-
### What happened?
In my app I am running into this. On ibis main, I can't repro. I'm guessing this has something to do with the combo of other libraries I have installed, eg pandas and duckdb. I fig…
-
Jupyter Notebook style
-
### Before You Report a Bug, Please Confirm You Have Done The Following...
- [X] I have updated to the latest version of the packages.
- [X] I have searched for both [existing issues](https://github.…
-
# Summary
This PR https://github.com/pytorch/pytorch/pull/108070 updates the pin in PyTorch from 3.1 to 3.2.1. We are currently failing for FlashAttention tests when doing this update.
The update …
-
### 🐛 Describe the bug
```python
In [52]: torch.pow(torch.tensor([1, 2]), True)
Out[52]: tensor([1, 2])
In [53]: torch.pow(torch.tensor([1, 2], device='cuda'), True)
---------------------------…
-
### 🐛 Describe the bug
`AdaptiveLogSoftmaxWithLoss` lacks of argument check on`div_value`, therefore it ends up with
C++ frontend: crash by FPE
Python frontend: ZeroDivisionError
Test code(C++):…
-
# Summary
When building with the following two settings on Main I am getting the following error:
``` Shell
/pytorch/torch/csrc/distributed/rpc/tensorpipe_agent.cpp
In file included from /home/dri…
-
### 🐛 Describe the bug
Hi! I can see that triton kernels work correctly only with eager mode. I am not able to compile functions containing triton kernels with torch compile. No matter what I would d…
-
### Required prerequisites
- [X] Make sure you've read the [documentation](https://nvidia.github.io/cuda-quantum/latest). Your issue may be addressed there.
- [X] Search the [issue tracker](https://g…
-
### 🐛 Describe the bug
Seems that the nested `true_fn` and `false_fn` subgraphs of `torch.cond` are not fully decomposed under the hood of the following `aot_autograd` call in Inductor:
https://…