I have been following below document from tensorflow-lite model maker object detection api.
https://www.tensorflow.org/lite/tutorials/model_maker_object_detection
Following the documentation in the advance usage section, I have tried to import my custom dataset by uploading those in the google drive.
Both of the image files and the annotation files I have been able to find from openimage dataset.
https://storage.googleapis.com/openimages/web/download.html
Below line I am able to run without issue.
object_detector.DataLoader.from_pascal_voc(image_loc, annotations_loc, label_map={1: "person", 2: "notperson"})
Image_loc contains images in .jpg format and annotation_loc contains data in pascal voc xml format.
Unfortunately the return object of the above line is of type DataLoader.
How may I divides the dataloader output in train and test and run the different models starting with efficientdet-lite0 on those data?
