Writing training data to a log file and plotting training curves

1

Hello everyone!!
How can I write my training data like the training loss, training error, validation error in a log file and then plot i the learning curve to visualize the performance of my model??
Please suggest . Also it would be very helpful, if there is some pre-implemented example related to that.

Thanks and regards

2

Save the stat of each epoch either in numpy array or in a list and save it. When the training process ends, plot the stat saved. I think it is pretty simple.

1 Like
3

You can do it with the tensorboard logger.

For more information, please refer to https://github.com/TeamHG-Memex/tensorboard_logger

I am using it and it works perfectly for me.

4 Likes
4

I think tensorboardX is better. You can find tutorial here.

5

Does TensorboardX work for Windows? I am using Jupyter via Anaconda on Windows OS.