Base Model
Basic Usages
-
Base Model
-
Ability to train a model on npy converted datasets
-
Ability to create predictions/deployment
-
Autosave functionality during training
Model Architecture
Basic architecture contained in the image_rec.py file
Contains it's own unique methods as well as acts as a baseline for use
Layers are as follows:
-
​ Convolutional Layer
-
Convolutional Layer
-
Convolutional Layer
-
Normalizer
-
Dense Layer
-
Dropout Layer
-
Dense Layer
-
Dropout Layer
-
Output Layer
Parameters can be adjusted as well to assist in learning
Training
Main function to train a model to a dataset
Programmer specifies parameters and the model will automatically train itself
Saving / Loading
Methods to save/load in models
Allows the programmer to save progress and use pre-trained models
Deployment
Prediction method for a fully-trained network
Takes in input data 'x' and will give a prediction based on the network