abskj / lossy-image-compression Goto Github PK

How large was your dataset, to be able to decompress almost perfectly? when I try to train with my own dataset I get something that looks like this :

I did not change your code at all, just an issue with training.
I am only assuming it is the problem of the dataset training and nothing else as I have not changed anything else

I need to train the model to got de-compressed images near the original with better quality how can I do that ? which params needs to change?

@abskj thank you so much for sharing this code.

I tried both codes in the 2 branches but autoencoder model in the master branch with less number of hidden layers gives me good images quality when decoding and it is better than model2, why this happen with me and what is the reason for that? is it normal ?

print('Train Loss:\t%.6f'%(total_loss*train_params['batch_size']/len(train_params['train_indices'])))

ZeroDivisionError: division by zero。

我不明白train_params['train_indices']在哪里赋值了，那么它的长度自然就是0.这个loss计算的定义是什么呢？
I don't understand train_ Where is params ['train_indices'] assigned, and its length is naturally 0. What is the definition of this loss calculation?