Model Distillation Python Packages

autodistill

Images to inference with no labeling (use foundation models to train supervised models).

pecca

Replace repetitive LLM calls with small, auditable ML models. Record → Shadow → Live, LLM kept as fallback.

llm-distiller

Model distiller automator

insurance-distill

GBM-to-GLM distillation for insurance pricing - surrogate factor tables for Radar/Emblem rating engines

autodistill-gcp-vision

Autodistill Google Cloud Vision module for use in training a custom, fine-tuned model.

darklight

A framework for knowledge distillation using TensorRT inference on teacher network

autodistill-rekognition

Use AWS Rekognition to train custom models that you own.

autodistll-grounded-sam

Images to inference with no labeling (use foundation models to train supervised models).