The Free Encyclopedia

Knowledge Distillation Lab (theLAB Genesis)

The Distillation Lab is a web app on your-gpu-host (/distillation/) that runs the full teacher → dataset → student pipeline from a browser. This documents the tooling; for the underlying concepts see Knowledge Distillation, End to End and Does Distillation Add Knowledge or Just Style?.

The pipeline

  1. Teacher — a large model generates high-quality completions for your prompts. The Lab's default teacher is glm-5.2:cloud.
  2. Dataset — those completions are captured as an instruction-tuning dataset (prompt + teacher response).
  3. Student — a small local model is fine-tuned on that dataset to imitate the teacher's behavior on the target task.

Mind the meter

glm-5.2:cloud burns Ollama Cloud Pro limits. Teacher generation is the expensive step — every dataset row is a cloud call. Check the usage.php ledger before and after a run so a large dataset doesn't blow the monthly quota.

Practical levers to control cost:

After the run

The usage.php ledger is your record of what each distillation run cost; review it as part of every session.