xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-01-07 18:43:04 +08:00
Code Issues Packages Projects Releases Wiki Activity

[BugFix] Illegal memory access for MoE On H20 (#13693)

Browse Source
This commit is contained in:
Zhonghua Deng 2025-02-24 23:37:32 +08:00 committed by GitHub
parent 781096e385
commit ccc00515fd
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/model_executor/layers/fused_moe/fused_moe.py
View File

@ -1271,7 +1271,7 @@ def fused_experts_impl(hidden_states: torch.Tensor,
# so the cache size and config are already set correctly and
# do not need to be adjusted.
intermediate_cache1 = intermediate_cache1[:tokens_in_chunk]
intermediate_cache2 = intermediate_cache2[:tokens_in_chunk]
intermediate_cache2 = intermediate_cache2[:tokens_in_chunk * topk_ids.shape[1]]
intermediate_cache3 = intermediate_cache3[:tokens_in_chunk]
config = get_config_func(tokens_in_chunk)