Getting Started
Start here if you're preparing your first run. This quickstart covers where to begin, what to prepare before payment, and what you receive after training completes. Use it to avoid retries and move faster.
Where to begin
- Go to Start Run, upload your dataset, and wait for validation results.
- Review eligibility and final price before payment.
What to prepare before starting
- Dataset file in .jsonl format (one JSON object per line). See Dataset Format.
- Dataset that fits BeaverYard's published size and token caps, plus the global 200,000 record-count cap and 20,000-character line-length cap. Final tier assignment happens automatically after upload and validation.
- High-quality training data for your own use case. We validate format and limits, but dataset quality remains your responsibility.
What you receive
- Downloadable artifacts (adapter_model.safetensors + adapter_config.json).
- Customer guides: Training Focus Summary and Usage Guidelines.