When Udacity set out to create an open source self driving car, they released an open dataset to help researchers. Unfortunately, that dataset had problems (but Roboflow fixed them and released an updated dataset).
If you'd like to train a model with the original annotation files, you'll need to convert them to a format that your model expects. Roboflow will do that for you.
Below, learn the structure of Udacity TXT.
1478019952686311006.jpg 950 574 1004 620 0 "car"
1478019952686311006.jpg 1748 482 1818 744 0 "pedestrian"
1478019953180167674.jpg 872 586 926 632 0 "car"
1478019953689774621.jpg 686 566 728 618 1 "truck"
1478019953689774621.jpg 716 578 764 622 0 "car"
1478019953689774621.jpg 826 580 880 626 0 "car"
1478019953689774621.jpg 1540 488 1680 608 1 "car"
1478019953689774621.jpg 1646 498 1848 594 1 "car"
1478019954186238236.jpg 662 562 710 616 1 "truck"
1478019954186238236.jpg 686 576 730 628 0 "car"With Roboflow supervision, an open source Python package with utilities for completing computer vision tasks, you can merge and split detections in Udacity TXT. Read our dedicated guides to learn how to merge and split Udacity TXT detections.
Below, see model architectures that require data in the Udacity TXT format when training a new model.
On each page below, you can find links to our guides that show how to plot predictions from the model, and complete other common tasks like detecting small objects with the model.