[Q] How can I integrate my plots instead of having separate plots?

[monajalal]

Also, I only have used 20 epochs for train/val. I am not sure why it is showing 100 epochs?

Here's what I've been using:

            wandb.log({"train loss": train_epoch_loss,
                       "val loss": val_epoch_loss,
                       "epoch": epoch})

            wandb.log({"train acc": train_epoch_acc,
                       "val acc": val_epoch_acc,
                       "epoch": epoch})

            wandb.log({"best val acc": val_best_acc, "epoch": epoch})

[exalate-issue-sync[bot]]

Leslie commented: Hello, what is on your charts are the steps not epochs. An epoch refers to one pass through your training data. If you’re doing batch gradient descent than each training iteration (sometimes referred to as training step) is equivalent to an epoch but most deep learning algorithms use mini-batch gradient descent so each epoch contains several iterations/steps. So you will have multiple steps per epoch. You can use define_metric (https://docs.wandb.ai/ref/python/run#define_metric) to indicate epochs

[exalate-issue-sync[bot]]

Leslie commented: Hi, I wanted to follow up with you regarding your support request as we have not heard back from you. Please let us know if we can be of further assistance or if your issue has been resolved.

[monajalal]

Hi Leslie, is there an option to meet with one of your engineers via zoom? My email is jalal@bu.edu Thank you, Mona Jalal

[exalate-issue-sync[bot]]

Leslie commented: Hi Mona, can you let me know through here what the issue you are running into is with as much detail as possible? Our engineers don't jump on calls with users. However, I can relay the information you tell me to them.

[exalate-issue-sync[bot]]

Leslie commented: Hi again Mona, we wanted to follow up with you regarding your support request as we have not heard back from you. Please let us know if we can be of further assistance or if your issue has been resolved.