save weights error

[NJUFlipped]

Traceback (most recent call last):
File "train.py", line 108, in
model.save_weights('%s/%d.weights' % (opt.checkpoint_dir, epoch))
File "F:\cmp\VehicleDetection\PyTorch-YOLOv3\models.py", line 306, in save_weights
self.header_info[3] = self.seen
File "D:\software\Anaconda3\lib\site-packages\torch\nn\modules\module.py", line 532, in getattr
type(self).name, name))
AttributeError: 'Darknet' object has no attribute 'seen'

[DeokyunKim]

I have same error.

[nationalflag]

You should load a pretrained weights such as yolov3.weights first in train.py.

[DeokyunKim]

@nationalflag Could you write the code ? I didnt understand. thanks !

[nationalflag]

model = Darknet(opt.model_config_path)
model.apply(weights_init_normal)
model.load_weights(opt.weights_path)

Darknet.seen is assigned in load_weights

[eriklindernoren]

Hi, this is fixed in 6727086. You don't have to load pretrained weights to be able to save the weights of the model during training with this commit.

[poornimajd]

Hey can the save darknet weights function in models.py be used to save the weights in pytorch format ,i.e with the extension .pt ,or how is that function used?