Performance of Pyranet on MPII

Naman-ntc / Pytorch-Human-Pose-Estimation

Implementation of various human pose estimation models in pytorch on multiple datasets (MPII & COCO) along with pretrained models

MIT License

472 stars 76 forks source link

Performance of Pyranet on MPII #5

Closed HenryJunW closed 5 years ago

HenryJunW commented 5 years ago

Hello, impressive work. I am wondering whether you can post the results on MPII of the Pyranet. If possible, would you mind posting the pre-trained model as well. Thanks.

Naman-ntc commented 5 years ago

Hi, I have tried training Pyranet on MPII dataset but was unable to reproduce official results. It might probably be difference in training routine or data augmentation or some subtle difference in architecture of model written in Lua-Torch or Pytorch. I will hopefully try to train it again sometime in future but it doesn't seem possible currently due to other commitments. Feel free to try to train it on your own!