huggingface/accelerate

Pull Requests Commits

S1ro1 committed 132 days ago

db02e4df

Remove deprecated FindTiedParametersResult (#3786)

cyyever committed 136 days ago

Verified 14383311

Revert "fix: correct dictionary unpacking in recursively_apply function (#3766)" (#3787)

SunMarc committed 137 days ago

Verified a737437c

rm mlflow (#3783)

SunMarc committed 137 days ago

Verified 6997855a

Add optional typing (#3769)

cyyever committed 137 days ago

Verified 401075ff

refactor: Use `with` in Accelerator.autocast()instead of ` __enter__()` and` __exit__()` for more elegant style. (#3767)

EquationWalker committed 137 days ago

Verified 8031e24e

fix: correct dictionary unpacking in recursively_apply function (#3766)

qgallouedec committed 137 days ago

Verified 3db9fb69

switch XPU ccl backend to torch-builtin xccl in test_zero3_integration (#3773)

yao-matrix committed 138 days ago

Verified fe795fd3

Lower complexity of get_balanced_memory by adding a set (#3776)

SamuelBarryCS committed 138 days ago

Verified 409b356f

enable 2 model hook ut cases on XPU (#3774)

yao-matrix committed 139 days ago

Verified 1b50d939

Added Tip for better rendering (#3781)

sergiopaniego committed 140 days ago

Verified e79f3836

fix Muti node CUDA error: invalid device ordinal #3775 (#3779)

RicardoDominguez committed 142 days ago

Verified 0cb1a334

use reset_peak_memory_stats on xpu (#3772)

yao-matrix committed 143 days ago

Verified dfdc2190

fix FSDP2 test case failure on XPU (#3771)

yao-matrix committed 143 days ago

Verified 45959d7b

Fix: typo makes tests fail (#3765)

S1ro1 committed 147 days ago

Verified 8b493524

fix: torch_npu import error (#3764)

yanyongyu committed 147 days ago

Verified 9ead94e5

feat: allow mixed precision policy as dtype (#3751)

kmehant committed 147 days ago

Verified a0bc36e8

Fix typos (#3753)

cyyever committed 148 days ago

Verified 8830e58a

make torch_native_parallelism examples device agnostic (#3759)

yao-matrix committed 148 days ago

Verified 40ebb4be

fix: model.set_requires_gradient_sync(False) should be called to turn off gradient synchronization in FSDP2 (#3762)

EquationWalker committed 149 days ago

Verified ec92b1af

CP docs typos fixed (#3761)

sergiopaniego committed 151 days ago

Verified 62ede1ed

fix: specify device for process_tensor in example usage (#3755)

qgallouedec committed 153 days ago

Verified 9f9c490c

xpu INT64 all_gather issue fixed in 2.9 (#3756)

yao-matrix committed 153 days ago

Verified 8b55e62b

Add bf16/fp16 support for amp with mps device (#3373)

SunMarc committed 159 days ago

Verified 0e4419b3

Add support for TE MXFP8 recipe in accelerate (#3688)

pstjohn committed 160 days ago

Verified 3b67c216

[ND Parallel] Update examples, cleanup (#3737)

S1ro1 committed 161 days ago

Verified 7b981788

fix: specify device_ids in torch.distributed.barrier for PartialState (#3744)

qgallouedec committed 161 days ago

Verified c4460e33

Protect import for device_mesh (#3742)

SunMarc committed 164 days ago

Verified 5dd3d0b6

Feat: add to_json (#3743)

S1ro1 committed 164 days ago

Verified 5fe4460c

fix: cpu ram efficient loading for nd or hsdp parallelisms (#3740)

kmehant committed 166 days ago

Verified 979d81e4

Older