https://github.com/csinva/tree-prompt-experiments
Create a tree of prompts during training that improves efficiency and accuracy.
https://github.com/csinva/tree-prompt-experiments
Last synced: over 1 year ago
JSON representation
Create a tree of prompts during training that improves efficiency and accuracy.
- Host: GitHub
- URL: https://github.com/csinva/tree-prompt-experiments
- Owner: csinva
- License: mit
- Created: 2023-02-07T21:48:45.000Z (over 3 years ago)
- Default Branch: main
- Last Pushed: 2023-12-13T22:40:46.000Z (over 2 years ago)
- Last Synced: 2025-03-17T19:11:38.908Z (over 1 year ago)
- Language: Jupyter Notebook
- Homepage:
- Size: 19.8 MB
- Stars: 5
- Watchers: 3
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: readme.md
- License: LICENSE
Awesome Lists containing this project
README
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`