MVIG-SJTU / AlphaPose

Real-Time and Accurate Full-Body Multi-Person Pose Estimation&Tracking System
http://mvig.org/research/alphapose.html
Other
8.02k stars 1.97k forks source link

load YOLO model error #131

Closed cswwp closed 6 years ago

cswwp commented 6 years ago

i have download the yolo mode file in baidu yun, and it recieved and error then run demo, it seems the model shape invalid, anyone help me?

This is the error info: conv_weights = conv_weights.view_as(conv.weight.data) RuntimeError: invalid argument 2: size '[128 x 384 x 1 x 1]' is invalid for input with 18431 elements at /opt/conda/conda-bld/pytorch_1532581333611/work/aten/src/TH/THStorage.cpp:84

cswwp commented 6 years ago

solved by download mode in Darknet github

Jeff-sjtu commented 6 years ago

Sorry for the wrong download link in Baidu Pan. It's for the old yolo version, we will update the link. Thanks.

asifahmedfuad commented 5 years ago

File "D:\YOLOv3\yolov3-master\models.py", line 342, in load_darknet_weights conv_w = torch.from_numpy(weights[ptr:ptr + num_w]).view_as(conv_layer.weight)

RuntimeError: shape '[128, 384, 1, 1]' is invalid for input of size 18431

Can anyone please suggest me a solution? Using cfg/yolov3-spp.cfg