Predictions from the trained model on the validation set look faded.
Any suggestion on how to tune would be helpful. Thanks
Glad that I was able to make a transition into making a clear and good prediction when I had changed my optimizer from sgd() to Adam(). I used the sgd() optimizer in Keras and it works well, but when I used the same parameters for the optimizer and model in torch it failed to work.
Can someone explain why this happened? Thanks