This project is about using keras to build models.
- What is Keras?
- What is a model?
- How to instantiate a model (2 ways)
- How to build a layer
- How to add regularization to a layer
- How to add dropout to a layer
- How to add batch normalization
- How to compile a model
- How to optimize a model
- How to fit a model
- How to use validation data
- How to perform early stopping
- How to measure accuracy
- How to evaluate a model
- How to make a prediction with a model
- How to access the weights/outputs of a model
- What is HDF5?
- How to save and load a model’s weights, a model’s configuration, and the entire model
| File | Description |
|---|---|
| 0-sequential.py | Builds a neural network with the Keras library. |
| 1-input.py | Builds a neural network with the Keras library, without using Sequential. |
| 2-optimize.py | Sets up Adam optimization for a keras model with categorical crossentropy loss and accuracy metrics. |
| 3-one_hot.py | Converts a label vector into a one-hot matrix. |
| 4-train.py | Trains a model using mini-batch gradient descent. |
| 5-train.py | Previous task, but also analyzes validaiton data. |
| 6-train.py | Previous task, but also trains the model using early stopping. |
| 7-train.py | Previous task, but also trains the model with learning rate decay. |
| 8-train.py | Previous task, but also saves the best iteration of the model. |
| 9-model.py | Saves and loads an entire model. |
| 10-weights.py | Saves and loads a model's weights. |
| 11-config.py | Saves and loads a model's configuration in JSON format. |
| 12-test.py | Tests a neural network. |
| 13-predict.py | Makes a prediction using a neural network. |