Open CoulsonZhao opened 4 years ago
right!
hi, I change your .pt model in BaiduYun into .weights (from torch model to yolo model) through convert.py. However, when I used
./darknet detector test cfg/oxfordhand.data cfg/prune_18.cfg prune_18.weights data/hand.jpg
in darknet, no hand target was detected. I have already generated oxfordhand.data oxfordhand.names and so on. Did you ever meet some problems like that? Thank you!
if you want to use the weight in darknet, you need to finetune (retrain) the weight in darknet
Got It!
for
python -c "from models import *; convert('cfg/yolov3.cfg', 'weights/last.pt')"
if you use the oxfordhand dataset, you'd better change that intopython -c "from models import *; convert('cfg/yolov3-spp-hand.cfg', 'weights/last.pt')"