addition-lstm

Model description

This repo contains the model which showcases the learning capabilities of LSTM using a simple example. A single-layer LSTM is made to learn to add two numbers, provided as strings. The model has been trained for adding two numbers where each number can have maximum of 5 digits.

Example: Input: "535+61" Output: "596"

Full credits to Smerity and others for this work.

Intended uses & limitations

More information needed

Training and evaluation data

The data consists of generation of two random 5 digit numbers as input and their sum as output. These numbers (and their sum) are encoded and fed as input to LSTM. The full data creation code is available within the example.

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.001
• train_batch_size: 32
• optimizer: {'name': 'Adam', 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-07, 'amsgrad': False}
• training_precision: float32
• num_epochs: 30

Training Metrics

Model Plot

<details> <summary>View Model Plot</summary>

</details>