- Switch different deep learning models and backends
(1) Current deep learning models available: CNN, LSTM
(2) Current backends available: Keras, Pyotrch, TensorFlow
- Data file location
- Data load function
- Explore the dataset
- Web service, flask server
Input: dataset & model name
Output: performance (accuracy, precision, recall, f1score)
- Main function
- K-fold cross-validation with the dataset
- Translate saved model file to .onnx file
- Original file using Keras
- Original file using Pytorch
- Original file using TensorFlow
- Manually separate train and test dataset
- Model architecture
Current architectures available:
CNN_Keras, LSTM_Keras, CNN_Pytorch, LSTM_Pytorch, CNN_TensorFlow, LSTM_TensorFlow
- Minst dataset using Pytorch
- Utilities