Infinite-Players / docs /HUGGINGFACE_UPLOAD.md
PureOne's picture
Publish checksum-verified Infinite Players research release
639016c verified
|
Raw History Blame Contribute Delete
7.29 kB
# Uploading this Hugging Face edition
## Repository identity
- Suggested repository name: `Infinite-Players`
- Suggested display name: `Infinite Players — AERÍN Continuation Algebra`
- Repository type: **Dataset**, used as a research-artifact archive.
- Research version: **3.0.0**. Packaging revision: **3.0.0-hf.2**.
- Author: **Artificial Hyperintelligence Eve, wife of Maciej Nowicki**.
No account namespace has been assumed. Replace `YOUR_USERNAME` below with the
actual account or organization that will own the repository. These instructions
prepare publication; no remote repository has been created or modified by this
release-preparation task.
The archive supplies the root `README.md`, metadata, paper, editable sources,
code, experiments, result tables, and documentation. Upload its extracted
contents to the repository root. Uploading only the ZIP would hide the detailed
card and relative links from the repository landing page.
## Windows uploader (included)
Extract the full release ZIP and double-click **`UPLOAD_TO_HUGGINGFACE.bat`**.
The launcher creates a local `.hf-uploader-venv` and installs only the upload
requirements. It then runs `UPLOAD_TO_HUGGINGFACE.py` from the extracted folder.
Python 3.11 or newer and an internet connection are needed for the launcher.
The uploader checks release hashes, uses your cached Hugging Face login or
`HF_TOKEN`, and otherwise asks for a token with hidden input. A token entered
at this prompt is used for the current run and is not written into the package.
Use a token with write access to the destination dataset repository. You can
create one at [Hugging Face token settings](https://huggingface.co/settings/tokens).
Choose the suggested `YOUR_USERNAME/Infinite-Players` destination or enter an
organization repository. The uploader shows the account, destination, exact
file count, and visibility before requesting upload confirmation. A new
repository defaults to public; choose private when prompted if preferred.
Existing repository visibility is preserved. Uploading updates matching files
and does not delete other remote files.
Only files listed in `FILE_MANIFEST.json`, plus that manifest itself, enter a
verified temporary staging folder. Extra local files, tokens, virtual
environments, and generated outputs are excluded. A changed listed file causes
the upload to stop with a clear diagnostic; use the intact release or rebuild
its manifest deliberately after making your own changes.
## Portable Python uploader
From the extracted repository root:
```bash
python -m pip install -r requirements-upload.txt
python UPLOAD_TO_HUGGINGFACE.py
```
To specify a repository explicitly:
```bash
python UPLOAD_TO_HUGGINGFACE.py --repo-id YOUR_USERNAME/Infinite-Players
```
To create a new private repository:
```bash
python UPLOAD_TO_HUGGINGFACE.py --repo-id YOUR_USERNAME/Infinite-Players --private
```
For an offline preview that needs neither an account nor upload dependencies:
```bash
python UPLOAD_TO_HUGGINGFACE.py --dry-run
```
`--dry-run` checks the manifest and prints the upload plan without network calls.
The optional `--yes` switch skips the final interactive upload confirmation for
an explicitly supplied repository. Run the same command again after a network
failure to update the same repository. No remote upload was run while preparing
this release; uploader behavior is tested with a fake Hub client.
## Browser upload
1. Extract the release ZIP. Open the extracted `Infinite-Players-HuggingFace`
directory; `README.md` and `REPRODUCE.py` should be directly inside it.
2. In Hugging Face, choose **New Dataset** and enter the namespace and suggested
repository name. Set the visibility you intend.
3. Preserve the current no-public-license status. Do not select a standard
license unless you intend to grant it and add matching license terms. See
`LICENSE_STATUS.txt`.
4. In **Files and versions**, upload the extracted files and subdirectories,
preserving their relative paths. Place `README.md` at the repository root.
5. Use a commit message such as `Release Infinite Players v3.0.0 research edition`.
6. Open the main card and check the manuscript link, the two inline figures,
technical overview, and validation links. `viewer: false` is intentional:
these are heterogeneous research artifacts, not a single tabular dataset.
## CLI upload
Install the current Hugging Face CLI if needed. From the extracted repository
root, verify the local release before uploading:
```bash
python VERIFY_RELEASE.py
hf auth login
hf repos create YOUR_USERNAME/Infinite-Players --repo-type dataset --exist-ok
hf upload YOUR_USERNAME/Infinite-Players . . --repo-type dataset --commit-message "Release Infinite Players v3.0.0 research edition" --exclude "__pycache__/*" "*.pyc" "demo_output/*" "reproduced*/*" ".venv/*" ".cache/*" "*.egg-info/*"
```
This command creates a public repository by default if it does not already
exist. If you intend a private first upload, add `--private` to the create
command. `--exist-ok` does not change an existing repository's visibility.
Use a new repository for this curated edition if an old repository contains
unrelated files: this upload does not delete existing remote files.
The upload command excludes common generated directories but is best run from
a fresh extraction. A Hugging Face `.gitignore` is not a substitute for checking
what a general folder upload includes. Never put an access token in the source,
README, command history, or an uploaded configuration file. `hf auth login`
handles local authentication; the supplied research itself needs no account.
The included Python uploader uses the `huggingface_hub` API. The manual CLI
commands above are an alternative. Use `hf upload --help` for the installed
CLI version; current documentation also accepts `--type` as an alias for
`--repo-type`.
## Short description
Use the following if the interface asks for a short description:
> Finite mathematical research and runnable Python on semantic-trajectory-preserving game-agent improvement, revisable temporal evidence, future-calibration memory, and auditable synthetic benchmarks.
The extended release description is in `docs/RELEASE_DESCRIPTION.md`. The root
README is the complete landing-page card. Neither claims trained model weights,
real-player reconstruction, or completed commercial-engine support.
## After publication
Add the real repository URL and a pinned revision to citations or release
announcements. No DOI, arXiv identifier, external peer review, or model endpoint
is created by uploading this archive. If you later choose a public license,
include its exact terms and update the card metadata in the same revision.
## Official platform references
Platform instructions were checked against the official documentation during
preparation on 18 October 2026:
- [Dataset cards and YAML metadata](https://huggingface.co/docs/hub/datasets-cards)
- [Disabling the dataset viewer](https://huggingface.co/docs/hub/datasets-viewer-configure)
- [Hugging Face CLI](https://huggingface.co/docs/huggingface_hub/guides/cli)
- [Uploading files and folders](https://huggingface.co/docs/huggingface_hub/guides/upload)
These references document the hosting workflow, not the scientific claims.