Closed VisImage closed 2 years ago
Hi, it hints that the jde.1088x608.uncertainty.pt might be corrupted. Can you try redownload and reupload it to the google drive?
Thank you for the suggestion. I re-download and re-upload the file jde.1088x608.uncertainty.pt and tried again. The same issue appears.
Actually, I noticed that my first customized cell also uses this same file " jde.1088x608.uncertainty.pt" and it worked fine. It might be another file corrupted.
! python3 scripts/demo_inference.py --cfg configs/coco/resnet/256x192_res50_lr1e-3_1x.yaml \ --checkpoint pretrained_models/fast_res50_256x192.pth --outdir /content/output_track \ --pose_track --sp --save_img --save_video --vis_fast --video /content/alphaPose_sample3.avi
Hi, we have fixed a bug of --detector tracker. You can try again now! @VisImage
And the file corruption of " jde.1088x608.uncertainty.pt" needs double check @VisImage, since the pytorch error message point to this file
Using the colab link provided in the repo, and went to the very end: works like a charm. :-)
"!python3 scripts/demo_inference.py with flag --pose_track --" on my own video file works fine. My own video file and osnet_ain_x1_0_msmt17_256x128_amsgrad_ep50_lr0 are uploaded to the colab folder.
"!python3 scripts/demo_inference.py with flag --detector tracker" hangs (frozen). file jde.1088x608.uncertainty.pt has been uploaded to //content/AlphaPose/detector/tracker/data/jde.1088x608.uncertainty.pt.
Two colab cells for the above two program-executions are shown below for your reference. Thank you.