The error about "ImportError: Skipping parameter for convW0!"

[impet14]

First of all, Thank you very much for the great contribution. Could you help me with my error? I have tried to use MSRA15 dataset for testing and use MSRA_network_prior_0.pkl and net_MSRA15_COM_AUGMENT.pkl for evaluating. Then I got the error as follow : Using gpu device 0: GeForce GTX 850M (CNMeM is disabled, cuDNN 5110) /home/john/worksp/deep-prior-pp/src/net/convpoollayer.py:261: UserWarning: DEPRECATION: the 'ds' parameter is not going to exist anymore as it is going to be replaced by the parameter 'ws'. pooled_out = pool_2d(input=conv_out, ds=poolsize, ignore_border=True, mode='max') Loading model parameters from ./eval/MSRA_network_prior_0.pkl Possibly not matching network configuration! ...... ..... ...... Traceback (most recent call last): File "test_realtimepipeline.py", line 81, in <module> rtp.processVideo(dev) File "/home/john/worksp/deep-prior-pp/src/util/realtimehandposepipeline.py", line 248, in processVideo self.initNets() File "/home/john/worksp/deep-prior-pp/src/util/realtimehandposepipeline.py", line 121, in initNets self.poseNet = ResNet(numpy.random.RandomState(23455), cfgParams=self.poseNet) File "/home/john/worksp/deep-prior-pp/src/net/resnet.py", line 340, in __init__ self.load(self.cfgParams.loadFile) File "/home/john/worksp/deep-prior-pp/src/net/netbase.py", line 474, in load raise ImportError("Skipping parameter for {}! Shape {} does not fit {}.".format(p.name, p.get_value().shape, v.shape)) ImportError: Skipping parameter for convW0! Shape (32, 1, 5, 5) does not fit (8, 1, 5, 5). Or anyone found the error same to mine, how do you solve it? I use theano 0.9.0, cuda 7.5, cudnn 5.1. Thank you very much. Sincerely,

[moberweger]

Hi @impet14 For MSRA, could you try PoseRegNet with type=11 instead of ResNet? The pretrained model could be for a different network.

[impet14]

Thank you very much.