bruxosdovfx - Camera H3 v33
Visual camera editor for MiniMax H3 inside ComfyUI.
Drag the camera, create keyframes, and generate camera-path prompts for H3.
Installation
cd ComfyUI/custom_nodes
git clone https://github.com/NyckM/3d-Camera-control-H3-Minimax.gitRestart ComfyUI and search for:
bruxosdovfx • Camera H3
https://github.com/NyckM/3d-Camera-control-H3-Minimax/tree/main
Models and Lora
https://huggingface.co/NyckM/Meridian_CameraH3_INT8_build_by_BruxosdoVFX/tree/main
Description
FAQ
Comments (7)
What? Do you really need a whole different minimax model to do that? 😢
no, you can use two ways
1 Uses the converted Meridian model, DMD LoRA and H3 3-step LoRA. (better quality)
2 Uses the original MiniMax H3 FL2VA base with the official Meridian Teacher → Turbo adapters. (you can use just loras on huggingface and the custom node, and your model)
I can’t find the ‘bruxosloadvideo’ node
https://github.com/NyckM/Video-Util-ComfYUI
i Use this for load video, because i can crop and rotate, inverse duration, but you can use just load video normal.
@BruxosDoVFX thkx, nice job!!
According to requirements, this Text Embed meridian_fixed_embed_124.safetensors suggested for the Quality.json workflow, should go to the Meridian text-cond folder - where is this folder? shall we create it somewhere? like models/diffusion_models/H3camera/ for MeridianH3Camera_int8_pruned.safetensors
will models/text_encoders/meridian_text_cond/ be correct?
btw, can we change this number of frames (124) somehow, and (if we can), where the other fixed text embed with different numbers of frames could be downloaded?