"AFEW dataset face alignment" and "accuracy on validation" problem.
yuexihang opened this issue · comments
Hello. I use the ./data/face_alignment_code/frame2face_afew.py to process the frame files from AFEW videos.
However, I found some face can't be detected, which causes that some aligned face folder was empty.
I want to know about whether you just use the ./data/face_alignment_code/frame2face_afew.py to process or not? Or do you have another ways?
What'more, I got the overfitting error after just several epochs, and the accuracy on validation is just around 41 (without the relation-attention) or 16 (with the relation-attention).
I use the code without any changing.
I have solved the problem.
Hello yuexihang, I meet the same overfitting error problem, could you please give a hint about how you solved the problem? thank you very much :)
Hello yuexihang, I meet the same overfitting error problem, could you please give a hint about how you solved the problem? thank you very much :)
Hello, I don't know if you have solved it.
I got the overfitting error for an interesting reason. Human face was not detected on some frames by the preprocessing method and I just choosed to ignore these frames, which causes the disalignment between emotion labels and frames. It means that a lot of frames have the incorrect emotion label.
Hello yuexihang, I meet the same overfitting error problem, could you please give a hint about how you solved the problem? thank you very much :)
Hello, I don't know if you have solved it. I got the overfitting error for an interesting reason. Human face was not detected on some frames by the preprocessing method and I just choosed to ignore these frames, which causes the disalignment between emotion labels and frames. It means that a lot of frames have the incorrect emotion label.
Hello, I'm new to this, could you let me know how you modified the code?
Hello yuexihang, I meet the same overfitting error problem, could you please give a hint about how you solved the problem? thank you very much :)
Hello, I don't know if you have solved it. I got the overfitting error for an interesting reason. Human face was not detected on some frames by the preprocessing method and I just choosed to ignore these frames, which causes the disalignment between emotion labels and frames. It means that a lot of frames have the incorrect emotion label.
Thank you very much that telling me the reason for the relation-attention. The old version will not have this kind of problem https://github.com/Open-Debin/Emotion-FAN/releases/tag/v1.0_august_2019. Also, the model can not be trained from scratch. please train it based on pretrained emotion model. Due to different Machines, I suggest you try different learning rates.
Hello. I use the ./data/face_alignment_code/frame2face_afew.py to process the frame files from AFEW videos. However, I found some face can't be detected, which causes that some aligned face folder was empty.
I want to know about whether you just use the ./data/face_alignment_code/frame2face_afew.py to process or not? Or do you have another ways?
What'more, I got the overfitting error after just several epochs, and the accuracy on validation is just around 41 (without the relation-attention) or 16 (with the relation-attention).
I use the code without any changing.
May I ask you for AFEW dataset? I sent the e-mail to the organization but they didn't reply me