xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-13 22:45:50 +08:00
Code Issues Packages Projects Releases Wiki Activity

[Bugfix] Fix sync_and_slice_intermediate_tensors (#21537)

Browse Source 
Signed-off-by: Rui Qiao <ruisearch42@gmail.com>
This commit is contained in:
Rui Qiao 2025-07-25 17:07:58 -07:00 committed by GitHub
parent 2eddd437ba
commit cea96a0156
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
vllm/v1/worker/gpu_model_runner.py
View File

@ -1270,7 +1270,7 @@ class GPUModelRunner(LoRAModelRunnerMixin, KVConnectorModelRunnerMixin):
if sync_self: if sync_self:
assert intermediate_tensors is not None assert intermediate_tensors is not None
for k, v in intermediate_tensors.items(): for k, v in intermediate_tensors.items():
is_scattered = "residual" and is_residual_scattered is_scattered = k == "residual" and is_residual_scattered
copy_len = num_tokens // tp if is_scattered else \ copy_len = num_tokens // tp if is_scattered else \
num_tokens num_tokens
self.intermediate_tensors[k][:copy_len].copy_( self.intermediate_tensors[k][:copy_len].copy_(