How to load pictures in C++ with Libtorch?

[RuFAI (John)]

Hi, the LibTorch library does not support any preprocessing functions for loading and resizing pictures etc. as torchvision does in python. What ist the best practice to do so in C++? I am aiming for an openCV approach. I this best practice or is there a better way?

Thank you

[RuFAI (John)]

Push. Nobody here who can answer the question?

[dambo (Shlomo)]

Libtorch C++ convert a Tensor to cv:Mat (single channel

From cv2 to TORCH: cv::cvtColor(frame, frame, CV_BGR2RGB); frame.convertTo(frame, CV_32FC3, 1.0f / 255.0f); auto input_tensor = torch::from_blob(frame.data, {1, frame_h, frame_w, kCHANNELS}); input_tensor = input_tensor.permute({0, 3, 1, 2}); From TORCh to cv2: input_tensor = input_tensor.to(at::kCUDA); torch::Tensor out_tensor = module->forward({input_tensor}).toTensor(); out_tensor = out_tensor.squeeze().detach().permute({1, 2, 0}); out_t…