---
pretty_name: jmtseng archival snapshot
license: other
size_categories:
- n>1T
---

# jmtseng archival snapshot

This public dataset repository stores an archival snapshot of
`/public/interns/hubin/jmtseng` as independent, uncompressed POSIX tar shards.
The source contains mixed research code, datasets, model weights, generated
artifacts, and Python environments. Licenses from each bundled upstream
project or asset continue to apply; this repository does not grant a blanket
license over third-party content.

The source is approximately 3.9 TB with more than 450,000 files. Ordinary
content is packed into approximately 2 GiB shards; individual source files
larger than that are stored in their own tar shard. Shards are grouped into
batch subdirectories for concurrent, resumable upload. See `RESTORE.md` for
restoration instructions.

Git metadata is excluded because repository history can retain deleted
credentials. Four archived third-party scripts containing hard-coded access
tokens are also excluded and listed in `REDACTED_PATHS.txt`.