Download code/README.md from PratikBuilds/modelclinic-experiment-001: direct link, hf CLI and curl.
- Browser
- Download file 1.47 kB
-
https://huggingface.co/datasets/PratikBuilds/modelclinic-experiment-001/resolve/main/code/README.md
- Command line
-
hf download hf://datasets/PratikBuilds/modelclinic-experiment-001/code/README.md
-
curl -L -o README.md https://huggingface.co/datasets/PratikBuilds/modelclinic-experiment-001/resolve/main/code/README.md
Packaged source code
This directory contains the analysis, scoring, intervention, and run source needed for the archive.
src/: portable checkpoint, scoring, paired-analysis, diagnostics, and reconstruction scripts.modal/: the retained Modal GPU runner and runtime-validation scripts.tests/: retained unit-test source.requirements-analysis.txt: a small local CPU analysis environment.requirements-inference.txt: top-level package versions used by the recorded Modal inference setup. It is not a full transitive lockfile.
The packaged code retains the research logic. Path changes made for this archive are limited to the package’s code/src locations, the organized experiments/ and results/ tree, and an environment-configurable Modal volume name. Each Modal script reads MODELCLINIC_MODAL_VOLUME; if unset it expects a volume named modelclinic-cache. Create and configure your own volume before running inference. The original local volume identifier is not published.
Modal jobs use the container path /cache as their mounted volume and /root/project for code copied into the container. Those are runtime container paths, not host machine paths. No Modal app, credential, user profile, or workspace identifier is included.
For the exact science settings and commands see REPRODUCIBILITY.md. The package’s analysis-only workflow does not import Modal, require a Hub token, or launch inference.