Error: evaluate.py about tensorflow-2.x-yolov3 HOT 3 CLOSED

Can you first download COCO test dataset and move it to TensorFlow-2.x-YOLOv3/model_data/coco/ folder.
Download original yolov3 or yolov4 weights and place it to model_data folder.
Change TRAIN_CLASSES' to "model_data/coco/coco.names" Change TEST_ANNOT_PATH` to "model_data/coco/val2017.txt"

If everything works, then change YOLO_CUSTOM_WEIGHTS to your weights_file_name_location
Let me know if it works.

Also it's by default everything is configured to train and evaluate custom mnist dataset. You can try to train it for several epochs to test it out

from tensorflow-2.x-yolov3.

Can you first download COCO test dataset and move it to TensorFlow-2.x-YOLOv3/model_data/coco/ folder.
Download original yolov3 or yolov4 weights and place it to model_data folder.
Change TRAIN_CLASSES' to "model_data/coco/coco.names" Change TEST_ANNOT_PATHto "model_data/coco/val2017.txt" If everything works, then changeYOLO_CUSTOM_WEIGHTSto yourweights_file_name_location`
Let me know if it works.
Also it's by default everything is configured to train and evaluate custom mnist dataset. You can try to train it for several epochs to test it out

Thank you. I followed your steps. Works fine.!!

from tensorflow-2.x-yolov3.

Hello @pythonlessons ,
I have the same error so I changed YOLO_CUSTOM_WEIGHTS = "checkpoints/yolov3_custom" from False. I am training on my own set of images & classes. After I changed to this, the "ValueError: 'audio' is not in list" is not seen, but I am getting the unboundlocalerror "local variable 'save_directory' referenced before assignment"

from tensorflow-2.x-yolov3.