AdamW with Trainer

You're beginning to train a Transformer model to simplify language translations. As a first step, you decide to use the AdamW optimizer as a benchmark and the Trainer interface for quick setup. Set up Trainer to use the AdamW optimizer.

AdamW has been pre-imported from torch.optim. Some training objects have been pre-loaded: model, training_args, train_dataset, validation_dataset, compute_metrics.

Deze oefening maakt deel uit van de cursus

Efficient AI Model Training with PyTorch

Cursus bekijken

Oefeninstructies

Pass the model parameters to the AdamW optimizer.
Pass the optimizer to Trainer.

Praktische interactieve oefening

Probeer deze oefening eens door deze voorbeeldcode in te vullen.

# Pass the model parameters to the AdamW optimizer
optimizer = ____(params=____.____())

# Pass the optimizer to Trainer
trainer = Trainer(model=model,
                  args=training_args,
                  train_dataset=train_dataset,
                  eval_dataset=validation_dataset,
                  ____=(____, None),
                  compute_metrics=compute_metrics)

trainer.train()

Code bewerken en uitvoeren