Autodistill is an ecosystem for using big, slower foundation models to train small, faster supervised models. Using autodistill and its associated packages, you can go from unlabeled images to inference on a custom model running at the edge with no human intervention in between.
Images to inference with no labeling (use foundation models to train supervised models).
Use Florence 2 to auto-label data for use in training fine-tuned object detection models.
Use Segment Anything 2, grounded with Florence-2, to auto-label data for use in training vision models.
A template for use in creating Autodistill Target Model packages.
Use PaliGemma to auto-label data for use in training fine-tuned vision models.
