signate-wandb-sync

A CLI tool to record SIGNATE competition scores to Weights & Biases (W&B).

Companion to signate-deploy — together they automate the full SIGNATE experiment tracking pipeline on GitHub Actions.

Installation

pip install signate-wandb-sync

Full Pipeline

[GitHub Actions]
  1. Download data via SIGNATE API  (signate-deploy)
  2. Run train.py — W&B run created, metrics logged
  3. Submit to SIGNATE              (signate-deploy)
  → W&B run URL printed to Actions log

[Local]
  4. Check score on SIGNATE leaderboard
  5. signate-wandb-sync score <W&B URL> --score 0.85 --rank 3

Add W&B to train.py

import wandb

run = wandb.init(project="my-signate-project", config={...})

# ... training and inference ...

wandb.log({"oof_score": oof_score})
print(f"W&B run URL: {run.url}")  # visible in Actions log
wandb.finish()

Set WANDB_API_KEY as a GitHub Secret:

- name: Train and predict
  env:
    WANDB_API_KEY: ${{ secrets.WANDB_API_KEY }}
  run: python train.py

Commands

signate-wandb-sync score

Log SIGNATE submission scores to a W&B run.

signate-wandb-sync score [RUN_ID] [OPTIONS]

Examples

# Full W&B URL (recommended - copy from the Actions log)
signate-wandb-sync score https://wandb.ai/your-entity/your-project/runs/abc123     --score 0.85 --rank 3

# With additional metrics
signate-wandb-sync score https://wandb.ai/your-entity/your-project/runs/abc123     --score 0.85 --rank 3     -m fbeta=0.85 -m recall=0.91

# Bare run ID (requires --project)
signate-wandb-sync score abc123 --project your-entity/your-project --score 0.85

Output:

Updated run: my-run-name (your-entity/your-project/abc123)
  submitted = True
  signate_score = 0.85
  signate_rank = 3

Windows

PYTHONUTF8=1 signate-wandb-sync score <run_id> --score 0.85

Authentication

W&B: run wandb login beforehand, or set WANDB_API_KEY environment variable.

Requirements

• Python 3.9+
• wandb

Related

• signate-deploy — Automate SIGNATE submission via GitHub Actions

License

MIT