Machine Learning 3: Model Evaluation

Summary

This lesson on ML model evaluation covered:

  1. Building an LSTM network and comparing it to our CNN
  2. Saving and loading trained models using model.save() and load_model()
  3. Understanding the output of our neural network, and using np.argmax to collapse probability distributions into a single, most-likely classification
  4. Comparing model predictions with correct classifications both by eye and via confusion matrix
  5. Wrapping model training in functions that track training progress and speed
  6. Automating the retraining of a model for different data splits to uncover which models show statistically significant improvement over another
  7. Using the recorded training history data to plot learning progress (and decide when no more epochs of training are needed)

Getting Help

Even if you’ve missed the session, you’ll be able to ask questions about whatever, whenever on our Discord. Feel free to drop your question in the #bootcamp channel or just message an instructor directly.

Recording

Ended up a bit pressed for time at the end there, but still made it through most of what you need to draw statistically significant conclusions about your models' performance and training.

Feedback

Don’t forget to fill in this week’s survey!