---
license: mit
---

Re-upload of the classic 2019 hit model, YOLACT, in case you wanted to relive real-time instance segmentation with a retro vibe.

These are the initial weights files described in the readme of the YOLACT github.

🔬 **Papers:**
- [YOLACT: Real-time Instance Segmentation (2019)](https://arxiv.org/abs/1904.02689)
- [YOLACT++: Better Real-time Instance Segmentation (2020)](https://arxiv.org/abs/1912.06218)
  
💻 **Run the code:** [dbolya/yolact](https://github.com/dbolya/yolact)
  
🚀 **Download this checkpoint:**
 - [resnet101_reducedfc.pth](https://huggingface.co/dbolya/yolact-initial-weights/resolve/main/resnet101_reducedfc.pth?download=true)
 - [resnet50-19c8e357.pth](https://huggingface.co/dbolya/yolact-initial-weights/resolve/main/resnet50-19c8e357.pth?download=true)
 - [darknet53.pth](https://huggingface.co/dbolya/yolact-initial-weights/resolve/main/darknet53.pth?download=true)