huggingface/accelerate

Pull Requests Commits

Refactor code to be more readable

muellerzr committed 1 year ago

739f64a4

BenjaminBossan committed 1 year ago

7baade72

[WIP] FEAT Decorator to purge accelerate env vars

BenjaminBossan committed 1 year ago

523b31d7

remove hook for bnb 4-bit (#3223)

SunMarc committed 1 year ago

Verified 8ade23cc

Fix `align_module_device`, ensure only cpu tensors for `get_state_dict_offloaded_model` (#3217)

kylesayrs committed 1 year ago

Verified c0552c90

[Utils] `align_module_device` (#3204)

kylesayrs committed 1 year ago

Verified bf4572b6

Update big_modeling.py (#3207)

a-r-r-o-w committed 1 year ago

Verified a4a44aca

add xpu (#3163)

faaany committed 1 year ago

Verified b0e5fd35

Models With Tied Weights Need Re-Tieing After FSDP Param Init (#3154)

fabianlim committed 1 year ago

Verified 8159c98d

fix bug (#3166)

faaany committed 1 year ago

Verified 497eb3cf

take `torch.nn.Module` model into account when moving to device (#3167)

faaany committed 1 year ago

Verified 87732a4c

eliminate dead code (#3198)

ji-huazhong committed 1 year ago

Verified ffbca159

Update transformers.deepspeed references from transformers 4.46.0 release (#3196)

loadams committed 1 year ago

Verified ba7ab93f

🚨 🚨 🚨 Goodbye Python 3.8! 🚨 🚨 🚨 (#3194)

muellerzr committed 1 year ago

Verified 85f35647

update Megatron-LM plugin code to version 0.8.0 or higher. (#3174)

eljandoubi committed 1 year ago

Verified 2f39575b

MLU devices : Checks if mlu is available via an cndev-based check which won't trigger the drivers and leave mlu (#3187)

huismiling committed 1 year ago

Verified 1ace241d

Fix typo (#3191)

kylesayrs committed 1 year ago

Verified 78e1bdd0

[docs] use nn.module instead of tensor as model (#3157)

faaany committed 1 year ago

Verified 4dda5797

docs: fix a wrong word in comment in src/accelerate/accelerate.py:1255 (#3183)

Rebornix-zero committed 1 year ago

Verified 1f4fbb77

[docs] update neptune API (#3181)

faaany committed 1 year ago

Verified c809f8e4

fix bnb (#3186)

eljandoubi committed 1 year ago

Verified 39dc2b12

[Utils] `has_offloaded_params` (#3188)

kylesayrs committed 1 year ago

Verified 735dfa30

enable cpu bnb distributed lora finetune (#3159)

jiqing-feng committed 1 year ago

Verified a84327e5

fix version check bug in `get_xpu_available_memory` (#3165)

faaany committed 1 year ago

Verified 292954b5

Remove broken dynamo test (#3155)

oraluben committed 1 year ago

Verified 0e61127b

Trigger `weights_only=True` by default for all compatible objects (#3036)

muellerzr committed 1 year ago

Verified 6f79b63b

Fixup Zero3 + `save_model` (#3146)

muellerzr committed 1 year ago

Verified 1d2ca747

support torch dynamo for deepspeed>=0.14.4 (#3069)

oraluben committed 1 year ago

Verified cba3f2d5

Adding multi gpu speech generation (#3149)

dame-cell committed 1 year ago

Verified f1f2b4d1

POC: Allow for a `data_seed` (#3150)

muellerzr committed 1 year ago

Verified fd9880da

Older