I intend to plot the learning curve for all three splits i.e training, validation, and test sets.
Learning curve: loss/accuracy on the y-axis and number of steps on the x-axis.
Tensorboard gives it for the training set (for every step and epoch).
How can I get this curve for val and test sets i.e it should evaluate the val and test set after every n steps and compute the loss/accuracy?
For the validation set, I can use val_check_interval=0.25 but is there a direct way to plot the learning curve
Also, how can achieve it for test set?