Ep. 3: Wrong model for batch norm, change function parameters

[ashwinvis]

• Errorneous instantiation in notebook and text. WeatherPredictionModel instead of WeatherPredictionModelBatchNorm. See https://learn.mimer-ai.eu/deep-learning-intro/3-monitor-the-model/#batchnorm-the-standard-scaler-for-deep-learning
• accumulate is not the best name for storing targets for inference. store_targets maybe?
• Arrange order of function parameters in a nicer way and make them consistent across episodes. See fit and fit_with tensorboard:

def fit(model, train_dl, loss_fn, optimizer, val_dl):
    ...

def fit_with_tensorboard(model, train_loader, val_loader, loss_fn, optimizer):
    ...