github/ColossalAI
1
0
Fork 0
mirror of https://github.com/hpcaitech/ColossalAI.git synced 2025-06-21 21:22:04 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix

Browse Source
This commit is contained in:
wangbluo 2024-10-14 18:07:56 +08:00
parent 3201377e94
commit fe9208feac
1 changed files with 1 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
colossalai/shardformer/layer/attn.py
View File

@ -623,7 +623,6 @@ class RingAttention(torch.autograd.Function):
is_packed: Optional[bool] = False,
inner_ring_group: Optional[dist.ProcessGroup] = None,
inter_ring_group: Optional[dist.ProcessGroup] = None,
tp_group: Optional[dist.ProcessGroup] = None,
):
cu_seqlens_q = cu_seqlens_kv = cu_seqlens
@ -1120,7 +1119,7 @@ class RingAttention(torch.autograd.Function):
if not is_packed:
dq, dk, dv = [x.view(b, sq, *x.shape[-2:]) for x in (dq, dk, dv)]
return (dq, dk, dv, None, None, None, None, None, None, None, None, None, None, None, None)
return (dq, dk, dv, None, None, None, None, None, None, None, None, None, None, None)
@staticmethod
def prepare_varlen_batch(