PlagueKind - Minimax H3 - Image Edit - Eros + Refmods help me get a 5090 <a target="_blank" rel="ugc" href="https://throne.com/plaguekind/item/073bace1-2e1a-4492-95e6-024b4da6459b">https://throne.com/plaguekind/item/073bace1-2e1a-4492-95e6-024b4da6459b</a><h1 id="disclaimer-read-this">DISCLAIMER READ THIS</h1>Update my nodepack BEFORE downloading and opening any workflow I release. I won't be responding to comments about the node errors like that anymore. ### V1 Notes - UPDATE NODEPACK TO V1.6.2 AND COMFYUI TO 0.44.0 or newer, and pytorch 2.13cu130 or newer.- For refmod use, lynnreal is required in the encoder slot, if only using pictures you can use the other. P-h3 must always be in decode.- Disable the image node with the bypasser or remove your images for T2I.- I kept the sparse node, not for the speed but for the ref enforcement option.- Preview-Image node at the end instead of Save, because I and probably you don't want a bunch of failed attempts piling up.- Create refmods in the subgraph.- Added lynnreal light vae, I didn't add sol vae because it was too soft.- Recommended 5-12 steps, with default settings.- Best use between 4-8MP- Overall function is similar to my Video WF, so you'll probably feel at home.<hr />### SupportIf you find this project useful and want to support development:Monero (XMR):<pre><code>865BrcfWLdwELwuq5faV1uVTbh93zVK6AUYLY2c3mX6sFfAGRfS6axe1kBTYYKuM7ccN7zBZDAZvnT7E4NKmUazySdbpc7p</code></pre>help me get a 5090 <a target="_blank" rel="ugc" href="https://throne.com/plaguekind/item/073bace1-2e1a-4492-95e6-024b4da6459b">https://throne.com/plaguekind/item/073bace1-2e1a-4492-95e6-024b4da6459b</a> ### Requirements- [ComfyUI-PlagueKind-Nodes](<a target="_blank" rel="ugc" href="https://github.com/PlagueKind/Comfyui-PlagueKind-Nodes">https://github.com/PlagueKind/Comfyui-PlagueKind-Nodes</a>)- [ComfyUI-KJNodes](<a target="_blank" rel="ugc" href="https://github.com/kijai/ComfyUI-KJNodes">https://github.com/kijai/ComfyUI-KJNodes</a>)- [Rgthree-comfy](<a target="_blank" rel="ugc" href="https://github.com/rgthree/rgthree-comfy">https://github.com/rgthree/rgthree-comfy</a>)- [ComfyUI-VideoHelperSuite](<a target="_blank" rel="ugc" href="https://github.com/Kosinkadink/ComfyUI-VideoHelperSuite">https://github.com/Kosinkadink/ComfyUI-VideoHelperSuite</a>)- [ComfyUI-MiniMaxH3Mod](<a target="_blank" rel="ugc" href="https://github.com/Luisacaotica/ComfyUI-MiniMaxH3Mod">https://github.com/Luisacaotica/ComfyUI-MiniMaxH3Mod</a>)- [ComfyUI_PowerShiftScheduler](<a target="_blank" rel="ugc" href="https://github.com/silveroxides/ComfyUI_PowerShiftScheduler">https://github.com/silveroxides/ComfyUI_PowerShiftScheduler</a>)---### Diffusion ModelMy Quant of Eros beta7, singularity v2, prism, and pika flex.[H3_Eros_Sing_PRISM_Pika_Parasyte_w6a8_g32](<a target="_blank" rel="ugc" href="https://huggingface.co/Plaguekind/H3_Eros_Sing_PRISM_Pika_Parasyte_w6a8_g32/resolve/main/H3_Eros_Sing_PRISM_Pika_Parasyte_w6a8_g32.safetensors">https://huggingface.co/Plaguekind/H3_Eros_Sing_PRISM_Pika_Parasyte_w6a8_g32/resolve/main/H3_Eros_Sing_PRISM_Pika_Parasyte_w6a8_g32.safetensors</a>)### VAEAudio VAE for refmod creation.[minimax_h3_audio_vae_fp32.safetensors](<a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/MiniMax-H3/resolve/main/vae/minimax_h3_audio_vae_fp32.safetensors">https://huggingface.co/Comfy-Org/MiniMax-H3/resolve/main/vae/minimax_h3_audio_vae_fp32.safetensors</a>)Encode VAE for refmod use.[minimax_h3_lynnreal_light_vae_int8_convrot](<a target="_blank" rel="ugc" href="https://huggingface.co/Kijai/MiniMax-H3-experimental/resolve/main/minimax_h3_lynnreal_light_vae_int8_convrot.safetensors">https://huggingface.co/Kijai/MiniMax-H3-experimental/resolve/main/minimax_h3_lynnreal_light_vae_int8_convrot.safetensors</a>)Decode VAE or Encode without refmod.[p-h3-image_vae_int8_convrot](<a target="_blank" rel="ugc" href="https://huggingface.co/PrunaAI/P-H3-Image-Comfy/resolve/main/vae/p-h3-image_vae_int8_convrot.safetensors">https://huggingface.co/PrunaAI/P-H3-Image-Comfy/resolve/main/vae/p-h3-image_vae_int8_convrot.safetensors</a>)Preview VAE[vae_approx/taeh3.safetensors](<a target="_blank" rel="ugc" href="https://huggingface.co/Kijai/MiniMax-H3-TAE/resolve/main/vae_approx/taeh3.safetensors">https://huggingface.co/Kijai/MiniMax-H3-TAE/resolve/main/vae_approx/taeh3.safetensors</a>)### Text Encoder [qwen3vl_32b_minimax_h3_native_w4w6_17.93gb](<a target="_blank" rel="ugc" href="https://huggingface.co/LokkenJP/TenEros_Max_optimized_learned/resolve/main/text_encoder/qwen3vl_32b_minimax_h3_native_w4w6_17.93gb.safetensors">https://huggingface.co/LokkenJP/TenEros_Max_optimized_learned/resolve/main/text_encoder/qwen3vl_32b_minimax_h3_native_w4w6_17.93gb.safetensors</a>)### Edit lora[p-h3-image_lora_bf16](<a target="_blank" rel="ugc" href="https://huggingface.co/PrunaAI/P-H3-Image-Comfy/resolve/main/loras/p-h3-image_lora_bf16.safetensors">https://huggingface.co/PrunaAI/P-H3-Image-Comfy/resolve/main/loras/p-h3-image_lora_bf16.safetensors</a>)---

PlagueKind - Minimax H3 - Image Edit - Eros + Refmods

### V1 Notes- For refmod use, lynnreal is required in the encoder slot, if only using pictures you can use the other. P-h3 must always be in decode.- Disable the image node with the bypasser or remove your images for T2I.- I kept the sparse node, not for the speed but for the ref enforcement option.- Preview-Image node at the end instead of Save, because I and probably you don't want a bunch of failed attempts piling up.- Create refmods in the subgraph.- Added lynnreal light vae, I didn't add sol vae because it was too soft.- Recommended 5-12 steps, with default settings.- Best use between 4-8MP- Overall function is similar to my Video WF, so you'll probably feel at home.

Files with SHA256: 5b0d96153cce8176e58cd7f5ba329d6eae22c16ab4b4e3e6ca38f0ac0a81a505