whulc
whulc
The channels number of the offsets is group*kernelsize*kernelsize*2, but what's the order of the channels? In other word, how the kernels of offsets flattened to 1D channel?
When I tried to train a model on multiple GPUs using nn.DataParallel, I met the error 'Running time error: argument are located on different GPUs', the error was located to...
I try to evaluate the FSNS using my own datasets. It seems that the image size must be 150*150, and there is something wrong with the .csv file, which I...
The offsets layer has num_group*kernel_width*kernel_height*2 channels, but how the 4D offsets flattened to 1D channel?
I have installed micmac and it works on some images. But for some other images, for a patch of size 800*800, it takes hours and doesn't get a output, the...
Hi I tried to use the model_inference.py. But met such error: `libprotobuf ERROR google/protobuf/text_format.cc:274] Error parsing text-format caffe.NetParameter: 1452:12: Message type "caffe.LayerParameter" has no field named "bn_param". Check failed: ReadProtoFromTextFile(param_file,...
I try to reproduce your experiment on icdar dataset, but meet the following problem:    It seems that the data is not imported successfully, but...
I tried to use opensfm to reconstruct some images takes by a car camera with accurate GPS. As the number of images is quite large, it takes a lot of...
To make dataset for training, I filtered the images with a Gaussian filter, and then down sampled the images to get low resolution images, But there is still a lot...
I have taken some photos and use colmap to reconstruct them. How can I use the code to generate the depthmap？