Hi, I am currently running demo.py with the aim of converting my own trained yolov4-csp model from darknet to pytorch. I wanted to clarify if the parameters for the post processing after detections were the same.
This is from darknet. It has three params, thresh, hier_thresh and nms.
def detect_image(network, class_names, image, thresh=.5, hier_thresh=.5, nms=.45):
"""
Returns a list with highest confidence class and their bbox
"""
pnum = pointer(c_int(0))
predict_image(network, image)
detections = get_network_boxes(network, image.w, image.h,
thresh, hier_thresh, None, 0, pnum, 0)
num = pnum[0]
if nms:
do_nms_sort(detections, num, len(class_names), nms)
predictions = remove_negatives(detections, class_names, num)
predictions = decode_detection(predictions)
free_detections(detections, num)
return sorted(predictions, key=lambda x: x[1])
For this repo, I see conf_thresh and nms_thresh as parameters.
Hi, I am currently running demo.py with the aim of converting my own trained yolov4-csp model from darknet to pytorch. I wanted to clarify if the parameters for the post processing after detections were the same.
This is from darknet. It has three params, thresh, hier_thresh and nms.
For this repo, I see conf_thresh and nms_thresh as parameters.
The provided values are 0.4 for conf_thresh and 0.6 for nms_thresh.
Can I clarify if conf_thresh corresponds to thresh, and nms_thresh corresponds to nms? Is there an equivalent for hier_thresh?