How to use pretrained models?

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

474 stars 75 forks source link

Closed hiepph closed 5 years ago

hiepph commented 5 years ago

Hi, your README doesn't include how to load pretrained models and infer with a test image. Can you update or show an example here?

Naman-ntc commented 5 years ago

Hi,

First you'll have to load a pretrained model loaded_model = Model.load_state_dict(torch.load(path))
Center crop the image (human body inside a 256x256 window centered)
Run evaluation using loaded_model(image) and use prediction as here depending on the model you load. That should be sufficient I believe