-
# TL;DR
Topics greatly improve the discoverability of repos; please add the short code from the table below to the topics of your repo so that ministries can use GitHub's search to find out what repo…
-
# TL;DR
Topics greatly improve the discoverability of repos; please add the short code from the table below to the topics of your repo so that ministries can use GitHub's search to find out what repo…
-
# TL;DR
Topics greatly improve the discoverability of repos; please add the short code from the table below to the topics of your repo so that ministries can use GitHub's search to find out what repo…
-
### 🐛 Describe the bug
Custom operators registered via `torch.library.custom_op` seem to be much slower than ops registered via `torch.Library.define` + `torch.Library.impl`
```
import torch
@…
-
# TL;DR
Topics greatly improve the discoverability of repos; please add the short code from the table below to the topics of your repo so that ministries can use GitHub's search to find out what repo…
-
# TL;DR
Topics greatly improve the discoverability of repos; please add the short code from the table below to the topics of your repo so that ministries can use GitHub's search to find out what repo…
-
# TL;DR
Topics greatly improve the discoverability of repos; please add the short code from the table below to the topics of your repo so that ministries can use GitHub's search to find out what repo…
-
### 🐛 Describe the bug
When I run this code on a tensor of type float16 of shape (1, 3, 256, 256) on torch.nn.InstanceNorm2d, I get a tensor with nan values from cpu but gpu can produce non-nan ten…
-
# TL;DR
Topics greatly improve the discoverability of repos; please add the short code from the table below to the topics of your repo so that ministries can use GitHub's search to find out what repo…
-
# TL;DR
Topics greatly improve the discoverability of repos; please add the short code from the table below to the topics of your repo so that ministries can use GitHub's search to find out what repo…