Yolov3: KeyError: 'training_results'

[Aria20155]

Hi, after I run train.py, I got this error:
(I downloaded yolov3-spp.pt from your google drive and opt.transfer=True )

Traceback (most recent call last):
File "/home/Aria/Desktop/pycharm-2019.1.3/helpers/pydev/pydevd.py", line 1758, in
main()
File "/home/Aria/Desktop/pycharm-2019.1.3/helpers/pydev/pydevd.py", line 1752, in main
globals = debugger.run(setup['file'], None, None, is_module)
File "/home/Aria/Desktop/pycharm-2019.1.3/helpers/pydev/pydevd.py", line 1147, in run
pydev_imports.execfile(file, globals, locals) # execute the script
File "/home/Aria/Desktop/pycharm-2019.1.3/helpers/pydev/_pydev_imps/_pydev_execfile.py", line 18, in execfile
exec(compile(contents+"\n", file, 'exec'), glob, loc)
File "/home/Aria/Documents/yolov3-master/train.py", line 352, in
accumulate=opt.accumulate)
File "/home/Aria/Documents/yolov3-master/train.py", line 127, in train
if chkpt['training_results'] is not None:
KeyError: 'training_results'

[glenn-jocher]

@Aria20155 ah yes this makes sense, we need to update all of the *.pt models in this Google Drive folder with the updated dictionaries. In the meantime you can use the darknet yolov3-spp.weights file, which has identical weights to the yolov3-spp.pt.

The repo will automatically download darknet weights as needed, i.e. simply run python3 detect.py.

[glenn-jocher]

@Aria20155 you might also be able to simply do this conversion yourself. If you have downloaded yolov3-spp.weights then you should be able to convert it to a correct *.pt model with the following python code:

from models import *; convert('cfg/yolov3-spp.cfg', 'weights/yolov3-spp.weights')

[glenn-jocher]

@Aria20155 the *.pt models have been corrected and are now being uploaded to the Google Drive folder. Please wait about 15 minutes and then try to download the updated models. Thank you!