huggingface/accelerate

Pull Requests Commits

Fix torch>=2.11 DeviceMesh deprecation warning (bookkeep flattened submeshes)

qgallouedec committed 75 days ago

bdca9b74

SunMarc committed 82 days ago

1d74e6dd

Fix flaky behavior in gated-model CLI test by making the test deterministic and independent of local Hugging Face cache state. (#4073)

sywangyi committed 89 days ago

Verified fb774426

SunMarc committed 90 days ago

67bf4d64

Fix all tests (#4072)

SunMarc committed 91 days ago

Verified a5cb37c1

Support dynamic batch size in BatchSamplerShard with even_batches (#3969)

yuxinyuan committed 91 days ago

Verified 14d7cdc7

fix(tracker): guard init_trackers and log against None kwargs (#4026)

xodn348 committed 92 days ago

Verified 77eb5cf6

fix(security): remediate workflow vulnerability in .github/workflows/build-docker-images-release.yml (#4069)

hf-security-analysis[bot] committed 92 days ago

Verified 379c8e00

Fix to-fsdp2: actually drop REMOVED / NOT_YET_IMPLEMENTED FSDP1 keys (#4065)

lollinng committed 92 days ago

Verified 61e64209

tests: Gracefully handle missing set_device for mps (#4028)

booxter committed 92 days ago

Verified b120c1f5

test: add regression test for no_split_module_classes accepting set type (#4048)

UFO0506 committed 92 days ago

Verified 418015aa

Bump the actions group with 8 updates (#4068)

dependabot[bot] committed 92 days ago

Verified 3f1acf85

docs: add aggregate profiler memory example (#4054)

aryanputta committed 92 days ago

Verified a2e5292f

Raise ValueError (not a bare string) in ParallelismConfig.get_device_mesh (#4064)

lollinng committed 92 days ago

Verified bd208073

DOC: document missing parameters in load_accelerator_state, find_executable_batch_size, and send_to_device (#4051)

kratos0718 committed 92 days ago

Verified 6fbd0ae2

chore: enable Dependabot weekly GitHub Actions bumps (#4049)

hf-dependantbot-rollout[bot] committed 92 days ago

Verified 4929d1ec

Fix MLflowTracker.store_init_configuration mutating the caller's config dict (#4046)

ATOM00blue committed 92 days ago

Verified fcb01d16

Fix implicit padding in `split_between_processes` when `apply_padding=False` and `num_samples < num_processes` (#4052)

3manifold committed 92 days ago

Verified bb803daf

docs: fix typos in docstrings, comments, and user docs (#4040)

mokashang committed 92 days ago

Verified 0413cf9f

Fix dtype_byte_size for FP8 fnuz / e8m0fnu dtypes (#4063)

lollinng committed 92 days ago

Verified 0dca40ab

fix(tracking): default step=None on tracker.log and accept extra kwargs in MLflowTracker (#4039)

1fanwang committed 92 days ago

Verified 27a142bc

fix(state): make MLU backend part of the _prepare_backend elif chain (#4057)

Anai-Guo committed 92 days ago

Verified 6b6d0ec6

fix notebook launcher cuda init (#4059)

SunMarc committed 92 days ago

Verified 1afc3d8c

Make accelerate work end-to-end on AMD ROCm (#4025)

Abdennacer-Badaoui committed 99 days ago

Verified 8d079672

Fix gloo backend error in test_load_checkpoint_and_dispatch_with_broadcast on XPU (#4056)

kaixuanliu committed 100 days ago

Verified 4ad096a3

Fix int8 offload hook detachment statistics restoration (#4044)

jiqing-feng committed 104 days ago

Verified 507ea7b2

feat(utils): add max reduction type (#4027)

imstevenpmwork committed 114 days ago

Verified cde3e58b

fix parameter fp32 (#4043)

jiqing-feng committed 114 days ago

Verified e1d39a3b

docs: Fix docstring of `fsdp2_prepare_auto_wrap_policy` (#4037)

slocoro committed 114 days ago

Verified 62ceff72

fix: add missing neuron device case (#4042)

michaelbenayoun committed 115 days ago

Verified 1fab6edd

Older