fix GLM4_MOE launch with compressed_tensor quant model#8456

zhyncs merged 1 commit intosgl-project:mainsgl-project/sglang:mainfrom

zminglei:fix-compressed_tensor_quantzminglei/sglang:fix-compressed_tensor_quantCopy head branch name to clipboard