I have a unmodifiable caffe prototxt, which includes ConvolutionDepthWise layer. Since tensorrt not support ConvolutionDepthWise directly, I want to add nvinfer::IConvolutionLayer with groups param in PluginFactory so that it can replace ConvolutionDepthWise Layer. Is there anyway to do that?
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| how weights are passed to plugin layer -- SampleFasterRCNN | 7 | 906 | October 18, 2021 | |
| tf.nn.depthwise_conv2d() | 0 | 292 | September 9, 2019 | |
| Depthwise_conv2d implementation | 0 | 367 | September 10, 2019 | |
| Tensor RT supports caffe model layers | 28 | 10735 | October 18, 2021 | |
| About the python API parameter and 2D depthwise convolution | 0 | 790 | November 27, 2018 | |
| TensorRT 3 RC and grouped convolutions | 6 | 3740 | October 30, 2018 | |
| Whats the different between Deconvolution groups and deconvolutional layers? | 4 | 1724 | October 18, 2021 | |
| How do I implement R-FCN in TensorRT? | 21 | 1415 | October 18, 2021 | |
| What is the recommended way to run inference using a non-convolutional model in DeepStream pipeline? | 6 | 771 | May 6, 2023 | |
| Does tensorrt4 support 3D convolution layer and 3D pooling layer of Keras? | 1 | 817 | December 3, 2018 |