minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

Fix TRTLLM MHA routing for draft extend (#24856)

Browse Source
This commit is contained in:
YAMY
2026-05-12 15:48:32 -07:00
committed by GitHub
parent d5f3254ed1
commit a4109e87ac
1 changed files with 1 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files
python/sglang/srt/layers/attention/trtllm_mha_backend.py
+1 -4
View File
@@ -848,10 +848,7 @@ class TRTLLMHAAttnBackend(FlashInferAttnBackend):
page_table = self._get_layer_page_table(layer, forward_batch) page_table = self._get_layer_page_table(layer, forward_batch)
if ( if forward_batch.forward_mode.is_target_verify():
forward_batch.forward_mode.is_target_verify()
or forward_batch.forward_mode.is_draft_extend_v2()
):
o = flashinfer.decode.trtllm_batch_decode_with_kv_cache( o = flashinfer.decode.trtllm_batch_decode_with_kv_cache(
query=q, query=q,
kv_cache=kv_cache, kv_cache=kv_cache,