Code for Hard Prompts Made Interpretable: Sparse Entropy Regularization for Prompt Tuning with RL

This code is built on the RLPrompt GitHub repository, which you can find "https://github.com/mingkaid/rl-prompt"

To use this code, you need the following versions of Python and PyTorch: Python >= 3.7 PyTorch >= 1.10.1 (install from the official website)

To install the core modules, run the following command:

pip install -e .

Refer to the examples folder for implementations of the following tasks, as described in our paper:

1. Few-shot classification
2. Text style transfer
3. Textual inversion from images