/tree-prompt-experiments

Create a tree of prompts during training that improves efficiency and accuracy.

Primary LanguageJupyter NotebookMIT LicenseMIT

Tree Prompting Experiments

Tree Prompting: Efficient Task Adaptation without Fine-Tuning, code for the Tree-prompt paper.

This repo contains code for reproducing experiments in the Tree-prompt paper. For a simple, easy-to-use interface, see https://github.com/csinva/tree-prompt.

Reproducing experiments

Organization

  • tprompt: contains main code for modeling (e.g. model architecture)
  • experiments: code for runnning experiments (e.g. loading data, training models, evaluating models)
  • scripts: scripts for running experiments (e.g. python scripts that launch jobs in experiments folder with different hyperparams)
  • notebooks: jupyter notebooks for analyzing results and making figures
  • tests: unit tests

Setup

  • clone and run pip install -e ., resulting in a package named tprompt that can be imported
    • see setup.py for dependencies, not all are required
  • example run: run python scripts/01_train_basic_models.py (which calls experiments/01_train_model.py then view the results in notebooks/01_model_results.ipynb