44 lines (32 loc) · 1.22 KB

beprepared!

0.2.3

Minor Web UI improvements
Multi-GPU support for captioning and watermark removal
Fix bug in Take in random mode where it would crash if n<len(dataset.images)

0.2.2

Implement beprepared cli
Change database format to make domain a first class concept
Update documentation to reflect the CLI

0.2.2

Implement beprepared cli
Change database format to make domain a first class concept
Update documentation to reflect the CLI

0.1.2

Support JoyCaption Alpha 2
Support Gemini Captioning
Support Florence2 Captioning
Support FilterByAspectRatio
Implement LLMCaptionVariations
Release GPU memory explicitly after using ML models
Add port configuration so more than one beprepared can run on the same machine at the same time
Fix GPU OOM errors related to Qwen VL 2
Implement SwinIR based upscaler
Add progress bars + console view to the web interface for generic nodes
Support enter keystroke to "Continue to next step" on HumanTag / HumanFilter
Fix websocket connection issues
Fix web interface usage outside of npm run dev
Improve logging

0.1.1

Fixed bugs, improved documentation, fixed consistency issues

0.1.0

Initial release to test PyPI publishing