Manuel de Prada Corral c338fd43b0 [cache refactor] Move all the caching logic to a per-layer approach (#39106) ...

* Squash for refactor: Replace monolithic cache classes with modular LayeredCache (#38077)

- Introduces CacheLayer and Cache base classes
- Ports Static, Dynamic, Offloaded, Quantized, Hybrid, etc. to use layers
- Implements method/attr dispatch across layers to reduce boilerplate
- Adds CacheProcessor hooks for offloading, quantization, etc.
- Updates and passes tests

* fix quantized, add tests

* remove CacheProcessorList

* raushan review, arthur review

* joao review: minor things

* remove cache configs, make CacheLayer a mixin (joaos review)

* back to storage inside Cache()

* remove cachebase for decorator

* no more __getattr__

* fix tests

* joaos review except docs

* fix ast deprecations for python 3.14: replace node.n by node.value and use `ast.Constant`

More verbose exceptions in `fix_docstring` on docstring formatting issues.

* Revert "back to storage inside Cache()"

This reverts commit 27916bc2737806bf849ce2148cb1e66d59573913.

* cyril review

* simplify cache export

* fix lfm2 cache

* HybridChunked to layer

* BC proxy object for cache.key_cache[i]=...

* reorder classes

* bfff come on LFM2

* better tests for hybrid and hybridChunked

* complete coverage for hybrid chunked caches (prefill chunking)

* reimplementing HybridChunked

* cyril review

* fix ci

* docs for cache refactor

* docs

* oopsie

* oopsie

* fix after merge

* cyril review

* arthur review

* opsie

* fix lfm2

* opsie2

2025-07-22 16:10:25 +02:00

..

bettertransformer

Use Python 3.9 syntax in tests (#37343)

2025-04-08 14:12:08 +02:00

commands

[serve] Add speech to text (/v1/audio/transcriptions) (#39434)

2025-07-17 14:29:57 +00:00

deepspeed

Remove script datasets in tests (#38940)

2025-06-25 14:31:20 +00:00

extended

Add Optional to remaining types (#37808)

2025-04-28 14:20:45 +01:00

fixtures

[tests] remove tests from libraries with deprecated support (flax, tensorflow_text, ...) (#39051)

2025-06-26 16:25:00 +01:00

fsdp

Gaudi3 CI (#38790)

2025-06-23 10:56:51 +02:00

generation

[cache refactor] Move all the caching logic to a per-layer approach (#39106)

2025-07-22 16:10:25 +02:00

models

[cache refactor] Move all the caching logic to a per-layer approach (#39106)

2025-07-22 16:10:25 +02:00

optimization

[tests] remove TF tests (uses of require_tf) (#38944)

2025-06-25 17:29:10 +00:00

peft_integration

Enable some ruff checks for performance and readability (#39383)

2025-07-17 13:21:59 +00:00

pipelines

Raise TypeError instead of ValueError for invalid types (#38660)

2025-07-21 12:42:00 +00:00

quantization

Remove residual quantization attribute from dequantized models (#39373)

2025-07-15 17:16:10 +02:00

repo_utils

[modular] speedup check_modular_conversion with multiprocessing (#37456)

2025-07-10 19:07:59 +01:00

sagemaker

Deprecate TF + JAX (#38758)

2025-06-11 17:28:06 +01:00

tensor_parallel

enable static cache on TP model (#39164)

2025-07-09 21:14:45 +00:00

tokenization

[tests] remove tests from libraries with deprecated support (flax, tensorflow_text, ...) (#39051)

2025-06-26 16:25:00 +01:00

trainer

Fix tests due to breaking change in accelerate (#39451)

2025-07-17 13:51:50 +01:00

utils

[cache refactor] Move all the caching logic to a per-layer approach (#39106)

2025-07-22 16:10:25 +02:00

__init__.py

…

causal_lm_tester.py

[Ernie 4.5] Add ernie text models (#39228)

2025-07-21 19:51:49 +02:00

test_backbone_common.py

Use Python 3.9 syntax in tests (#37343)

2025-04-08 14:12:08 +02:00

test_configuration_common.py

Update composition flag usage (#36263)

2025-04-09 11:48:49 +02:00

test_feature_extraction_common.py

Use Python 3.9 syntax in tests (#37343)

2025-04-08 14:12:08 +02:00

test_image_processing_common.py

Fix overriding Fast Image/Video Processors instance attributes affect other instances (#39363)

2025-07-12 23:39:06 +00:00

test_image_transforms.py

Raise TypeError instead of ValueError for invalid types (#38660)

2025-07-21 12:42:00 +00:00

test_modeling_common.py

Kernels flash attn (#39474)

2025-07-22 15:41:06 +02:00

test_pipeline_mixin.py

Enable some ruff checks for performance and readability (#39383)

2025-07-17 13:21:59 +00:00

test_processing_common.py

[chat template] return assistant mask in processors (#38545)

2025-07-18 12:23:20 +00:00

test_sequence_feature_extraction_common.py

[tests] remove TF tests (uses of require_tf) (#38944)

2025-06-25 17:29:10 +00:00

test_tokenization_common.py

Fix pylint warnings (#39477)

2025-07-21 12:38:05 +00:00

test_tokenization_mistral_common.py

Add voxtral (#39429)

2025-07-18 00:02:04 +00:00

test_training_args.py

Fix TrainingArguments.torch_empty_cache_steps post_init check (#36734)

2025-03-17 16:09:46 +01:00

test_video_processing_common.py

Fix overriding Fast Image/Video Processors instance attributes affect other instances (#39363)

2025-07-12 23:39:06 +00:00