Hyper-AI/Qwen3.5-9B-fp8
Image-Text-to-Text • 10B • Updated • 323 • 1
qwen3.5-fp8
fp8 quant for qwen3.5 models, nearly half memory decrease, speedup 30%, vllm serve can run
fp8 quant for qwen3.5 models, nearly half memory decrease, speedup 30%, vllm serve can run