Autonomous Generation of Reactions and Species-VAE

Machine Learning Project on generating Chemical Reactions using Auto Encoder Networks

Dependencies

python = ">=3.7.1, <3.9"

keras = "^2.6.0"

numpy = "1.19.5"

tensorflow = "^2.6.0"

tensorflow-addons = "^0.14.0"

scipy = "^1.7.1"

pathlib = "^1.0.1"

ipykernel = "^6.4.1"

pandas = "^1.3.3"

Installation

Step 1.

Clone the repository locally

Step 2.

Create a virtual environment with the necessary dependencies.

Step 3.

Create the tokenized dataset by running src/create_tokenized_dataset.py

Train the model

To train the model and generate new equations run

python main.py

This will save the model, encoder and generator separately inside the models directory. This allows for the user to load the generator at any time to continue generating new equations.