k-eak

> ##### Top > * channel: 512 > * n_block: 4 > * n_res_block: 5 > * res_channel: 512 > * n_cond_res_block: 0 > * n_out_res_block: 5 > * attention:...

@rosinality when I check the sample.py I noticed that F.one_hot function seem to be taking too much of time (190 seconds for top-level with batch size:32). I tried to change...

@rosinality Thank you for the suggestion. I replaced the one-hot and now I update it after each sample with the scatter function. Although this improved the speed compared one_hot, the...

@rosinality oh my bad, I needed to add torch.cuda.synchronize(). So my method does not change the speed that much and mostly saves a couple of seconds for large batches. I...

I think you need to shuffle the data as the score is calculated for 50 images each time. If all 50 images belong to the same class, the score will...

Dear sukritshankar, Thank you for your prompt response. I think I managed to get it work. However, I am now getting another error: I0316 10:54:52.287003 21115 layer_factory.hpp:77] Creating layer data...

@Feywell Hi Feywell, I did not fix that error (I am not familiar with Caffe) but I just switched to Tensorflow which is much more easier to configure.