ml 0.1.0

A collection of machine learning utilities

ml

A collection of machine learning utilities.

Installation

pip install ml

A collection of utilities for ML/NLP research.

Utilities

ml.snapshot

Isolate experiment code on shared / networked filesystems (e.g. Slurm) by copying the current directory into a per-experiment snapshot and running a command inside the copy. Editing your source afterwards won't affect queued or in-flight jobs.

As a library (standard library only):

from ml.snapshot import snapshot

snapshot_dir = snapshot(
    "python train.py",
    exclude=["*.log", "wandb"],
    experiment_id="42",
)

As a CLI (requires the cli extra: pip install 'ml[cli]'):

ml-snapshot --experiment-id 42 --exclude '*.log' 'python train.py'

The base directory defaults to $SNAPSHOT_DIR (or snapshotted_experiments), and exclude patterns fall back to the comma-separated $SNAPSHOT_EXCLUDE.

Development

This project uses uv for dependency management.

# Install dependencies (including dev tools)
uv sync

# Run the test suite
uv run pytest

# Lint and format
uv run ruff check .
uv run ruff format .

License

Apache License 2.0