束利阳

PyTorch 1.0.1 support CTCLoss itself. Use `torch.nn.CTCLoss`

Try this, i rewrite it in python3 and fix bug. https://github.com/Holmeyoung/crnn_pytorch

Hi, you can try this, i rewrite it to support Chinese. https://github.com/Holmeyoung/crnn_pytorch

label里的字符在alphabets.py里没出现。

Hi, the images used in pre-trained model are generated by me and they all look very well. Just try to make more training data, it will work!

Hi, the images from CTPN should look like your training data. If your training data is generated by yourself and looks very nice, while the images out from CTPN are...

```python if __name__ == "__main__": for epoch in range(params.nepoch): train_iter = iter(train_loader) i = 0 while i < len(train_loader): cost = train(crnn, criterion, optimizer, train_iter) loss_avg.add(cost) i += 1 if...

Of course you can. Thank you~

Hi, in fact, it only depends on the output lstm T length.

> I got the answer from your reply from #17 > > > You need to calculate it. After conv and pool what's the image width. The image width will...