🚀 Features
- Try using all of the available inference providers 12, Thanks knorth55
- Support <model_name>[:latest] syntax, rename model tags to :latest 17
✨ Enhancement
- Add benchmark.py and make only CUDA enable when available (ignore TensorRT, Azure, CoreML, which are not-tested / not-benefit) 15
💬 Other
- Run CI on macOS and Windows 13
- Fix README (output extension must be .png) 14
- Install onnxruntime-gpu for non-Darwin platform 16
---
- For my daily development update, check [my Twitter/X](https://twitter.com/wkentaro_).
- If you'd like to support this project, check [my sponsoring page](https://github.com/sponsors/wkentaro).