transformers
[`Gemma`] final fixes to the modeling
#29729
Merged

[`Gemma`] final fixes to the modeling #29729

ArthurZucker merged 21 commits into main from fix-gemma-bis
ArthurZucker
danielhanchen gelu_pytorch_tanh
14645206
danielhanchen Force config.hidden_act to be approx gelu
606463fd
danielhanchen Merge branch 'huggingface:main' into main
16ed142b
danielhanchen Gemma bug fixes
03139e69
danielhanchen Merge branch 'huggingface:main' into main
ca3cae37
danielhanchen Merge branch 'huggingface:main' into main
73c24f61
danielhanchen force_use_exact_gelu
2b8c7f10
danielhanchen Update configuration_gemma.py
c1b8bef8
danielhanchen Update modeling_gemma.py
32656ccf
danielhanchen Merge branch 'huggingface:main' into main
9aa08bf7
ArthurZucker update
557c8fc1
ArthurZucker update for simpler handling
3d5abdbe
ArthurZucker nit
7b3e8e8f
ArthurZucker nit
667ca19f
ArthurZucker ArthurZucker requested a review from LysandreJik LysandreJik 2 years ago
ArthurZucker fixpup
98f54a07
HuggingFaceDocBuilderDev
ArthurZucker update
a54068ea
ArthurZucker also update the jax modeling!
496c8073
ArthurZucker add `"gelu_pytorch_tanh": partial(nn.gelu, approximate=True),`
9a52bbb0
ArthurZucker fixup
840b47f5
ArthurZucker fix order
8f0a9771
ArthurZucker
ArthurZucker commented on 2024-03-19
ArthurZucker act vs act_fn
e99ab0b7
younesbelkada
younesbelkada approved these changes on 2024-03-19
ArthurZucker ArthurZucker merged 8e2fc52e into main 2 years ago
ArthurZucker ArthurZucker deleted the fix-gemma-bis branch 2 years ago
ArthurZucker

Login to write a write a comment.

Login via GitHub

Assignees
No one assigned
Labels
Milestone