This annotation format consists of two files: a text file containing all the annotations and a labelmap file mapping the integer class IDs to their human-readable names.
But you don't really need to worry about the details because Roboflow can both import and export this format!
Below, learn the structure of YOLO Keras TXT.
000001.jpg 178,84,230,143,1 111,144,134,174,1
000007.jpg 280,127,337,208,1 8,125,86,235,1 115,139,180,230,1 375,152,410,219,1 412,148,474,233,1 336,148,387,223,1 247,124,294,203,1 197,177,231,227,1 174,156,201,219,1 477,143,499,223,1
000012.jpg 75,20,174,128,1 284,131,303,155,1 381,189,407,215,1 242,32,258,55,1
000013.jpg 0,100,176,331,1
000014.jpg 437,174,487,233,1 326,184,379,248,1 228,249,310,352,1 372,120,402,169,1 326,137,364,185,1
000016.jpg 129,28,239,140,1
000017.jpg 150,102,184,144,1
000018.jpg 128,38,149,61,1
000019.jpg 254,70,283,107,0 107,71,137,101,0 181,66,215,111,0 145,66,174,104,0 303,61,349,121,0
000020.jpg 95,32,123,65,1 149,93,176,120,1
head
helmet
person
With Roboflow supervision, an open source Python package with utilities for completing computer vision tasks, you can merge and split detections in YOLO Keras TXT. Read our dedicated guides to learn how to merge and split YOLO Keras TXT detections.
Below, see model architectures that require data in the YOLO Keras 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.