SUMIN/HuggingFace_transformer
1
0
Fork 0
Code Issues Pull Requests Actions 7 Packages Projects Releases Wiki Activity
17,433 Commits 5 Branches 239 Tags
62ab94dea8015440a6617afe0ec28d346976d884
Commit Graph

17433 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
dependabot[bot]
62ab94dea8 Bump tornado from 6.4.1 to 6.4.2 in /examples/research_projects/visual_bert (#34887) 
Bump tornado in /examples/research_projects/visual_bert

Bumps [tornado](https://github.com/tornadoweb/tornado) from 6.4.1 to 6.4.2.
- [Changelog](https://github.com/tornadoweb/tornado/blob/v6.4.2/docs/releases.rst)
- [Commits](https://github.com/tornadoweb/tornado/compare/v6.4.1...v6.4.2)

---
updated-dependencies:
- dependency-name: tornado
  dependency-type: direct:production
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
 2024-11-25 12:54:55 +00:00
Meliksah Turker
c50b5675d6 prepare_fa2_from_position_ids function bugfix (#33269) 
contiguous() is called before view() for key and value within prepare_fa2_from_position_ids function
 2024-11-25 13:51:26 +01:00
VictorAtIfInsurance
a0f4f3174f allow unused input parameters passthrough when chunking in asr pipelines (#33889) 
* allow unused parameter passthrough when chunking in asr pipelines

* format code

* format

* run fixup

* update tests

* update parameters to pipline in test

* updates parametrs in tests

* change spelling in gitignore

* revert .gitignore to main

* add git ignore of devcontainer folder

* assert asr output follows expected inference output type

* run fixup

* Remove .devcontainer from .gitignore

* remove compliance check
 2024-11-25 11:36:44 +01:00
kang sheng
4dc1a69349 Sum gathered input tokens (#34554) 
* sum gathered input tokens

* ruff line-length is 119, format the code

---------

Co-authored-by: kangsheng <kangsheng@meituan.com>
 2024-11-25 11:27:13 +01:00
Raushan Turganbay
1e492afd61 🔴 Mllama: fix base prefix (#34874) 
fix base prefix
 2024-11-25 11:20:20 +01:00
Arthur
857d46ca0c [Deberta/Deberta-v2] Refactor code base to support compile, export, and fix LLM (#22105) 
* some modification for roadmap

* revert some changes

* yups

* weird

* make it work

* sttling

* fix-copies

* fixup

* renaming

* more fix-copies

* move stuff around

* remove torch script warnings

* ignore copies

* revert bad changes

* woops

* just styling

* nit

* revert

* style fixup

* nits configuration style

* fixup

* nits

* will this fix the tf pt issue?

* style

* ???????

* update

* eval?

* update error message

* updates

* style

* grumble grumble

* update

* style

* nit

* skip torch fx tests that were failing

* style

* skip the failing tests

* skip another test and make style
 2024-11-25 10:43:16 +01:00
Raushan Turganbay
098962dac2 BLIP: fix generation after hub update (#34876) 
* fix blip generation

* dont remove it yet

* Update src/transformers/models/blip_2/modeling_blip_2.py

Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

* address comments

* modular

---------

Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
 2024-11-25 10:41:55 +01:00
Raushan Turganbay
c1a8520419 Cache: init empty cache when use_cache (#34274) 
* fix

* fix tests

* fix copies

* add docs

* Revert "add docs"

This reverts commit 32d35634f12ba02781d2ebdee0c8dcfbe992a7b9.

* qwen move deltas

* mllama can potentiall fullgraph compile

* enable mllama compile and fix tests

* remove mllama fixes
 2024-11-25 10:11:33 +01:00
Dmitry Rogozhkin
1339a14dca Add safe_globals to resume training on PyTorch 2.6 (#34632) 
Starting from version 2.4 PyTorch introduces a stricter check for the objects which
can be loaded with torch.load(). Starting from version 2.6 loading with weights_only=True
requires allowlisting of such objects.

This commit adds allowlist of some numpy objects used to load model checkpoints.
Usage is restricted by context manager. User can still additionally call
torch.serialization.add_safe_globals() to add other objects into the safe globals list.

Accelerate library also stepped into same problem and addressed it with PR-3036.

Fixes: #34631
See: https://github.com/pytorch/pytorch/pull/137602
See: https://pytorch.org/docs/stable/notes/serialization.html#torch.serialization.add_safe_globals
See: https://github.com/huggingface/accelerate/pull/3036

Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
 2024-11-25 10:03:43 +01:00
jeongin601
318fe25f22 Fix: Enable prefill phase key value caching of nemotron/minitron models (#34742) 
* modeling nemotron kv caching bugfix

Signed-off-by: jeongin601 <0200angela@gmail.com>

* test file deleted

Signed-off-by: jeongin601 <0200angela@gmail.com>

* code refinement

Signed-off-by: jeongin601 <0200angela@gmail.com>

* remove unused variables

Signed-off-by: jeongin601 <0200angela@gmail.com>

* import block sorted

* removed deprecation warning

Signed-off-by: jeongin601 <0200angela@gmail.com>

* removed support for tuple shape past_key_values

Signed-off-by: jeongin601 <0200angela@gmail.com>

* Update conditional statement for cache initialization

Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

---------

Signed-off-by: jeongin601 <0200angela@gmail.com>
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
 2024-11-25 09:45:35 +01:00
Yoni Gozlan
3a8eb74668 Fix support for image processors modifications in modular (#34866) 
* add fix and examples

* fix camel case naming
 2024-11-22 18:14:24 -05:00
Mohamed Mekkouri
54be2d7ae8 Bitnet test fix to avoid using gated model (#34863) 
small test fix
 2024-11-22 17:18:49 +01:00
Benjamin Bossan
286ffaaf0a [CI] Skip EETQ tests while package is broken with latest transformers (#34854) 
* CI Skip EETQ tests while package is broken

EETQ tries to import the shard_checkpoint function from transformers but
the function has been removed. Therefore, trying to use EETQ currently
results in an import error. This fix results in EETQ tests being skipped
if there is an import error.

The issue has been reported to EETQ:

https://github.com/NetEase-FuXi/EETQ/issues/34

* Raise helpful error when trying to use eetq

* Forget to raise the error in else clause
 2024-11-22 17:13:30 +01:00
Andrés Marafioti
861758e235 smol improvements to support more flexible usage (#34857) 
* smol improvements to support more flexible usage

* ruff
 2024-11-22 16:34:38 +01:00
Nadav Timor
42b36d7395 Speculative decoding: Test the target distribution (to prevent issues like #32867) (#34553) 
* Update test_utils.py

* formatting

* Update test_utils.py

* formatting

* formatting

* Update test_utils.py

* formatting

* Update test_utils.py

* formatting

* format

* comments at standard positions
 2024-11-22 16:02:37 +01:00
Arthur
597efd21d2 Auto compile when static cache (#34247) 
* generate with compile

* nits

* simple

* generate with compile

* nits

* simple

* safe

* style

* Update src/transformers/generation/utils.py

Co-authored-by: Cyril Vallez <cyril.vallez@huggingface.co>

* remove TOKENIZER forked warning

---------

Co-authored-by: Cyril Vallez <cyril.vallez@huggingface.co>
 2024-11-22 15:33:35 +01:00
Konrad Kalita
d9e6f307e7 Remove quantization related config from dequantized model (#34856) 
* Remove quantization related config from dequantized model

* Fix whitespace
 2024-11-22 10:06:29 +01:00
Logan Adams
1867be666d Update checks for torch.distributed.tensor to require torch >= 2.5 (#34816) 
* Update checks for torch.distributed.tensor

* Update PR with feedback

* Formatting fix for import order

* Remove unused function
 2024-11-22 10:05:26 +01:00
Raushan Turganbay
6a912ff2c5 Watermarking: fix order (#34849) 
fix watermarking order
 2024-11-22 08:25:14 +01:00
Cyril Vallez
4e90b99ed9 Refactor StarCoder2 using modular (#34015) 
* Create modular_starcoder2.py

* Update modular_starcoder2.py

* update

* finalize modular

* revert # no-unravel

* Add support

* style

* Update modular_model_converter.py

* update docstring
 2024-11-21 14:52:39 +01:00
Jonathan Mamou
18871599c9 Fix heuristic scheduling for UAG (#34805) 
* fix heuristic schedule

* fix style

* fix format
 2024-11-21 14:46:35 +01:00
AbdelKarim ELJANDOUBI
d6a5c23f71 Fix ds nvme (#34444) 
* skip nested deepspeed.zero.Init call

* make fixup

* solve conflict

* solve conflict

* put back local

* use context mangers instead of local thread

* Skip recursive calls to deepspeed.zero.Init

* Skip recursive calls to deepspeed.zero.Init

* back to old notebooks

* make style
 2024-11-21 13:52:22 +01:00
Vladislav Bronzov
ae5cbf804b Improve gguf tensor processing (#34515) 
* add tensor processing system to separate logic for models

* format refactoring

* small fix

* make some methods private

* move custom methods to processors

* refactor tensor processing

* format fix
 2024-11-21 13:40:49 +01:00
farrosalferro
c57eafdaa1 Add Nemotron GGUF Loading Support (#34725) 
* Add Nemotron GGUF Loading Support

* fix the Nemotron architecture assignation

---------

Co-authored-by: Marc Sun <57196510+SunMarc@users.noreply.github.com>
 2024-11-21 11:37:34 +01:00
Quentin Gallouédec
d4e1acbb7c Change logging level from warning to info for max_steps overriding num_train_epochs (#34810) 
Update trainer.py
 2024-11-21 11:37:02 +01:00
Raushan Turganbay
28fb02fc05 VLMs: enable generation tests - last batch (#34484) 
* add tests for 3 more vlms

* fix fuyu back

* skip test
 2024-11-21 11:00:22 +01:00
Yih-Dar
40821a2478 Fix CI slack reporting issue (#34833) 
* fix

* fix

* fix

* fix

* fix

---------

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
 2024-11-20 21:36:13 +01:00
Marc Sun
3cb8676a91 Fix CI by tweaking torchao tests (#34832) 2024-11-20 20:28:51 +01:00
Corentin Royer
bf42c3bd4b Fix hyperparameter search when optuna+deepseed (#34642) 
* Fix hyperparameter search when optuna+deepseed

* Adding free_memory to the search setup

---------

Co-authored-by: Corentin-Royer <corentin.royer@ibm.com>
 2024-11-20 18:02:58 +01:00
Marc Sun
67890de3b8 Torchao weights only + prequantized compability (#34355) 
* weights only compability

* better tests from code review

* ping torch version

* add weights_only check
 2024-11-20 17:24:45 +01:00
Tibor Reiss
f297af55df Fix: take into account meta device (#34134) 
* Do not load for meta device

* Make some minor improvements

* Add test

* Update tests/utils/test_modeling_utils.py

Update test parameters

Co-authored-by: Marc Sun <57196510+SunMarc@users.noreply.github.com>

* Make the test simpler

---------

Co-authored-by: Marc Sun <57196510+SunMarc@users.noreply.github.com>
 2024-11-20 11:32:07 +01:00
Phillip Kuznetsov
8cadf76e1c fix(DPT,Depth-Anything) torch.export (#34103) 
* Fix torch.export issue in dpt based models

Signed-off-by: Phillip Kuznetsov <philkuz@gimletlabs.ai>

* Simplify the if statements

Signed-off-by: Phillip Kuznetsov <philkuz@gimletlabs.ai>

* Move activation definitions of zoe_depth to init()

Signed-off-by: Phillip Kuznetsov <philkuz@gimletlabs.ai>

* Add test_export for dpt and zoedepth

Signed-off-by: Phillip Kuznetsov <philkuz@gimletlabs.ai>

* add depth anything

Signed-off-by: Phillip Kuznetsov <philkuz@gimletlabs.ai>

* Remove zoedepth non-automated zoedepth changes and zoedepth test

Signed-off-by: Phillip Kuznetsov <philkuz@gimletlabs.ai>

* [run_slow] dpt, depth_anything, zoedepth

Signed-off-by: Phillip Kuznetsov <philkuz@gimletlabs.ai>

---------

Signed-off-by: Phillip Kuznetsov <philkuz@gimletlabs.ai>
 2024-11-20 11:31:21 +01:00
kjohew
9d16441e4f Fix the memory usage issue of logits in generate() (#34813) 2024-11-20 11:25:37 +01:00
Raushan Turganbay
9470d65324 Fix low memory beam search (#34746) 
* fix

* higher max positions in tests
 2024-11-20 07:46:35 +01:00
Raushan Turganbay
145fbd46cb LLaVA OV: fix unpadding precision (#34779) 
* fix

* propagate

* type check
 2024-11-20 07:46:13 +01:00
wwwbai
3033509327 Translate attention.md into Chinese (#34716) 
* try

* tryagain

* tryagggain

* translated

* translated2

* Update docs/source/zh/attention.md

Co-authored-by: Huazhong Ji <hzji210@gmail.com>

---------

Co-authored-by: Huazhong Ji <hzji210@gmail.com>
 2024-11-19 10:03:12 -08:00
Merve Noyan
befbbf2f98 Added image-text-to-text pipeline to task guide (#34783) 
* Added image-text-to-text pipeline to task guide

* Update docs/source/en/tasks/image_text_to_text.md

Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>

* Update docs/source/en/tasks/image_text_to_text.md

Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>

* Update docs/source/en/tasks/image_text_to_text.md

Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>

* Update docs/source/en/tasks/image_text_to_text.md

Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>

* Merge codeblocks

---------

Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com>
 2024-11-19 09:49:10 -08:00
Yih-Dar
469eddbe2d Fix check_training_gradient_checkpointing (#34806) 
fix

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
 2024-11-19 17:48:34 +01:00
Yih-Dar
05ebe8b9b0 Run test_medium_seamless_m4t_pt in subprocess to avoid many failures (#34812) 
* fix

* fix

* fix

---------

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
 2024-11-19 17:32:10 +01:00
Yoni Gozlan
eedc113914 Add Image Processor Fast Deformable DETR (#34353) 
* add deformable detr image processor fast

* add fast processor to doc

* fix copies

* nit docstring

* Add tests gpu/cpu and fix docstrings

* fix docstring

* import changes from detr

* fix imports

* rebase and fix

* fix input data format change in detr and rtdetr fast
 2024-11-19 11:18:58 -05:00
Yoni Gozlan
b99ca4d28b Add support for OpenAI api "image_url" input in chat for image-text-to-text pipeline (#34562) 
* add support for openai api image_url input

* change continue to elif

* Explicitely add support for OpenAI/TGI chat format

* rewrite content to transformers chat format and add tests

* Add support for typing of image type in chat templates

* add base64 to possible image types

* refactor nesting
 2024-11-19 11:08:37 -05:00
dependabot[bot]
15dd625a0f Bump aiohttp from 3.10.2 to 3.10.11 in /examples/research_projects/decision_transformer (#34792) 
Bump aiohttp in /examples/research_projects/decision_transformer

Bumps [aiohttp](https://github.com/aio-libs/aiohttp) from 3.10.2 to 3.10.11.
- [Release notes](https://github.com/aio-libs/aiohttp/releases)
- [Changelog](https://github.com/aio-libs/aiohttp/blob/master/CHANGES.rst)
- [Commits](https://github.com/aio-libs/aiohttp/compare/v3.10.2...v3.10.11)

---
updated-dependencies:
- dependency-name: aiohttp
  dependency-type: direct:production
...

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
 2024-11-19 16:08:07 +00:00
Wang, Yi
dc42330388 fix crash in tiiuae/falcon-11B-vlm image-to-text generation (#34728) 
Signed-off-by: Wang, Yi <yi.a.wang@intel.com>
 2024-11-19 16:51:32 +01:00
David Zhang
427b62ed1a Fix post process function called in the instance segmentation example of mask2former (#34588) 
* Fix post process function called in the instance segmentation example of mask2former

* fix description and additional notes for post_process_instance_segmentation of maskformers

* remove white space in maskformers post_process_instance_segmentation doc

* change image.size[::-1] to height and width for clarity in segmentation examples
 2024-11-19 16:49:25 +01:00
jp
fdb9230485 Add do_convert_rgb to vit (#34523) 
* Add: do_convert_rgb

* Add: doc string

* Update src/transformers/models/vit/image_processing_vit.py

Co-authored-by: Pavel Iakubovskii <qubvel@gmail.com>

* Update src/transformers/models/vit/image_processing_vit.py

Co-authored-by: Pavel Iakubovskii <qubvel@gmail.com>

* Update src/transformers/models/vit/image_processing_vit.py

Co-authored-by: Pavel Iakubovskii <qubvel@gmail.com>

* Add: do_convert_rgb to fast

* Add: convert_to_rgb

---------

Co-authored-by: Pavel Iakubovskii <qubvel@gmail.com>
 2024-11-19 16:48:05 +01:00
Tibor Reiss
7b9e51c1a0 Feature: print tokens per second during training (#34507) 
* Log tokens per second during training

* Nitpicks

* Move logic into _maybe_log_save_evaluate

* Use speed_metrics
 2024-11-19 16:46:04 +01:00
Phillip Kuznetsov
5fa4f64605 🚨🚨🚨 fix(Mask2Former): torch export 🚨🚨🚨 (#34393) 
* fix(Mask2Former): torch export

Signed-off-by: Phillip Kuznetsov <philkuz@gimletlabs.ai>

* revert level_start_index and create a level_start_index_list

Signed-off-by: Phillip Kuznetsov <philkuz@gimletlabs.ai>

* Add a comment to explain the level_start_index_list

Signed-off-by: Phillip Kuznetsov <philkuz@gimletlabs.ai>

* Address comment

Signed-off-by: Phillip Kuznetsov <philkuz@gimletlabs.ai>

* add torch.export.export test

Signed-off-by: Phillip Kuznetsov <philkuz@gimletlabs.ai>

* rename arg

Signed-off-by: Phillip Kuznetsov <philkuz@gimletlabs.ai>

* remove spatial_shapes

Signed-off-by: Phillip Kuznetsov <philkuz@gimletlabs.ai>

* Use the version check from pytorch_utils

Signed-off-by: Phillip Kuznetsov <philkuz@gimletlabs.ai>

* [run_slow] mask2former

Signed-off-by: Phillip Kuznetsov <philkuz@gimletlabs.ai>

---------

Signed-off-by: Phillip Kuznetsov <philkuz@gimletlabs.ai>
 2024-11-19 16:44:53 +01:00
huismiling
581524389a MLU devices : Checks if mlu is available via an cndev-based check which won't trigger the drivers and leave mlu (#34326) 
* add Cambricon MLUs support

* fix mlu device rng state

* up for quality check

* up mlu to support fp16

* fix mlu device dependency error

* fix mlu device dependency error

* enable mlu device for bf16

* fix mlu device memory tracker

* Cambricon support SDPA and flash_attn

* MLU devices : Checks if `mlu` is available via an `cndev-based` check which won't trigger the drivers and leave mlu
 2024-11-19 16:37:39 +01:00
Cyril Vallez
e3a5889ef0 Modular fix (#34802) 
* Modular fix

* style

* remove logger warning

* Update modular_model_converter.py
 2024-11-19 16:08:57 +01:00
Marc Sun
ce1d328e3b Fix cache_utils for optimum.quanto kvcache quantization (#34750) 
* add co-author

Co-authored-by: w3rew <w3rew@users.noreply.github.com>

* fix docs

* fix cache

* remove print

---------

Co-authored-by: w3rew <w3rew@users.noreply.github.com>
 2024-11-19 14:16:34 +01:00