Upload README.md with huggingface_hub

README.md

@@ -0,0 +1,44 @@
+---
+tags:
+- photogen
+- storymaker
+- ip-adapter
+- sd-xl
+- character-consistency
+- face-id
+license: other
+---
+
+# Photogen Checkpoints
+
+Trained character adapter weights for [Photogen](https://github.com/AshimDangol/Photogen) (StoryMaker-based character-consistent image generation).
+
+## Checkpoints
+
+| Step | File | Description |
+|------|------|-------------|
+| 50   | checkpoint-50/mask.bin | Early checkpoint (50 steps) |
+| 100  | checkpoint-100/mask.bin | Medium checkpoint (100 steps) |
+| 150  | checkpoint-150/mask.bin | Recommended checkpoint (150 steps) |
+| 200  | checkpoint-200/mask.bin | Extended checkpoint (200 steps) |
+
+## Usage
+
+Download to your project root and use with the Photogen pipeline:
+
+`ash
+huggingface-cli download DeathAngel00/Photogen-checkpoints checkpoint-150/mask.bin --local-dir checkpoints --local-dir-use-symlinks False
+`
+
+Then in code:
+
+`python
+from pipeline_sdxl_character import StableDiffusionXLCharacterPipeline
+
+pipe = StableDiffusionXLCharacterPipeline.from_pretrained('huaquan/YamerMIX_v11', torch_dtype=torch.float16)
+pipe.load_character_adapter('laion/CLIP-ViT-H-14-laion2B-s32B-b79K', 'checkpoints/checkpoint-150/mask.bin', scale=0.8, lora_scale=0.8)
+`
+
+## License
+
+Same as the original StoryMaker project.