Mehran Maghoumi
Mehran Maghoumi
> It may be helpful to provide such loss class for those who are more comfortable with this pattern such as yourself. Thanks for bringing this up. No problem! That...
@mblondel First, thanks again for your great paper! I've been extensively using it for my research. I just publicly released a CUDA implementation of your algorithm for PyTorch that can...
I can confirm that this project works under Ubuntu 16.04, with CUDA v8.0 on a GTX 1080. I simply followed the build instructions and everything worked out of the box....
Heck, I was even able to get to work under Bash for Windows 10! In case anybody else wants to get it to work with Bash for Windows 10, you...
I used the one that's includes in this repo. On Dec 19, 2017 2:18 PM, "Neo" wrote: > @Maghoumi Hi Maghoumi, I also has the same > environment with you...
Yes this is expected and was discussed previously in #9 and #10. Leaving this issue open for anyone else in the future who might have the same question.
Please see #9, #10 and #27 I think I should add a note about negative loss values in the readme file to avoid confusion in the future.
What do you mean exactly? You mean loading the dataset?
This might be related to another issue we recently investigated where the memory usage went extremely high with 8 workers, but not with 4 workers. Ryan suspected some change on...
Hi @praxi-roshan, thanks for opening this issue and providing detailed repro steps. I just attempted running the tutorial on the latest `main` commit with a freshly setup environment and it...