I have my own MobileNet caffe model, and I encountered with the problem of custom layer of DepthWise Convolution. I saw you change your MobileNet model's dw_conv into group_conv, so I mimic your move and run with TensorRT 7.2.
It works, however, it has outputs like Weights for layer conv2_1/dw doesn't exist.. Is this the problem of the model or the problem of this move? Have you met with the same problem?
Hope to hear from you, thanks.
I have my own MobileNet caffe model, and I encountered with the problem of custom layer of DepthWise Convolution. I saw you change your MobileNet model's dw_conv into group_conv, so I mimic your move and run with TensorRT 7.2.
It works, however, it has outputs like
Weights for layer conv2_1/dw doesn't exist.. Is this the problem of the model or the problem of this move? Have you met with the same problem?Hope to hear from you, thanks.