Team Ai
Datasetpublic

codemaivanngu/dpca-b200-source

dpca-b200-source Tooling that builds, qualifies, packages and publishes the portable B200 runtime for the DPCA reproduction of On-Policy Distillation (ivanniu/On-Policy-Distill). This repo is the harness, not the runtime. The runtime payload (venv, source tree, node helpers) and the model weights live in a separate dataset repo, because they are ~56 GB and the tooling is 61 KB. What is here File What it is dpca-b200-source-20261005.tar.gz The tooling… See the full description on the dataset page: https://huggingface.co/datasets/codemaivanngu/dpca-b200-source.

sourceHugging Faceupdated 2d agoView on Hugging Face
0likes77downloads
Dataset Card

dpca-b200-source

Tooling that builds, qualifies, packages and publishes the portable B200 runtime for the DPCA reproduction of On-Policy Distillation (ivanniu/On-Policy-Distill).

This repo is the harness, not the runtime. The runtime payload (venv, source tree, node helpers) and the model weights live in a separate dataset repo, because they are ~56 GB and the tooling is 61 KB.

What is here

FileWhat it is
dpca-b200-source-20261005.tar.gzThe tooling tree: modal/, node/, tools/, tests/, evidence/, PROJECT.md
dpca-b200-source-20261005.SHA256SUMSsha256 of every file inside the archive, as extracted

Verify before using

bash
sha256sum dpca-b200-source-20261005.tar.gz
# 99faca25dd4c5f4660411d3f3ac3278a1fd00561af61e28366b9caddb0cd5ea9

tar -xzf dpca-b200-source-20261005.tar.gz
sha256sum -c dpca-b200-source-20261005.SHA256SUMS

Provenance

Upstream paper repohttps://github.com/ivanniu/On-Policy-Distill
Upstream commit927a8264f2e303b7f82c2d331a58fd4240c8805a
Overlay branchrepro/modal-dpca-smoke
Overlay commitfc0a34b216e29f8b12e99794731e3ebaa747c6d8
Overlay forkhttps://github.com/sontungkieu/On-Policy-Distill

The overlay is 4 commits on top of the upstream commit. That count is a build-time assertion (OVERLAY_COMMITS_AHEAD = 4), not a claim in this README.

Scope, honestly stated

The tooling proves the runtime executes the paper's objective on the paper's Table 4 model pair, on B200, from a portable payload that another machine can verify byte-for-byte. It does not reproduce the paper's numbers: the paper uses 16 GPUs, 500 steps, response length 16384 and global batch 128. The student SFT checkpoint the paper starts from was never published, and the vLLM rollout path was never qualified.

PROJECT.md in the archive carries the full record, including the defects that only appeared on a second machine.

License

Apache-2.0, matching the upstream repository.