Hi, there, vllm-omni is trying to support quantization version of qwen3-omni, just wondering what's the accuracy of this fp8 ckpt?
· Sign up or log in to comment
