LoggerML

Rich machine learning logger in the console.

Tired of watching a dull, unreadable or uninformative terminal screen all day long to monitor your ML training? Use LoggerML to brighten up your day in colour and save coding time!

Github project: valentingol/logml.

Log your Machine Learning training in the console in a beautiful way using rich✨ with useful information but with minimal code.

Simply add the a logml.Logger object in your training loop to have a complete overview of your training in the console.

And then, you can simply customize your display to your liking, for instance like this:

Have a look on the Quick start section!

Contents:

• Installation
• Quick start
  • Minimal usage
  • Save the logs
  • Advanced usage
• Why not use tqdm.tqdm?
• Features
  • Regex for log config (styles, sizes and average)
  • Don’t know the number of batches in advance?
  • Use LoggerML with experiment tracker (like Weights and Biases)
• Logml API
  • logml.logger
    • Logger
      • Logger.log()
      • Logger.tqdm()
      • Logger.new_epoch()
      • Logger.start_epoch()
      • Logger.new_batch()
      • Logger.start_batch()
      • Logger.detach()
      • Logger.stop()
      • Logger.start()
      • Logger.reset()
      • Logger.get_vals()
  • logml.time_utils
    • sec_to_timestr()
    • get_time_range()
• How to contribute
  • Todo list
• License