-
Platforms: linux
This test was disabled because it is failing in CI. See [recent examples](https://hud.pytorch.org/flakytest?name=test_gelu_dynamic_shapes_cpu&suite=DynamicShapesCpuTests&limit=100) a…
-
### 🐛 Describe the bug
It seems that tensor.to(device) isn't copying the data correctly when it is moved from one GPU to another GPU.
Copying to/fro a GPU and a CPU seems to work as intended.
M…
-
1) microsoft/Phi-3-small-128k-instruct is not running on cpu (ICELAKE or Graviton)
the script which i was using :
![Screenshot 2024-07-26 165001](https://github.com/user-attachments/assets/e0b2…
-
There's CI for windows, but it fails. The only available builds are for `pytorch-cpu`, [v1.1.0](https://anaconda.org/conda-forge/pytorch-cpu/files) more than 1.5 years ago, nothing for the newer `pyto…
-
### 🐛 Describe the bug
https://github.com/pytorch/vision/pull/8463 introduced a platform/compiler specific cast. See this comment for more details. https://github.com/numpy/numpy/issues/23481#issue…
-
I tried to run this example https://github.com/intel/neural-compressor/blob/master/examples/3.x_api/pytorch/cv/static_quant/main.py,
and I got an error in https://github.com/intel/neural-compressor/b…
-
### System Info
```Shell
- `Accelerate` version: 0.29.2
- Platform: Linux-5.15.0-106-generic-x86_64-with-glibc2.35
- `accelerate` bash location: /usr/local/venv/bin/accelerate
- Python version: 3.…
-
This should reduce disk space usage as well as image build times.
-
I don't see any related MPS `max` reproducers quite this simple on the issue tracker, so figured this might help. Working on an M3 Max, Python `3.12` with `torch` `2.3.1`, I saw the `max` discrepancy …
-
### 🐛 Describe the bug
In some cases certain ops are slow. The failure mode does not always trigger, but its relatively easy to trigger this behavior if run in rapid succession.
Take the followi…