Pharrell Yang

The keypoints may look like below. 

Hi, I just ran the ```pose_estimation.py``` to do sanity check and I also got this error. ```Traceback (most recent call last): File "hand_net.py", line 180, in print(PoseModel(19, 6, True, False))...

Hi, I still got error when compiled with CUDA. Here is the error message. -- Does not need to define long separately. -- std::exception_ptr is supported. -- NUMA is not...

Hi, can you list the commands for building with CUDA? I followed the installation guide but failed. Here is the error message I got. I don't have caffe2 installed on...

Thanks, Chayoung! I will take a look at it. On Wed, Jan 3, 2018 at 12:57 PM, Chayoung You wrote: > Hi, it's because whiteglass is using jekyll-archives > gem...

Also, I noticed that you didn't implement nn.Dropout2d for now while I have this op in my graph. So, any advice on this? Thanks!

@samhodge Hi, can you list the commands for building ATen with CUDA on Ubuntu? I followed the installation guide but failed. Here is the error message I got. -- Does...

I was wondering is it related to the precision difference when I converted pth.tar to h5?

Thanks, @warmspringwinds . I forgot to switch to eval mode and now the results are equal. Cheers! There is another problem with cpp version, which speed is few times slower...

@warmspringwinds Yes, the method to benchmark cpp version is the same as yours. Here is the code snippet: ```cpp auto net = torch::resnet18(); net->load_weights("../checkpoints/resnet18_no_drop.h5"); net->cuda(); // puts net on cuda...