * adds `_REAL_CHECKPOINT_FOR_DOC` * removes refs from pr tests * reformat to make ruff happy --------- Co-authored-by: Sanchit Gandhi <93869735+sanchit-gandhi@users.noreply.github.com> |
||
|
Yih-Dar
|
30e92ea323
|
Trigger corresponding pipeline tests if tests/utils/tiny_model_summary.json is modified (#27693)
* fix --------- Co-authored-by: ydshieh <ydshieh@users.noreply.github.com> |
||
|
Arthur
|
651408a077
|
[Styling] stylify using ruff (#27144)
* 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> |
||
|
Susnato Dhar
|
e1c3ac2551
|
Add Phi-1 and Phi-1_5 (#26170)
* only dir not even init * init * tokenizer removed and reference of codegen added * modeling file updated a lot remaining app_rotary_emb * conversion script done * conversion script fixed, a lot of factoring done and most tests pass * added token_clf and extractive_QA_head * integration tests pass * flash attn tests pass! * config done * more docs in modeling file * some style fix * style and others * doc test error fix * more doc fix * some attention fixes * most fixes * style and other fixes * docs fix and config * doc fix * some comments * conversion script updated * conversion script updated * Revert "conversion script updated" This reverts commit e92378c54084ec0747041b113083d1746ecb6c7f. * final comments * add Phi to language_modeling.md * edit phi.md file * rebase and fix * removed phi-1.5 example * changed model_type from 'phi'->'mixformer-sequential' * small change * small change * revert \small change * changed mixformer-sequential->phi * small change * added phi-1.5 example instead of phi-1 * doc test might pass now * rebase and small change * added the dropout layer * more fixes * modified .md file * very very small doc change |