search

Daniil-Osokin / lightweight-human-pose-estimation.pytorch

Fast and accurate human pose estimation in PyTorch. Contains implementation of "Real-time 2D Multi-Person Pose Estimation on CPU: Lightweight OpenPose" paper.
Apache License 2.0
2.09k stars 474 forks source link

Evaluation error #251

Closed ghost closed 2 years ago

ghost commented 2 years ago

I have encountered an error while trying to val.py on coco val2017 : 1/ I run this : python val.py --labels /content/lightweight-human-pose-estimation.pytorch/data/coco/annotations/person_keypoints_val2017.json --images-folder /content/lightweight-human-pose-estimation.pytorch/data/coco/val2017 --checkpoint-path /content/lightweight-human-pose-estimation.pytorch/checkpoint_iter_370000.pth

2/ after execution, I obtained this TypeError: 'numpy.float64' object cannot be interpreted as an integer

Daniil-Osokin commented 2 years ago

Hi! Do you need any help with this?

ghost commented 2 years ago

Yes. I didn't find the cause of this error ?

ghost commented 2 years ago

It works fine now after reimplementing the evaluation method. Thanks

Daniil-Osokin commented 2 years ago

Cool!