Qwen3.6-35B-A3B
Alibaba (Qwen)CUDA graph / Mamba cache size error: reduce --max-cudagraph-capture-size(default 512).
optionalinference servingin vLLMvLLM
1 more from this model
reduce --max-cudagraph-capture-size(default 512)
usedunclearin Qwen3.6-35B-A3BQwen Team