TDA4VM: Custom Dataset compilation using Edgeai-Benchmark

I am currently working with TDA4VM for doing object detection using Yolov5s6 model. I have trained with my custom dataset and converted into onnx format with prototxt file. I have to compile my model using benchmark Repo. The current scripts accepts coco Dataset. But I have dataset in Yolo format. How can I can convert the dataset loaders to accept yolo fomat annotation? Is there any way to compile model using my custom dataset using benchmark Repo? If yes kindly provide the possible ways.

Thank you