Skip to content

[WIP] Subblock exp#66

Closed
sergeyf wants to merge 324 commits into
mainfrom
subblock_exp
Closed

[WIP] Subblock exp#66
sergeyf wants to merge 324 commits into
mainfrom
subblock_exp

Conversation

@sergeyf
Copy link
Copy Markdown
Collaborator

@sergeyf sergeyf commented Mar 17, 2026

Not done...

sergeyf and others added 30 commits September 10, 2021 09:05
Be more explicit about use_cache to avoid
No reference features for production model dump
As in, all the papers from the test blocks, not just the ones that get sampled
into the test triplets.  This is because we might want to run full S2AND later
with a model that was trained on the triplets, so if a test-block paper made it
into the triplet train set we'd have a problem.
atalyaalon and others added 27 commits April 21, 2026 11:37
- Move data/ into s2and/data/ so files ship with pip install
- Add [tool.setuptools.package-data] for data/*.json and data/*.txt
- Replace hand-rolled config loading in plotting_utils.py, scripts,
  and Dockerfile with canonical CONFIG from s2and.consts
- Update Dockerfile to remove redundant COPY lines and target
  S3 downloads to s2and/data/
- Bump version to 0.49.0 across all packages
Bundle data files in pip package and fix hardcoded paths
Call the reranker dataset staging and raw-similarity modules directly from the official rebuild path. Move staging-specific coverage out of the official-stack wrapper tests and keep S2AND relabel preflight covered at the staging boundary.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

6 participants