StevenLiuWen / ano_pred_cvpr2018

Official implementation of Paper Future Frame Prediction for Anomaly Detection -- A New Baseline, CVPR 2018
433 stars 126 forks source link

Training of U-Net #53

Closed kakashiayush closed 2 years ago

kakashiayush commented 3 years ago

can someone tell me that how U-Net is training in this whole code because i am not able to see any code which is used to train U-Net which is predicting a new frame. In research paper they have told that they are training their generator.

StevenLiuWen commented 3 years ago

Hi, @kakashiayush,in this paper, the generator is the U-Net, and they are the same thing. https://github.com/StevenLiuWen/ano_pred_cvpr2018/blob/master/Codes/models.py#L15

ayushdgupta commented 3 years ago

1-> I already studied this in your paper but my question is you are generating future frames without training your neural network but in your paper you have written about training your generator, I'm not getting this point that how you are training your generator??? 2-> And one more thing that you have used tf.train(-----------------------------------) in your code in train.py file. what is that line is doing (i mean that which kind of model or which model that line is training). Please clear these 2 douts of mine.