DexForceAI/embodichain_reports

EmbodiChain Technical Report Assets

Public asset repository: DexForceAI/embodichain_reports.

This repository stores report-specific benchmark inputs, published measurement artifacts, figures, and videos for EmbodiChain technical reports. It starts with documentation and an empty asset index: no asset bundles are available yet. The download design specifies future integration; the Python helper and benchmark download commands are not implemented by this initialization.

Ownership

• Benchmark scripts, suite configurations, and runtime implementations remain in DexForce/EmbodiChain.
• General robot and scene assets continue to use the existing EmbodiChain asset resolver and embodichain_data.
• Model weights continue to use their existing model repositories and download helpers. This repository does not duplicate those assets or change their access requirements.
• This repository owns report-specific inputs and supporting results/media. Downloading benchmark inputs will not automatically download results or media.

The three initialization files are maintained together in scripts/benchmark/tech_reports/assets/ in the code repository. Publish those three files to this dataset's root together; asset directories are created only when real files are uploaded. The code repository copy of index.json is the empty bootstrap template; after assets are published, the dataset index at the selected commit is authoritative and must not be replaced with that empty template during documentation updates.

Benchmark source assets

The code repository reserves these directories under scripts/benchmark/tech_reports/assets/ for small authored inputs, input descriptions, and references:

shared/
loading/
rendering/
physics/
expert_generation/
demonstrations/

These match the L/R/P/G/D families in issue #679. Put future experiment-specific material under the appropriate family and experiment ID; shared definitions belong in shared/. The directories initially contain only Git placeholders. They are source organization, separate from the dataset's report layout below. Publish the three named bootstrap files explicitly, not the entire source directory or its .gitkeep files.

Large input binaries remain in the selected asset repository and download to the external cache. Generated run artifacts and reports belong in outputs/benchmarks/. The first real input publication will define the matching bundle IDs and hashes; this scaffold does not supply datasets or measured results.

Layout

README.md
DOWNLOAD_DESIGN.md
index.json
reports/<report-id>/
  README.md
  inputs/<benchmark-id>/...
  results/<benchmark-id>/...
  figures/...
  videos/...

The reports/ subtree above is a convention, not a claim that those files already exist. Hugging Face may also maintain repository metadata such as .gitattributes.

Access

The dataset is public and supports anonymous downloads. Publishing changes requires an account or token with write access to this repository. For the project's existing huggingface_hub>=0.35.3,<1.0 dependency, publishers can use hf auth login or provide HF_TOKEN through the execution environment. Credentials must not be written into report files, commands recorded in reports, or benchmark provenance. Downloads use https://huggingface.co directly.

A client can retrieve the current empty index without logging in using the existing Hugging Face SDK; this is not the proposed EmbodiChain download helper:

from huggingface_hub import HfApi, hf_hub_download

repo_id = "DexForceAI/embodichain_reports"
endpoint = "https://huggingface.co"
revision = HfApi(endpoint=endpoint, token=False).dataset_info(repo_id).sha
index_path = hf_hub_download(
    repo_id=repo_id,
    repo_type="dataset",
    filename="index.json",
    revision=revision,
    endpoint=endpoint,
    token=False,
)

For reproducible benchmark runs, use a tested full commit SHA instead of discovering the latest revision. See the version and cache contract.

Publish report assets

1. Assign a stable report ID and benchmark ID. Prepare the input files and supporting outputs under the corresponding report directories.
2. Add reports/<report-id>/README.md recording the report title/version, EmbodiChain Git commit, benchmark/suite identity, exact run commands, seeds, hardware/software environment, input sources, and output interpretation.
3. Record each file's source and license in that report README, with attribution or license files where required. No blanket asset license is assigned by this initialization; the code repository's license does not automatically license report data or third-party media.
4. Add input bundles to index.json following the index contract. Compute byte sizes and SHA-256 values from the actual uploaded bytes. Keep results/media outside input bundles.
5. Publish inputs, report metadata, and their index entries in one dataset commit. Review files and checksums at that commit before declaring it usable.
6. Record the verified dataset commit in the report release metadata and in the future benchmark integration's tested default. An optional release tag is a human-readable alias; record the resolved full SHA with results.

Later corrections create a new dataset commit and a new validated pin. Report publication, migration of historical assets, and implementation of the download interfaces are separate follow-up work.