Skip to content

README.md

Latest commit

 

History

History
executable file
·
32 lines (21 loc) · 1.11 KB

README.md

File metadata and controls

executable file
·
32 lines (21 loc) · 1.11 KB
Raw

Cog-SDXL

Replicate demo and cloud API

This is an implementation of the SDXL as a Cog model. Cog packages machine learning models as standard containers.

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/zeke.zip -i use_face_detection_instead=True

Update notes

2021-08-12

  • 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