BetterTransformer support training & autocast for all archs #1225
support training
212aa3a8
encoders and encoder+decoder all work
abd89201
warning about training decoders with padding
db0c5618
leave to an other PR the backward for some archs
bac435b8
nit
d1f160a1
fix tests
c70a3dbb
hopefully tests pass
dd675955
fix
0fcdff81
fxmarty
merged
38061a66
into main 2 years ago
Assignees
No one assigned
Login to write a write a comment.
Login via GitHub