Accessing training data in custom callbacks

I’m working on super resolution generative adversarial network in which I’m trying to visualize data at the end of every epochs using custom callbacks

But things is that I’m kinda structs in developing custom callbacks , I don’t know how to access training data in custom callbacks

I have looked at on_train_batch_begin() method but i want the picture or image at the end of epoch

Is anyone have done that ? then please let me know

Thanks :blush:

Instead of using the entire training dataset, you can set aside a few samples that you would like to visualize over time and use them as needed. The following examples provide good references:

1 Like

If you like the Tensorboard you can use callback with
your image summaries Displaying image data in TensorBoard  |  TensorFlow

1 Like