Hands-On Machine Learning for Algorithmic Trading

How to train the model

Model training consists of an outer loop for each epoch, that is, each pass over the training data, and an inner loop over the batches produced by the DataLoader. That executes the forward and backward passes of the learning algorithm. Some care needs to be taken to adjust data types to the requirements of the various objects and functions; for example, labels need to be integers and the features should be of type float, as follows:

for epoch in range(num_epochs):    print(epoch)    for i, (features, label) in enumerate(dataloader):                features = Variable(features.float())                 label = Variable(label.long())        # Initialize the hidden weights        optimizer.zero_grad()                  # Forward pass: compute output given features        outputs = net(features) ...