Skip to content

Commit ed79577

Browse files
committed
fix: run lmcache on dsv4 tep agentic
1 parent c149b9d commit ed79577

1 file changed

Lines changed: 1 addition & 2 deletions

File tree

.github/configs/nvidia-master.yaml

Lines changed: 1 addition & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -1781,8 +1781,7 @@ dsv4-fp4-b200-vllm-agentic:
17811781
# Experimental LMCache MP offload. LMCacheMPConnector currently requires
17821782
# vLLM's hybrid KV manager to be disabled, so this is not an HMA/CSA/HCA
17831783
# parity run against the no-offload path.
1784-
- { tp: 8, offloading: lmcache-mp, conc-list: [1, 2, 4, 8, 12, 16] }
1785-
- { tp: 8, ep: 8, dp-attn: true, offloading: none, conc-list: [12, 16, 24, 32, 48, 64] }
1784+
- { tp: 8, ep: 8, dp-attn: true, offloading: lmcache-mp, conc-list: [12, 16, 24, 32, 48, 64] }
17861785

17871786
dsv4-fp4-b200-trt:
17881787
image: ghcr.io#semianalysisai/trtllm-deepseek-v4:feat-deepseek_v4-9aa3715

0 commit comments

Comments
 (0)