ErasureResearch/ErasingDiffusionModels

Model Upload and Contribution Guide

This guide assumes you have a Hugging Face account and access to the ErasingResearch organization.

Prerequisites

• Account: Ensure you’ve joined the ErasingResearch organization.
• Directory Structure: Store all models in a single directory (models can be organized in subfolders).
• Authentication: If prompted for login details, use:
  • Username: Your Hugging Face username
  • Password: Your Hugging Face write token

Setup

1. Create a .env File

In the project’s root directory, create a .env file with your Hugging Face API key:

HUGGINGFACE_API_KEY="your_write_token"

2. Enter Virtual Environment

Activate the Python environment with Pipenv:

pipenv shell

3. Initialize Git LFS

Run Git Large File Storage (LFS) initialization:

git lfs install

4. Upload Models

Run the upload script, specifying the path to your models folder:

python3 upload.py "path/to/models/folder"

Contributing Code

To add new code or update existing files without downloading large model weights:

1. Pull Latest Code

Pull the latest version from the repository to sync any recent changes (this step uses LFS but doesn’t download model weights):

git pull

2. Make Your Changes

Add or update code as needed.

3. Push Changes

Commit and push your changes:

git push origin