We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
There was an error while loading. Please reload this page.
1 parent dc687d6 commit 0fa2f9fCopy full SHA for 0fa2f9f
python/sglang/srt/model_executor/cuda_graph_runner.py
@@ -134,7 +134,8 @@ def get_batch_sizes_to_capture(model_runner: ModelRunner):
134
)
135
136
gpu_mem = get_device_memory_capacity()
137
- if gpu_mem is not None and gpu_mem > 81920:
+ # Batch size of each rank will not become so large when DP is on
138
+ if gpu_mem is not None and gpu_mem > 81920 and server_args.dp_size == 1:
139
capture_bs += list(range(160, 257, 8))
140
141
if max(capture_bs) > model_runner.req_to_token_pool.size:
0 commit comments