-
### 🐛 Describe the bug
The `read_image` func ignores `ImageReadMode.GRAY` when reading WEBP images. It produces tensors with 3 color channels instead of 1.
Example image: [here](https://drive.goog…
-
The following code snippet demonstrates an issue with `torch.nn.functional.linear(...)` and `torch.nn.functional.conv2d(...)` when using automatic broadcasting on the GPU. Specifically, the outputs of…
-
### 🐛 Describe the bug
I was trying to install torchvision==0.19.0 from source.
After download zip file, for command: ` python setup.py install`, I got an error:
`error: ‘class torch::Library’ has…
-
### 🐛 Describe the bug
I have a fine-tuned TinyLlama/TinyLlama-1.1B-Chat-v1.0 model.
I created the checkpoint file using the following:
torch.save(model.state_dict(), "/opt/ml/model/model.pth")…
-
### 🐛 Describe the bug
I'm not sure if this is the intended behavior or a bug.
Using the nightly release (`torch-2.6.0.dev20241115+cu124`), when I run the following code (lifted out of [dynamo/t…
-
### 🐛 Describe the bug
Currently, torch.cuda doesn't support changing `CUDA_VISIBLE_DEVICES` on the fly, a demo:
```python
import os
import torch
os.environ["CUDA_VISIBLE_DEVICES"] = ""
prin…
-
### Your current environment
The output of `python collect_env.py`
```text
Collecting environment information...
PyTorch version: N/A
Is debug build: N/A
CUDA used to build PyTorch: N/A
ROC…
-
### 🐛 Describe the bug
I am building PyTorch from source on an aarch64 system and have noticed performance differences of Hugging face models(Bert, resnet etc) when running on 64 threads/cores, compa…
-
### 🐛 Describe the bug
fbgemm's CMakelists.txt has a -Werror set but when building it throws errors about VLAs being a clang extension, even though I'm using clang. Removing the Werror from fbgemm'…
-
### 🐛 Describe the bug
FP8 basic creation operations such as zeros / ones / full work under eager mode but not backend=inductor compile.
same error as https://github.com/pytorch/pytorch/issues/12…