RuntimeError： batched_detect_faces use_origin_size=False

xinntao / facexlib

FaceXlib aims at providing ready-to-use face-related functions based on current STOA open-source methods.

MIT License

791 stars 140 forks source link

RuntimeError： batched_detect_faces use_origin_size=False #50

Open azuredsky opened 2 months ago

azuredsky commented 2 months ago

bboxes = det_net.batched_detect_faces(batch_frames, conf_threshold=0.8, nms_threshold=0.15, use_origin_size=False)

File "python3.10/site-packages/facexlib/detection/retinaface.py", line 320, in batched_detect_faces frames = frames - self.mean_tensor RuntimeError: The size of tensor a (4) must match the size of tensor b (3) at non-singleton dimension 1

woctezuma commented 2 months ago

Maybe related:

https://stackoverflow.com/q/58496858/376454
https://github.com/StoryMY/take-off-eyeglasses/issues/1

TL;DR: RGBA vs. RGB.