<h2 id="base-minimaxh3-(turbo-merge)">Base MinimaxH3 (Turbo Merge)</h2>This model is unchanged/untrained from the base model listed. The model was quantized from the source FP32/BF16 model. BOTH THE FL2VA and REF2VA models are 8 step models.<h3 id="requirements:">Requirements:</h3><ol><li>Update <a target="_blank" rel="ugc" href="https://developer.nvidia.com/cuda-downloads">CUDA to 13.4.2</a></li><li>Update <a target="_blank" rel="ugc" href="https://pytorch.org/">Pytorch to 13.2</a></li><li>Update at minimum comfy-aimdo, comfy-kitchen</li><li>Basic <a target="_blank" rel="ugc" href="https://civarchive.com/api/download/models/3337482?fileId=3224058">workflow</a> shows how to use both first frame text guided and first frame last frame with comfy kitchen backend node to speed up generation by 60% (This prevents pyattention fallback which is slow)</li></ol>pytorch version: 2.13.0+cu132xformers version: 0.0.35Using xformers attentionComfyUI version: 0.35.0comfy-aimdo version: 0.5.5comfy-kitchen version: 0.2.34Consider --disable-dynamic-vram if you are having OOM issues after a few generation or crash when trying to use comfy kitchen vs pyattentionNote: INT Convrot Group size is 64 vs 256 on the comfy supplied modelsNF4 version of the reference to video/audio model works but the quality is very low

Base MinimaxH3 (Turbo Merge 8-Step)

baseMinimaxh3TurboMerge_comfyRef2vaPruned_mxfp8.safetensors

baseMinimaxh3TurboMerge_comfyRef2vaPruned_nvfp4.safetensors

Files with SHA256: ac9dc5a105b9e84ec2ef65f8459809588d21cab85c74c4b11b4b0ab99dc85740