Conditional Generation for NER with GPT

Python Package Requirements are:

• openai
• transformers
• datasets
• tqdm
• pinecone

In order to create JSONs for CONLL and WNUT, use the data.py file. In order to use KNN, one will have to setup a Pinecone Index. To do so, use the embed.py file.

Use the respective files for each method

• baseline.py
• knn.py
• taskmod.py
• domain_context.py
• qa.py

Refer to the notebooks for each method if something is amiss or not working as we spent the majority of our time using those.

Use metric.py or metric.ipynb