Datasets:
Formats:
json
Languages:
English
Size:
< 1K
Tags:
recursive-self-improvement
program-synthesis
term-rewriting
automated-reasoning
formal-methods
benchmark
| # Hugging Face dataset upload | |
| Extract the complete ZIP before running the uploader. Python 3.10 or newer is required. The research itself does not require Hugging Face access. | |
| ## Windows | |
| Double-click `UPLOAD_TO_HF.bat`. It selects an available Python 3.10+ installation, creates `.venv-hf`, installs `huggingface_hub`, and launches the uploader. Enter a repository ID such as `YOUR_USERNAME/nhais-dq-ii-autonomous-rsi` when asked. | |
| The uploader uses `HF_TOKEN`, an existing Hugging Face login, or a hidden token prompt. A token must have write access to the chosen dataset repository. It is never saved by the uploader. Do not put a token inside this release folder. | |
| New repositories are **private by default**. Double-click `UPLOAD_PUBLIC_TO_HF.bat` to opt into a public upload, or use this terminal command: | |
| ```bat | |
| UPLOAD_TO_HF.bat --repo-id YOUR_USERNAME/nhais-dq-ii-autonomous-rsi --public | |
| ``` | |
| Existing repository visibility is preserved. Upload to an existing public repository also requires `--public`. This flag does not change an existing private repository to public; that can be done in its Hugging Face settings after review. | |
| ## Python on any platform | |
| ```bash | |
| python -m venv .venv-hf | |
| # Activate the environment using the command appropriate for your shell. | |
| python -m pip install -r requirements-upload.txt | |
| python upload_to_hf.py --verify-only | |
| python upload_to_hf.py --repo-id YOUR_USERNAME/nhais-dq-ii-autonomous-rsi | |
| ``` | |
| For unattended use, pass `--non-interactive` with `--repo-id` and supply authentication through `HF_TOKEN` or an existing Hub login. The uploader does not accept a token on the command line, preventing accidental token exposure in shell history. | |
| ## What is uploaded | |
| Only files listed in `release-manifest.json`, plus the manifest itself, are copied into a temporary upload snapshot. Their SHA-256 hashes and sizes are checked before network writes. Unlisted local files, virtual environments, caches, and credentials are excluded. Symlinks and paths outside the extracted release are rejected. | |
| The dataset card is the root `README.md`. Matching remote paths are replaced by this release; remote files absent from the release are retained. Re-running the command is safe for resuming an interrupted upload. Do not edit release files before uploading, because hash verification will reject changes. Keep a fresh extracted copy if you want to run experiments that regenerate files. | |
| ## Troubleshooting | |
| - **Integrity check failed:** extract a fresh ZIP; run experiments in a separate copy. | |
| - **401:** supply a valid token. | |
| - **403:** check write permission for the dataset repository or organization. | |
| - **404:** confirm the namespace, dataset repository ID, and token scope. | |
| - **Concurrent update:** inspect the repository and retry. | |
| - **No Python found:** install Python 3.10+; no particular Python 3.12 installation is required. | |
| Official API references checked for this uploader: | |
| - https://huggingface.co/docs/huggingface_hub/package_reference/hf_api | |
| - https://huggingface.co/docs/huggingface_hub/guides/upload | |
| - https://huggingface.co/docs/huggingface_hub/guides/repository | |
| The uploader supports both the current `visibility` and legacy `private` repository-creation keyword. No upload was performed during release preparation. | |