This repository contains the code supporting the DETR base model for use with Autodistill.

DETR is a transformer-based computer vision model you can use for object detection. Autodistill supports training a model using the Meta Research Resnet 50 checkpoint.

Read the full Autodistill documentation.

Read the DETR Autodistill documentation.

To use DETR with autodistill, you need to install the following dependency:

pip3 install autodistill-detr

from autodistill_detr import DETR

# load the model
target_model = DETR()

# train for 10 epochs
target_model.train("./roads", epochs=10)

# run inference on an image
target_model.predict("./roads/valid/-3-_jpg.rf.bee113a09b22282980c289842aedfc4a.jpg")

This project is licensed under an Apache 2.0 license. See the Hugging Face model card for the DETR Resnet 50 model for more information on the model license.

We love your input! Please see the core Autodistill contributing guide to get started. Thank you 🙏 to all our contributors!