Making large AI models cheaper, faster and more accessible
[shardformer]fix flash attention, when mask is casual, just don't unpad it (#5084)
* fix flash attn * fix fix
F
flybird11111 committed
aae496631c4acac03d8707216958184290a018fd
Parent: 75af66c
Committed by GitHub <noreply@github.com>
on 11/22/2023, 8:00:07 AM