/cog-sdxl

Stable Diffusion XL training and inference as a cog model

Primary LanguagePython

Cog-SDXL

Replicate demo and cloud API

This is an implementation of Stability AI's SDXL as a Cog model.

Development

Follow the model pushing guide to push your own fork of SDXL to Replicate.

Basic Usage

for prediction,

cog predict -i prompt="a photo of TOK"
cog train -i input_images=@example_datasets/__data.zip -i use_face_detection_instead=True
cog run -p 5000 python -m cog.server.http

Update notes

2023-08-17

  • ROI problem is fixed.
  • Now BLIP caption_prefix does not interfere with BLIP captioner.

2023-08-12

  • Input types are inferred from input name extensions, or from the input_images_filetype argument
  • Preprocssing are now done with fp16, and if no mask is found, the model will use the whole image

2023-08-11

  • Default to 768x768 resolution training
  • Rank as argument now, default to 32
  • Now uses Swin2SR caidas/swin2SR-realworld-sr-x4-64-bsrgan-psnr as default, and will upscale + downscale to 768x768