Skip to content

GitLab

"vscode:/vscode.git/clone" did not exist on "193fbb0bce05625f77cd60840adf4358dbc67449"
Unverified Commit 3d7cdb2e authored by fzyzcjy's avatar fzyzcjy Committed by GitHub
Browse files

Fix MTP error when enabling two-batch overlap (#7569)

parent 604efe07
Hide whitespace changes
Inline Side-by-side
Showing with 1 addition and 1 deletion
python/sglang/srt/model_executor/cuda_graph_runner.py
View file @ 3d7cdb2e
...@@ -168,7 +168,7 @@ def get_batch_sizes_to_capture(model_runner: ModelRunner): ...@@ -168,7 +168,7 @@ def get_batch_sizes_to_capture(model_runner: ModelRunner):
capture_bs += [model_runner.req_to_token_pool.size] capture_bs += [model_runner.req_to_token_pool.size]
if server_args.enable_two_batch_overlap: if server_args.enable_two_batch_overlap:
capture_bs = [bs for bs in capture_bs if bs >= 2] capture_bs = [bs for bs in capture_bs if bs % 2 == 0]
if server_args.cuda_graph_max_bs: if server_args.cuda_graph_max_bs:
capture_bs = [bs for bs in capture_bs if bs <= server_args.cuda_graph_max_bs] capture_bs = [bs for bs in capture_bs if bs <= server_args.cuda_graph_max_bs]
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment