search

yhenon / pytorch-retinanet

Pytorch implementation of RetinaNet object detection.
Apache License 2.0
2.14k stars 665 forks source link

How to use the pre-trained model? #81

Open handsomefengfeng opened 5 years ago

handsomefengfeng commented 5 years ago

I've downloaded the state dict and pytorch model. I'm wondering the effect on VOC. What should i do next? Put the downloaded dict and model into the project file？ Thanks.

zabulskyy commented 5 years ago

I recommend you to use PyTorch model.

You should have something like

weights = "path/to/weights/retinanet.weights" retinanet = torch.load(weights)

where retinanet should be usable model