加载模型出错：Missing key和Unexpected key about fots.pytorch

[niloofarAzari]

@niloofarAzari @AndosonLiu 请问 这个问题有解决么，望解答 谢谢

comment line 20 and 21 in eval.py and try again. I mean these two lines:
if with_gpu:
model.parallelize()

[niloofarAzari]

i have the same problem, may someone help please?

[SealQ]

@niloofarAzari @AndosonLiu 请问 这个问题有解决么，望解答 谢谢

[SealQ]

@niloofarAzari Hello, this problem has been solved, it can be run, thank you! I'd like to ask you another question. I have the following situations when I use my own data for training. Please give me your advice!

[niloofarAzari]

@niloofarAzari Hello, this problem has been solved, it can be run, thank you! I'd like to ask you another question. I have the following situations when I use my own data for training. Please give me your advice!

about "poly in wrong direction" warning: poly coordinates in your dataset gt files should be clockwise and start with top left corner
and about the results you get, continue training and see if you get better results!

[SealQ]

@niloofarAzari Hello, this problem has been solved, it can be run, thank you! I'd like to ask you another question. I have the following situations when I use my own data for training. Please give me your advice!

about "poly in wrong direction" warning: poly coordinates in your dataset gt files should be clockwise and start with top left corner
and about the results you get, continue training and see if you get better results!

Hello, it's adjusted, but there will still be a situation where val_precious and val_recall is 0. Didn't this happen when you were training?

[niloofarAzari]

@SealQ hi, have you succeeded to train this model and get better precious and recall?
i tried to train it on ICDAR2015 but the result on detection was not satisfactory.