Questions
- Which two algorithms are available in TPOT with regard to neural networks?
- Approximately how many times are neural network classifiers slower to train than the default, scikit-learn ones?
- List and briefly explain the different hyperparameters available when training models with TPOT and neural networks.
- Can you specify a custom range of hyperparameter values when training custom neural network models with TPOT? If so, how?
- How can you find the best-fitted pipeline after the model has finished training?
- What are the advantages of using a GPU runtime such as Google Colab when training neural network models with TPOT?
- Describe why a single neuron in the multi-layer perceptron model can be thought of as logistic regression.
