| # Upload this release to Hugging Face |
|
|
| Recommended repository name: **EVE-AUREOLITH-Omega** |
| Repository type: **Dataset** |
| Display title: **EVE–AUREOLITH Ω: Observable-Kernel Realizability Under Acoustic Actuation** |
|
|
| This folder is ready to upload. Its root `README.md` is the Hugging Face card; the intact manuscript, code, and certificates are in `research/`. Upload the contents of this folder, so that `README.md` sits at the repository root. Preparation has not created or published a remote repository. |
|
|
| ## Rights decision |
|
|
| The source release has no selected redistribution license. Its rights statement and this status are preserved. Decide the intended permissions before public distribution. If you choose a license later, update the applicable rights files and card metadata, and regenerate the release integrity manifest; do not merely tag the entire bundle with a license that covers only part of it. |
|
|
| ## Windows: automatic upload helper |
|
|
| 1. Extract the ZIP completely. Keep its folder structure intact. |
| 2. Open `UPLOAD_WINDOWS.bat` from the extracted folder. |
| 3. The helper uses the Windows Python launcher, installs `huggingface_hub` if missing, and requests sign-in through the official library if needed. Enter your token only into that authentication prompt. Never place a token in the repository. |
| 4. Enter `YOUR_USERNAME/EVE-AUREOLITH-Omega` when asked for the repository ID. |
|
|
| **Running the helper uploads the files and creates a public dataset repository if that name is unused.** For a private new repository, open a terminal in this folder and run: |
|
|
| ```bat |
| py -m pip install --upgrade huggingface_hub |
| py upload_to_huggingface.py --repo-id YOUR_USERNAME/EVE-AUREOLITH-Omega --private |
| ``` |
|
|
| For an existing repository, its existing visibility is retained. The helper adds or replaces matching file paths; it does not delete other remote files. Use a fresh name for this release if you want a clean standalone repository. |
|
|
| A local-only preview, with no login or network request: |
|
|
| ```bat |
| py upload_to_huggingface.py --repo-id YOUR_USERNAME/EVE-AUREOLITH-Omega --dry-run |
| ``` |
|
|
| ## Official CLI alternative |
|
|
| From this folder: |
|
|
| ```bash |
| python -m pip install --upgrade huggingface_hub |
| hf auth login |
| hf upload YOUR_USERNAME/EVE-AUREOLITH-Omega . . --repo-type dataset --commit-message "EVE-AUREOLITH Omega research release 1.0.0-hf.1" |
| ``` |
|
|
| The current official CLI uses `hf`; an old `huggingface-cli` command is unnecessary. The Python uploader is convenient when the `hf` executable is not on Windows PATH. |
|
|
| ## Browser alternative |
|
|
| Create a Dataset repository at https://huggingface.co/new-dataset and upload this folder's contents through its file uploader, preserving all subfolders. Use the root `README.md` as the dataset card. You can instead keep the manuscript PDF and source research ZIP as downloadable artifacts, but a ZIP-only upload will not expose the card or source tree correctly. |
|
|
| ## Check before and after upload |
|
|
| ```bash |
| python verify_release.py |
| ``` |
|
|
| After upload, check the rendered card, manuscript link, figures, and nested source files. The heterogeneous research collection intentionally has no automatic table viewer or fabricated train/test splits. No trained model, inference widget, or Space configuration is required. |
|
|
| Research reproduction writes diagnostic outputs. Use a working copy for those runs and a fresh extraction for an unchanged release upload. The checksum manifest establishes file integrity, not the correctness of the mathematical theorems. |
|
|
| Official documentation consulted for this edition: |
|
|
| - Dataset cards: https://huggingface.co/docs/hub/datasets-cards |
| - Scientific/custom-format datasets: https://huggingface.co/docs/hub/datasets-upload-guide-llm |
| - CLI authentication and uploads: https://huggingface.co/docs/huggingface_hub/guides/cli |
| - License metadata: https://huggingface.co/docs/hub/repositories-licenses |
| |
