Note-Liu
Note-Liu
I also want to ask this question.
> When i use convert_onnx.py to convert crowdhuman.pth or mot_half.pth, it's failed. Then I find the description in the annotation:" Script to convert a trained CenterNet model to onnx, currently...
Traceback (most recent call last): File "main.py", line 71, in trainer.train_one_epoch(epoch) File "/home//W-Net-PyTorch/src/trainer.py", line 96, in train_one_epoch cls_enc_s, File "/pytorch/lib/python3.6/site-packages/torch/nn/modules/module.py", line 889, in _call_impl result = self.forward(*input, **kwargs) File "/home/W-Net-PyTorch/src/loss_func.py",...
> 比赛的数据能分享一下吗？谢谢。结束之后才看到的比赛，数据没法下载了，望能够提供数据连接谢谢。 > > 你好，这边比赛的数据集有点大，不太好分享，你可以直接换成其他数据集来做相关实验，如果还是需要，我这边想办法传到网盘分享。 请问有传数据集吗？
同求数据集
I have also encountered this problem. Have you resolved it？
> @Goldatch 您好，感谢您对本项目的关注，我们文章提到的数据集对于字体的选择是没有特殊的考虑的，我们在实验过程中是直接在公开字库中进行随机选取的；当然依照我们的经验，当训练数据的字体风格越多样，字的复杂度多样，模型的鲁棒性会更好～ When using TTF to generate font images, does the absence of certain Chinese characters in a TTF file affect training? Must font images for each style be...
I also want to know.