Adam Harrison

Commenting now with my personal account, I didn't realize I was using my work account, and we are moving away from that domain anyway, so it will soon be defunct....

This bug just bit me, still seems to be present. Reading and writing an untouched file nifti file should work, especially given the ubiquity of the format.

Hi @hjmjohnson Here is the metadata for the image I am loading and then just saving (based off of how `sitk` loads it). Anything pop out at you? It seems...

Thanks @gdevenyi - for some reason if I load and save the image, and then load the image in a viewer like itkSnap or mitkWorkbench, the saved image is reflected...

@blowekamp Thanks, I had tried that, but no difference. I think I've narrowed down what's happening here. In the upstream process of creating the image I was working with downstream,...

Haha well excellent timing on my part. That's very good to see :)

I also get this problem with my own dataset. The only way I could get the loss to stop exploding was to use learning rates of 1*10e-8, but I feel...

@brisker Notice the error says: "Could not open or find file ../../data/HED-BSDS/train/aug_gt_scale_0.5/157.5_1_0/159045.png" - the problem is it can't find an image in your training list. The image_labelmap_data_layer will keep running,...

@brisker For one reason or the other, it's unable to open the file. I would assume that something is wrong with the path - what do you have as root_folder...

@brisker The path to the train_pair.lst is given in your train_val prototx file, whereas the paths to the images are given in train_pair.lst and your root_folder - so the path...