cinjon

I also ran into the initial problem above, but have changed mmcv a bit so reverting to a prior version wasn't ideal. I changed `model = MMDistributedDataParallel(model.cuda())` to ``` model...

Actually, this is really a problem with `workflow`. When it is has more than one array item like in https://github.com/open-mmlab/mmaction/blob/master/configs/TSN/tsn_kinetics400_2d_rgb_r50_seg3_f1s1.py#L110, it fails because we aren't building the dataloaders for the...

I think this might be related to #15629, which has a similar level of response :/. @bschorchit any idea when we'll get support for this bug?

For reference, I tried the same settings using this library https://github.com/cyvius96/prototypical-network-pytorch and got the expected results. Having a lot of trouble getting miniimagenet to work with this library.

Also having this issue.

Is this correct? ``` {'params': {'backbone': {'kernel_size': 3, 'output_dim': 1024, 'input_dim': 3, 'stride': 2, 'padding': 1, 'out_img_size': 18}, 'primary_capsules': {'kernel_size': 3, 'stride': 2, 'input_dim': 1024, 'caps_dim': 64, 'num_caps': 16, 'padding':...

Do you mind verifying that it is actually correct please? That would be quite helpful so that I don't spend a week debugging when it's not actually right.

With the config as is, it reaches only 73% accuracy on the diverse multi mnist test set. That's notably lower than the result in the paper. It also is at...

Hi, was this implemented?

Sort of? I can get similar results to the original using the TF code that @wzmsltw sent (see below). However, trying to turn that into Pytorch has proven oddly difficult...