search

potterhsu / easy-faster-rcnn.pytorch

An easy implementation of Faster R-CNN (https://arxiv.org/pdf/1506.01497.pdf) in PyTorch.
MIT License
165 stars 57 forks source link

Problems with infer #8

Open Codermay opened 5 years ago

Codermay commented 5 years ago

Thanks for your good implementation. I have run your code on my own dataset. And I want to know, do your code provide multiple images inferring?

potterhsu commented 5 years ago

No, but you can easily extend for your needs.

PaperLamp commented 5 years ago

Thanks for your codes. May I ask for your help about the following problem: When I try to evaluate or infer, I meet this error: 1.evaluate: size mismatch for detection._proposal_transformer.bias: copying a param with shape torch.Size([368]) from checkpoint, the shape in current model is torch.Size([84]). 2.infer: size mismatch for rpn._anchor_transformer.weight: copying a param with shape torch.Size([48, 512, 1, 1]) from checkpoint, the shape in current model is torch.Size([36, 512, 1, 1]). Thank you

PeterBishop0 commented 5 years ago

same problems,have you worked it out?@PaperLamp

PeterBishop0 commented 5 years ago

@potterhsu

potterhsu commented 5 years ago

These sorts of errors usually occurred when your configs between training and inference are inconsistent, for example: