Add a short README

README.md

@@ -21,3 +21,47 @@ tags:
 - minimax-h3
 - pruna-ai
 ---
+
+# p-h3-image for ComfyUI
+
+p-h3-image, based on MiniMax-H3, makes images from text and edits images, with up to 9 input images. This repository
+has its ComfyUI files.
+
+## Use it
+
+1. Install the nodes (ComfyUI 0.43 or newer):
+
+   ```bash
+   git clone https://github.com/PrunaAI/comfyui-p-h3-image ComfyUI/custom_nodes/comfyui-p-h3-image
+   ```
+
+2. Download the files into `ComfyUI/models`:
+
+   ```bash
+   hf download PrunaAI/p-h3-image-comfy --include "diffusion_models/p-h3-image_bf16.safetensors" "text_encoders/p-h3-image_text_encoder_bf16.safetensors" "text_encoders/p-h3-image_prompt_upsampler_bf16.safetensors" "vae/p-h3-image_vae_bf16.safetensors" --local-dir ComfyUI/models
+   ```
+
+   Less VRAM: take the `w4a8` diffusion model and text encoder and the `int8_convrot` VAE instead.
+
+3. Open a workflow from `workflows/` (also under Templates in ComfyUI), type a short request and run.
+
+Keep the prompt upsampler in the graph: it turns a short request, in any language, into the full prompt p-h3-image
+needs. The workflows default to 8 megapixels and 8 steps.
+
+Already have MiniMax-H3 (FL2VA) in ComfyUI? Add `loras/p-h3-image_lora_bf16.safetensors` and use the `lora` workflows.
+
+## Files
+
+```
+ComfyUI/models/
+├── diffusion_models/  p-h3-image_{bf16,int8_convrot,w4a8}.safetensors
+├── text_encoders/     p-h3-image_text_encoder_{bf16,int8_convrot,w4a8}.safetensors
+│                      p-h3-image_prompt_upsampler_{bf16,int8_convrot,w4a8}.safetensors
+├── vae/               p-h3-image_vae_{bf16,int8_convrot}.safetensors
+└── loras/             p-h3-image_lora_bf16.safetensors
+```
+
+## License
+
+The p-h3-image weights are under the MiniMax H3 Community License ([LICENSE](LICENSE)), the prompt upsampler under
+Apache 2.0 ([LICENSE-APACHE](LICENSE-APACHE)). See [NOTICE](NOTICE).