This project uses OpenAI's GPT-3.5 Turbo to solve mathematical word problems. It takes a set of prompts in the form of word problems, processes them using a predefined prompt structure, and provides answers based on the model's predictions.
-
Configure the Model: Edit
config.jsonto set the model parameters, API key, and file paths. -
Prepare Test Data: Ensure your test data is in
data/test.jsonlwith questions and ground truth answers. -
Run the Main Script: Execute
main.pyto load the model, process data, make predictions, and evaluate performance.
The model output, including predicted answers, will be stored in output/answers.json after each prediction.
Model performance can be evaluated using metrics such as accuracy. The evaluation script is included in utils.py.
- Python 3.x
- OpenAI GPT-3.5 Turbo API key
This project is built on OpenAI's GPT-3.5 Turbo and follows the prompt-based appr