-
## 🚀 Feature
Provide a `pin_memory_` method on tensors (note the trailing underscore) which operates in-place.
## Motivation
Pinning memory using the current `pin_memory` method creates a copy of…
-
### Description
When running examples on my MAC notebook, it got the correct path of the native library but failed to load it.
Here's the output of `uname -a` of my notebook.
```
Darwin U-0R7T…
-
### Describe the issue
Several operators that should support bfloat16 do not do so with the CUDA execution provider. This was noticed with `ReduceMean`, but here is the complete list:
`Abs`, `ArgM…
-
When I run `cmake ../crsc`
Heres the output ;
```
CMake Error at /home/zen/anaconda3/envs/tf/share/cmake-3.26/Modules/CMakeDetermineCompilerId.cmake:751 (message):
Compiling the CUDA compil…
-
**Your question**
How can I profile bubble time in pipeline parallelism?
-
The Python SWIG client exposes the setDefaultStream() function on the GPU resources object but it does not seem to provide any type conversion options to pass in a Python equivalent to the C++ cudastr…
-
### Describe the issue
We exported the Huggingface transformer model [OneFormer](https://huggingface.co/docs/transformers/model_doc/oneformer) into onnx.
Opset 20 failed with the error:
```
O…
tikr7 updated
5 months ago
-
Tried tensorflow and torch with tinygrad
still getting this error with llama 8b 3.1 and llama 8b as well.
Apparently this is an opencl compile error for bfloat16 data type
Sorry, I am not a kernel …
-
### Your current environment
```text
Collecting environment information...
PyTorch version: 2.4.0
Is debug build: False
CUDA used to build PyTorch: 12.1
ROCM used to build PyTorch: N/A
OS: Ub…
htlou updated
1 month ago
-
### 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…