Udacity Self Driving Car Dataset fixed-large
Export Created
5 years ago
2020-02-11 7:03am
Export Size
15000 images
Annotations
obstacles
Available Download Formats
COCO JSON
COCO JSON annotations are used with EfficientDet Pytorch and Detectron 2.
COCO-MMDetection
COCO-MMDetection JSON annotations are used with MMDetection.
CreateML JSON
CreateML JSON format is used with Apple's CreateML and Turi Create tools.
PaliGemma JSONL
PaliGemma format is used with Google's Multimodal Vision Model.
Pascal VOC XML
Common XML annotation format for local data munging (pioneered by ImageNet).
YOLO Darknet TXT
Darknet TXT annotations used with YOLO Darknet (both v3 and v4) and YOLOv3 PyTorch.
YOLO v3 Keras TXT
TXT annotations used with YOLOv3 Keras.
YOLO v4 PyTorch
Darknet TXT annotations used with YOLOv4 PyTorch (deprecated).
MT-YOLOv6
MT-YOLOv6 TXT annotations used with meituan/YOLOv6.
YOLO v5 PyTorch
TXT annotations and YAML config used with YOLOv5.
YOLO v7 PyTorch
TXT annotations and YAML config used with YOLOv7.
YOLOv8
TXT annotations and YAML config used with YOLOv8.
YOLOv9
TXT annotations and YAML config used with YOLOv9.
YOLOv11
TXT annotations and YAML config used with YOLOv11.
Tensorflow Object Detection CSV
CSV format used with Tensorflow (usually converted before training so you probably want to export as a TFRecord instead unless you need to inspect the human-readable CSV).
RetinaNet Keras CSV
A custom CSV format used by Keras implementation of RetinaNet.
Multiclass Classification
Converts your object detection dataset into a classification dataset CSV.
OpenAI CLIP Classification
Converts your object detection dataset a classification dataset for use with OpenAI CLIP.
Tensorflow TFRecord
TFRecord binary format used for both Tensorflow 1.5 and Tensorflow 2.0 Object Detection models.
Train/Test Split
Your images are split at upload time. Learn more.
Preprocessing Options
Applied to all images in dataset
Auto-Orient
Preprocessing can decrease training time and increase inference speed. Learn more on our blog.
Augmentation Options
Randomly applied to images in your training set
No augmentation steps were applied.
Augmentations create new training examples for your model to learn from. Learn more on our blog.