Project code for LIN205A Project -- UC Davis, Spring 2022

Single Word Model

There is a Google Colab demonstration notebook on how to run inference using our final model. There is also the data generation notebook

Multi-word span

There is a single notebook for generating the context span data and training the model. No inference.

Running the nearest pronunciation FST

See the single word model demonstration notebook. Didn't want to turn Kenji's code into a package and fiddle with making the FST pickling work. So I had to port most of the working code into google colab to make sure it worked there.

To-Do

• Finish Generative Grammar Vectors
• Geneate FST
• Decide on a dataset(s) for classification
• Generate Sequence Classifier
• Write Report
• Tidy up Presentation