Lightning AI Studios: Never set up a local environment again →

Log in or create a free Lightning.ai account to track your progress and access additional course materials  

Deep Learning Fundamentals

  • Pages

    • Final certification exam

    Deep Learning Fundamentals > Unit 5 > Unit 5.6
    Course Progress:

    5.6 The Benefits of Logging Your Model Training

    Slides

    References

    Code

    Additional resources if you want to learn more

    For more information about other types of loggers available via the Lightning Trainer, I recommend checking the official documentation here.

    Log in or create a free Lightning.ai account to access:

    • Quizzes
    • Completion badges
    • Progress tracking
    • Additional downloadable content
    • Additional AI education resources
    • Notifications when new units are released
    • Free cloud computing credits

    Quiz: 5.6 The Benefits of Logging Your Model Training (Part 1)

    Logging is a good way for reducing overfitting.

    Incorrect. Logging can help you detect overfitting more easily, but you still have to address and reduce it by some other means.

    Correct. Logging can help you detect overfitting more easily, but you still have to address and reduce it by some other means.

    Please answer all questions to proceed.

    Quiz: 5.6 The Benefits of Logging Your Model Training (Part 2)

    In order to log our model training, we have to specify what we want to log in the

    Incorrect. The trainer will always log the model automatically.

    Correct. If we want to log anything in addition to the loss returned by the train_step method, we have to specify that via the self.log code in the LightningModule.

    Please answer all questions to proceed.

    Quiz: 5.6 The Benefits of Logging Your Model Training (Part 3)

    One advantage of the CSVLogger is that it makes it a tad easier to plot the log files via Pandas and matplotlib. What is, on the other hand, an advantage of using TensorBoard for logging?

    Correct. One advantage of the TensorBoard logger is that you can view the progress more easily live as the model trains.

    Incorrect. I would argue CSV files are always easier to archive as they are a more human-readable format and software-independent.

    Please answer all questions to proceed.
    Watch Video 1

    Unit 5.6

    Videos
    Follow along in a Lightning Studio

    DL Fundamentals 5: PyTorch Lightning

    Sebastian
    Launch Studio →
    Questions or Feedback?

    Join the Discussion