Mwiza

Hi, When removing first order effects from the uncentered second order ALE, the paper published by Zhu & Apley subtracts the first order effects of each dependent variable from the...

BART: disable returning kv states since there exists an on device cache

1

# What does this PR do? Given that the kv cache is on device, there is no need to return `past_key_values`. However, this would require overriding the forward methods of...

Dont use constant mask if ynumel potentially overflows ygrids

6

If (ynumel / YBLOCK) > get_max_ygrids(), the z dimension will be used if znumel is None. However, if (ynumel / YBLOCK) % get_max_ygrids() != 0, there will be program launches...

### 🐛 Describe the bug Trying to generate a pointwise add kernel with 3 tiling groups fails. Reproducer: ```python import torch from torch._inductor.utils import run_and_get_triton_code from torch._inductor import config import...

### 🐛 Describe the bug Is symbolic execution expected to fail for kernels that have fallthrough registrations? Reproducer with a fallthrough for the mul operator: ```python import torch import torch.fx.experimental...