error occurs running on gpu when i transplant the code to windows

[yeyuxmf]

I parse models/resnet/caffenet_train_iter_2000000.caffemodel under the caffe. I only transplant the part of sparse matrix calculations to caffe(for example, the caffe_gpu_sparse_dense2csr function interface),but CUSPARSE_CHECK () reports error when running to the caffe_gpu_sparse_mmcsr function interface。What parts do I need to transplant to run correctly? I need yorur help,Thank you.

[jspark1105]

Thanks for reporting this issue. Unfortunately, SkimCaffe doesn't support doing inference in GPU and I'm happy to help if anyone wants to contribute that part. The first step would be making it work correctly on GPU (perhaps following what is being done on CPU code path) and the next step would be implementing a customized CUDA kernel to take the full advantage of sparsity (because the current cusparse kernels are mostly optimized for sparse matrices found in scientific applications).