A simple CNN-LSTM deep neural model using Tensorflow to classify emotions from speech dataset
- Reading the 'train.json' file
- Rescaling the features
- Creating train & validation data splits
- Defining the DNN model
- Model training
- Evaluating Results
Install dependencies using:
pip3 install -r requirements.txt
- The notebook is executed online on kaggle.com with a GPU accelerator.
- To run it locally on a PC, 16GB of RAM is needed.
- Boost the 'Validation Accuracy' by Model modification or Data Augmentation
- email: navalekanishk@gmail.com