Extras#

Testing output with backbones from the openpifpaf-extras package (install with pip3 install openpifpaf-extras).

Prediction#

Here, we try the pre-trained swin_s model.

%%bash
python -m openpifpaf.predict coco/000000081988.jpg --checkpoint=swin_s --decoder=cifcaf:0 --image-output=coco/000000081988.jpg.swin_s.predictions.jpeg
INFO:__main__:neural network device: cpu (CUDA available: False, count: 0)
INFO:openpifpaf.predictor:neural network device: cpu (CUDA available: False, count: 0)
INFO:openpifpaf.decoder.cifcaf:annotations 5: [14, 13, 13, 14, 13]
INFO:openpifpaf.predictor:batch 0: coco/000000081988.jpg
src/openpifpaf/csrc/src/cif_hr.cpp:102: UserInfo: resizing cifhr buffer
src/openpifpaf/csrc/src/occupancy.cpp:53: UserInfo: resizing occupancy buffer
IPython.display.Image('coco/000000081988.jpg.swin_s.predictions.jpeg')
_images/780fe970f0177a4e45e1263a767ad22198ddb489893d8a77b6c883fdfca822ff.jpg