I have fine tuned SAM2 on building rooftop images and I want to check the IoU value to see if the model has been trained well and assess its precision. Is there a code to evaluate the IoU value or any precision/accuracy metric?
I have fine tuned SAM2 on building rooftop images and I want to check the IoU value to see if the model has been trained well and assess its precision. Is there a code to evaluate the IoU value or any precision/accuracy metric?
You can use supervision’s MeanAveragePrecision with metric_target = MetricTarget.MASKS