Repository navigation
Ratchet 0.5.0
Ratchet is an OpenAI-compatible gateway for license-clean distillation on open models. A small Nemotron 3 Nano 4B student with a LoRA adapter answers first; Nemotron 3 Ultra on Nebius Token Factory is the teacher. Every number below can be recomputed from the saved answers in evidence/ (see docs/VERIFY.md).
Results (measured, strict scoring)
Text-to-SQL (headline)
| Untrained | Teacher (Ultra) | Adapter | |
|---|---|---|---|
| Exam 1 (172 questions) | 12.8% | 25.6% | 39.0% |
| Exam 2 (200 questions, new industries) | 12.5% | 15.0% | 33.0% |
At the 95% precision bar the student answers 1 of 86 (exam 1) and 0 of 100 (exam 2) questions by itself, so most traffic still goes to the teacher.
Tool calling (second result)
| Untrained | Adapter | Teacher (Ultra) | |
|---|---|---|---|
| Hard exam, 150 questions, 20 unseen tools | 62.7% | 90.7% | 91.3% |
| BFCL, 300 questions | 77.3% | 80.3% | 88.3% (adapter is lower than Ultra) |
Adapters (SHA-256 of adapter_model.safetensors)
| Adapter | SHA-256 |
|---|---|
| Text-to-SQL, round 1 | a883bcf6bbf4152efb5a874520b8fa1a09d9bead55c172507725dc255f18236a |
| Text-to-SQL, round 2 | 38f99d441e0527b75fdf8fd9d10acbf4473dcf70db7febe3ad73fccd631a6aa7 |
| Tool calling, round 3 | fde47de3913fdf14d29328a174b5a92411c3f058989017c7709b4c572db324b9 |
What is real, and what is not
- Adapters were trained on a borrowed GPU, on synthetic data labelled by Nemotron 3 Ultra on Nebius Token Factory (OpenMDW-1.1).
- Teacher calls run on Nebius Token Factory. The live demo and the simulation run on Cloudflare.
- The Nebius AI Cloud path (Serverless Endpoint for the student, Serverless Job for Night School, a CPU VM and Object Storage) is scripted and tested offline, ready to run with
make nebius-planandmake nebius-up. - Simulation numbers are labelled Simulation.
- Not an official product of Nebius or NVIDIA, and not endorsed by either.
Try it
- Demo video (2:59): https://youtu.be/GIeSmDtopfs
- Live demo (real Token Factory calls, free, no sign-up): https://live.ratchet.dualmindlab.tech
- Judges page: https://ratchet.dualmindlab.tech/judges
- Simulation: https://demo.ratchet.dualmindlab.tech
- Docs: https://docs.ratchet.dualmindlab.tech
- Hugging Face collection (2 adapters, 6 datasets, 2 demos): https://huggingface.co/collections/harshbhx/ratchet
- Recompute the numbers: docs/VERIFY.md
What is new in 0.5.0
See the changelog: live and simulation demos side by side, a waking-up state for cold starts, Scout example questions and a runs-left counter, a one-command Nebius path (make nebius-plan), evidence, link and secret checks in CI, and post-deploy smoke tests.
License: Apache-2.0 for the code. Model weights and data keep their own licenses.