Mixed precision training with Trainer

You want to simplify your training interface for mixed precision training. As an alternative to training loops with Accelerator, you've decided to build Trainer since your model doesn't require custom training loops. Set up Trainer to use mixed precision training! The exercise will take some time to run with the call to trainer.train().

This exercise is part of the course

Efficient AI Model Training with PyTorch

View Course

Exercise instructions

Enable mixed precision training for Trainer.

Hands-on interactive exercise

Have a go at this exercise by completing this sample code.

training_args = TrainingArguments(
    output_dir="./results",
    evaluation_strategy="epoch",
    # Enable mixed precision training
    ____=____
)
trainer = Trainer(model=model,
                  args=training_args,
                  train_dataset=dataset["train"],
                  eval_dataset=dataset["validation"],
                  compute_metrics=compute_metrics)
trainer.train()

Edit and Run Code