Attempt to build sage attention 2.2+ it speeds up MiniMax significantly.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/ee02eee7-3e71-441a-9725-4fdc9c17cf29/original=true/ee02eee7-3e71-441a-9725-4fdc9c17cf29.jpeg" />Tested on rtx5070ti.<h3 id="small-info:-bypass-params-main-every-time-ollama-is-active">Small info: bypass PARAMS / MAIN every time Ollama is active</h3>Diffusion model:<ul><li><a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/MiniMax-H3/blob/main/diffusion_models/minimax_h3_ref2va_pruned_int8_convrot.safetensors">https://huggingface.co/Comfy-Org/MiniMax-H3/blob/main/diffusion_models/minimax_h3_ref2va_pruned_int8_convrot.safetensors</a></li></ul>Text encoder:<ul><li><a target="_blank" rel="ugc" href="https://huggingface.co/sakamakismile/Qwen3-VL-32B-Heretic-MiniMax-H3-NVFP4/blob/main/qwen3vl_32b_heretic_minimax_h3_nvfp4.safetensors">https://huggingface.co/sakamakismile/Qwen3-VL-32B-Heretic-MiniMax-H3-NVFP4/blob/main/qwen3vl_32b_heretic_minimax_h3_nvfp4.safetensors</a></li></ul>VAEs:<ul><li><a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/MiniMax-H3/tree/main/vae">https://huggingface.co/Comfy-Org/MiniMax-H3/tree/main/vae</a></li></ul>Ollama LLM (gemma 4 12B it) for prompt helper:<ul><li><a target="_blank" rel="ugc" href="https://huggingface.co/mradermacher/gemma-4-12B-it-uncensored-heretic-GGUF/blob/main/gemma-4-12B-it-uncensored-heretic.Q8_0.gguf">https://huggingface.co/mradermacher/gemma-4-12B-it-uncensored-heretic-GGUF/blob/main/gemma-4-12B-it-uncensored-heretic.Q8_0.gguf</a></li></ul>You can add as many references as needed..<h2 id="down-below-is-old-version-content."> DOWN BELOW IS OLD VERSION CONTENT.</h2><h2 id="disable-comfyui_smznodes-if-you-have-this-extension!!!"> DISABLE comfyui_smznodes if you have this extension!!!</h2> (It breaks any LTXV workflow.)Diffuser model (pick one):<ul><li><a target="_blank" rel="ugc" href="https://huggingface.co/vantagewithai/LTX2.3-10Eros-1.4-GGUF/blob/main/10Eros_v1.4-Q4_K_S.gguf">https://huggingface.co/vantagewithai/LTX2.3-10Eros-1.4-GGUF/blob/main/10Eros_v1.4-Q4_K_S.gguf</a></li></ul><ul><li><a target="_blank" rel="ugc" href="https://huggingface.co/unsloth/LTX-2.3-GGUF/blob/main/ltx-2.3-22b-dev-Q4_K_S.gguf">https://huggingface.co/unsloth/LTX-2.3-GGUF/blob/main/ltx-2.3-22b-dev-Q4_K_S.gguf</a></li></ul>Text encoders (both):<ul><li><a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/ltx-2/blob/main/split_files/text_encoders/gemma_3_12B_it_fp8_scaled.safetensors">https://huggingface.co/Comfy-Org/ltx-2/blob/main/split_files/text_encoders/gemma_3_12B_it_fp8_scaled.safetensors</a></li><li><a target="_blank" rel="ugc" href="https://huggingface.co/unsloth/LTX-2.3-GGUF/blob/main/text_encoders/ltx-2.3-22b-dev_embeddings_connectors.safetensors">https://huggingface.co/unsloth/LTX-2.3-GGUF/blob/main/text_encoders/ltx-2.3-22b-dev_embeddings_connectors.safetensors</a></li></ul>VAEs:<ul><li><a target="_blank" rel="ugc" href="https://huggingface.co/Kijai/LTX2.3_comfy/tree/main/vae">https://huggingface.co/Kijai/LTX2.3_comfy/tree/main/vae</a></li></ul>Latent upscaler:<ul><li><a target="_blank" rel="ugc" href="https://huggingface.co/Lightricks/LTX-2.3/blob/main/ltx-2.3-spatial-upscaler-x2-1.1.safetensors">https://huggingface.co/Lightricks/LTX-2.3/blob/main/ltx-2.3-spatial-upscaler-x2-1.1.safetensors</a></li></ul>LoRA (distilled LoRA is needed, do not disable):<ul><li><a target="_blank" rel="ugc" href="https://huggingface.co/Kijai/LTX2.3_comfy/blob/main/loras/ltx-2.3-22b-distilled-lora-dynamic_fro09_avg_rank_105_bf16.safetensors">https://huggingface.co/Kijai/LTX2.3_comfy/blob/main/loras/ltx-2.3-22b-distilled-lora-dynamic_fro09_avg_rank_105_bf16.safetensors</a></li></ul>Enjoy!

MiniMax H3 with SEEDVR2 upscaler + Ollama prompt helper | 16GB VRAM 32GB RAM

Added ollama prompt helper, tuned optimization/quality.

Files with SHA256: 05998ce294ecd78710ad0c6445f140ffd97fa5cb686c25fd6ccad4fa54d3f9e6