NEW v1.2 - Simpler and Cleaner with RefMods.Generate LOW RESOLUTION in a first Pass, then upscale and refine when happy with the low quality resolution results.Features:<ul><li>Add RefMods or image references.</li><li>Enable or disable the high-resolution upscaling and refinement in the 2nd Pass.</li><li>Keep the same seed to produce small videos first to check for audio quality and composition before enabling the 2nd pass to upscale to higher resolution</li><li>Add your desired Loras, individually to each pass to ensure the Lora weights are applied accordingly</li></ul>MODELS<ul><li>Fast H3 OmniRef Checkpoint: <a target="_blank" rel="ugc" href="https://huggingface.co/sandpies/FastH3-OmniRef-V2-Hybrid-int8-ComfyUI/tree/main/diffusion_models">https://huggingface.co/sandpies/FastH3-OmniRef-V2-Hybrid-int8-ComfyUI/tree/main/diffusion_models</a></li><li>OR EROS: <a target="_blank" rel="ugc" href="https://civarchive.com/models/2851079/h3-eros-max?modelVersionId=3429412">https://civarchive.com/models/2851079/h3-eros-max?modelVersionId=3429412</a></li><li>Text Encoder for 50xx Series Nvidia GPUs: <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/MiniMax-H3/blob/main/text_encoders/qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors">https://huggingface.co/Comfy-Org/MiniMax-H3/blob/main/text_encoders/qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors</a></li><li>Text Encoder for older GPUs: <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/MiniMax-H3/blob/main/text_encoders/qwen3vl_32b_minimax_h3_int8_convrot.safetensors">https://huggingface.co/Comfy-Org/MiniMax-H3/blob/main/text_encoders/qwen3vl_32b_minimax_h3_int8_convrot.safetensors</a></li><li>Video VAE: <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/MiniMax-H3/blob/main/vae/minimax_h3_video_vae_int8_convrot.safetensors">https://huggingface.co/Comfy-Org/MiniMax-H3/blob/main/vae/minimax_h3_video_vae_int8_convrot.safetensors</a></li><li>Audio VAE: <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/MiniMax-H3/blob/main/vae/minimax_h3_audio_vae_fp32.safetensors">https://huggingface.co/Comfy-Org/MiniMax-H3/blob/main/vae/minimax_h3_audio_vae_fp32.safetensors</a></li><li>H3 Latent Upscaler: <a target="_blank" rel="ugc" href="https://github.com/LBH-123-AI/Comfyui_Minimax_h3_latent_Upscaler">https://github.com/LBH-123-AI/Comfyui_Minimax_h3_latent_Upscaler</a></li></ul>Custom Nodes<ul><li><a target="_blank" rel="ugc" href="https://github.com/rgthree/rgthree-comfy">https://github.com/rgthree/rgthree-comfy</a></li><li><a target="_blank" rel="ugc" href="https://github.com/kijai/ComfyUI-KJNodes">https://github.com/kijai/ComfyUI-KJNodes</a></li><li><a target="_blank" rel="ugc" href="https://github.com/xmarre/Comfyui_Minimax_h3_latent_Upscaler-Plus">https://github.com/xmarre/Comfyui_Minimax_h3_latent_Upscaler-Plus</a></li></ul>

MMH3 Ref2VA - (simple) 2 Pass Workflow

Added RefMods. Additional major improvements for single-model-use and simplification. Back to Power Lora Loader. No more seperate Lora strengths for the 2 passes.

Files with SHA256: dda83960433e7d46b9d36e7e6e5d07bca81ca6d9fbe7cacfa1d48bebeb68c662