-
### 🐛 Describe the bug
When running distributed (or multi-process) jobs on AMD GPU systems with GPU isolation (i.e., the `*_VISIBLE_DEVICES` environment variable), PyTorch still tries to set the se…
-
When trying to write code in a `.cu` file the Intellisense only partially works, for example if I try compile this code as a `.cpp` file then every header is correctly recognized and quick fixes work
…
-
### 🐛 Describe the bug
If I follow the minimal C++ example (https://pytorch.org/cppdocs/installing.html#minimal-example), but use CUDA stream or graph functionality through ATen, the build fails:
…
-
### 🐛 Describe the bug
For some inputs, torch.remainder and torch.fmod produce wrong results, especially for integer datatype. When converting the int32 input to float32, they can produce correct res…
-
### Your current environment
The output of `python collect_env.py`
```text
PyTorch version: 2.4.0+cu121
Is debug build: False
CUDA used to build PyTorch: 12.1
ROCM used to build PyTorch: N…
-
### 🐛 Describe the bug
Description:
I encountered an issue while using the ExecuTorch runtime library, which was built from source with the following parameters:
- **ANDROID_NDK**: 26.2.1139434…
-
### Describe the issue
inference speed is very slow when using fp16 while using fp 32 is normal
### To reproduce
inference speed is very slow when using fp16 while using fp 32 is normal
### Urge…
-
## 描述问题/Describe the bug
我把训练好的模型部署到一台win10系统的机器中，执行eval或者inference推断时都会报错：
cudaErrorNoKernelImageDevice: no kernel image is available for execution on the device.
个人感觉是cuda driver 版本、cudatoo…
-
### Checklist
- [X] 1. I have searched related issues but cannot get the expected help.
- [X] 2. The bug has not been fixed in the latest version.
- [X] 3. Please note that if the bug-related iss…
-
when executing `sudo ./build.sh` i get this output
pytorch 1.4.0
cuda-10.2.89-3
cudnn-7.6.5.32-3
on 5.4.24-1-MANJARO
GPU: NVIDIA GeForce GTX 1070
```
Need pytorch>=1.0.0
./build.sh: line 4: ac…