I am working on continual learning and I train a model on multiple tasks. Each task is trained for a number of epochs. This means that I have a nested loop like the following:
for task_idx in range(num_tasks):
train_model_on_task()
and the train function is:
def train_model_on_task():
for epoch in range(num_epochs):
avg_loss = TrainEpoch()
avg_loss_eval = EvalEpoch()
I would like to see the acc anfd loss for the train and eval dataset of all the tasks but I don’t know how to give these commands to the tensoboard. Any help would be appreciated.