karylab_agents/vllm
1
0
Fork 0
forked from Karylab-cklius/vllm
Code Pull Requests 2 Actions 1 Packages Activity

Revert "Temporary disable persistent topk (#41442)"

Browse Source 
This reverts commit 05ebca5250.
This commit is contained in:
khluu
2026-05-02 00:52:06 -07:00
parent 87494544b1
commit a4debbd5bd
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
vllm/model_executor/layers/sparse_attn_indexer.py
+1 -1
View File
@@ -320,7 +320,7 @@ def sparse_attn_indexer(
num_rows = logits.shape[0]
topk_indices = topk_indices_buffer[:num_padded_tokens, :topk_tokens]
if current_platform.is_cuda() and topk_tokens in (512, 2048):
if current_platform.is_cuda() and topk_tokens in (512, 1024, 2048):
workspace_manager = current_workspace_manager()
(topk_workspace,) = workspace_manager.get_simultaneous(
((RADIX_TOPK_WORKSPACE_SIZE,), torch.uint8),