johndpope commited on
Commit
3085e84
·
verified ·
1 Parent(s): a707c5a

Card: ComfyUI nodes and workflow

Browse files
Files changed (1) hide show
  1. README.md +4 -0
README.md CHANGED
@@ -36,4 +36,8 @@ Each folder has `adapter.safetensors` (LoT heads + fitted extent bank), `lora.sa
36
 
37
  **Status: research.** Coarse regions are still softer than dense H3. Renders, eval curves and caveats are in [johndpope/MiniMax-H3 issue 1](https://github.com/johndpope/MiniMax-H3/issues/1). Code, and training from your own mp4s: [`scripts/lot`](https://github.com/johndpope/MiniMax-H3/tree/main/scripts/lot) ([cheatsheet](https://github.com/johndpope/MiniMax-H3/blob/main/scripts/lot/CHEATSHEET.md)).
38
 
 
 
 
 
39
  Derived from MiniMax H3 weights; use is governed by the [MiniMax H3 Community License Agreement](https://huggingface.co/MiniMaxAI/MiniMax-H3/blob/main/LICENSE).
 
36
 
37
  **Status: research.** Coarse regions are still softer than dense H3. Renders, eval curves and caveats are in [johndpope/MiniMax-H3 issue 1](https://github.com/johndpope/MiniMax-H3/issues/1). Code, and training from your own mp4s: [`scripts/lot`](https://github.com/johndpope/MiniMax-H3/tree/main/scripts/lot) ([cheatsheet](https://github.com/johndpope/MiniMax-H3/blob/main/scripts/lot/CHEATSHEET.md)).
38
 
39
+ ## ComfyUI (one image)
40
+
41
+ Nodes: **MiniMax H3 LoT Apply** + **MiniMax H3 LoT Unscale Latent** in [ComfyUI-MiniMax-H3-Image-Lane](https://github.com/johndpope/ComfyUI-MiniMax-H3-Image-Lane/tree/lot) (branch `lot`), with the workflow `workflows/t1_lot_image_api.json` (LoT and dense A/B from one prompt). Put `run4_nikki_distill/adapter.safetensors` in `models/lot/` and `run4_nikki_distill/lora.safetensors` in `models/loras/`, and use the int8 ConvRot FL2VA DiT the adapters were trained on. Unscale Latent is required before VAE Decode.
42
+
43
  Derived from MiniMax H3 weights; use is governed by the [MiniMax H3 Community License Agreement](https://huggingface.co/MiniMaxAI/MiniMax-H3/blob/main/LICENSE).