younesbelkada/transformers

Pull Requests Commits

younesbelkada committed 2 years ago

d84bc577

younesbelkada committed 2 years ago

a4bad5c3

Merge remote-tracking branch 'upstream/main' into add-efficient-sam-clean

younesbelkada committed 2 years ago

6b5484a8

fix resuming from ckpt when using FSDP with FULL_STATE_DICT (#27891)

pacman100 committed 2 years ago

Verified 238d2e3c

[docs] MPS (#28016)

stevhliu committed 2 years ago

Verified ebfdb9ca

[docs] Trainer (#27986)

stevhliu committed 2 years ago

Verified 0d63d177

Fix Vip-llava docs (#28085)

younesbelkada committed 2 years ago

Verified 1faeff85

Fix wrong examples in llava usage. (#28020)

Lyken17 committed 2 years ago

Verified ffa04def

Fix `low_cpu_mem_usage` Flag Conflict with DeepSpeed Zero 3 in `from_pretrained` for Models with `keep_in_fp32_modules`" (#27762)

kotarotanahashi committed 2 years ago

Verified 29a1c1b4

Update fixtures-image-utils (#28080)

lhoestq committed 2 years ago

Verified 26ea725b

Fix bug for checkpoint saving on multi node training setting (#28078)

dumpmemory committed 2 years ago

Verified 1c286be5

make torch.load a bit safer (#27282)

julien-c committed 2 years ago

Verified dec84b32

Make GPT2 traceable in meta state (#28054)

kwen2501 committed 2 years ago

Verified 74cae670

[LLaVa] Add past_key_values to _skip_keys_device_placement to fix multi-GPU dispatch (#28051)

aismlv committed 2 years ago

Verified e2b6df79

Skip M4T `test_retain_grad_hidden_states_attentions` (#28060)

ylacombe committed 2 years ago

Verified deb72cb6

[`Mixtral`] update conversion script to reflect new changes (#28068)

younesbelkada committed 2 years ago

Verified d269c4b2

doc: Correct spelling mistake (#28064)

caiyili committed 2 years ago

Verified 70a127a3

Remove SpeechT5 deprecated argument (#28062)

ylacombe committed 2 years ago

Verified c817c17d

[Flax LLaMA] Fix attn dropout (#28059)

sanchit-gandhi committed 2 years ago

Verified 6af3ce77

[Flax BERT] Update deprecated 'split' method (#28012)

sanchit-gandhi committed 2 years ago

Verified 7e876dca

[`Modeling` / `Mixtral`] Fix GC + PEFT issues with Mixtral (#28061)

younesbelkada committed 2 years ago

Verified e737446e

[`FA-2`] Fix fa-2 issue when passing `config` to `from_pretrained` (#28043)

younesbelkada committed 2 years ago

Verified 1e209317

Remove warning when Annotion enum is created (#28048)

amyeroberts committed 2 years ago

Verified 1a585c12

younesbelkada committed 2 years ago

169a7142

add new model-like

younesbelkada committed 2 years ago

8ee6712f

Replace build() with build_in_name_scope() for some TF tests (#28046)

Rocketknight1 committed 2 years ago

Verified 3060899b

Proper build() methods for TF (#27794)

Rocketknight1 committed 2 years ago

Verified 050e0b44

[Seamless] Fix links in docs (#27905)

sanchit-gandhi committed 2 years ago

Verified 52c37882

Generate: Mistral/Mixtral FA2 cache fix when going beyond the context window (#28037)

gante committed 2 years ago

Verified 388fd314

Fixed spelling error in T5 tokenizer warning message (s/thouroughly/t… (#28014)

jeddobson committed 2 years ago

Verified 0ede7626

Older