-
### 🐛 Describe the bug
```python
import torch
@torch.compile(backend="eager")
def f(x):
if x.numel() >= 1024:
x += 5
return x
else:
x = x * 2
return…
-
### 🐛 Describe the bug
By testing with the following code, `angle_pi` and `angle_pi_jit` output different results
````python
import torch
angle = torch.zeros(10,device='cpu').uniform_(-7, -3)
…
-
The behavior of the CutMix transform on labels was changed significantly from v1.3.2 to v1.4.0rc1. This change originates from #7813, in which I have left a comment, but I'm creating an issue to incre…
-
### Your current environment
```
Collecting environment information...
INFO 05-23 16:19:36 pynccl.py:58] Loading nccl from library librccl.so.1
/opt/conda/envs/py_3.9/lib/python3.9/site-packages/t…
-
I've just installed CUDA 12.6 on a fresh system. Turns out, R Torch only supports 11.7/.8 for now. Are there any near-term plans to extend support to newer CUDA versions, or should I downgrade CUDA …
-
### 🐛 Describe the bug
**When trying to compile a simple function that uses `cpu` tensors, torch inductor initialises a context on `cuda:0`**.
If used in a multiprocessing context (E.G.) a torch d…
-
### System Info
- `transformers` version: 4.41.0
- Platform: Linux-5.4.0-162-generic-x86_64-with-glibc2.35
- Python version: 3.10.12
- Huggingface_hub version: 0.23.4
- Safetensors version: 0.4…
-
## 🐛 Bug
#### When concentration parameters a, b are equal and go to 0, sample values are wrong.
## To Reproduce
Steps to reproduce the behavior:
copy paste the following in a notebook:
…
-
When both use cpu in HalfCheetah-v2, the pytorch occupy about 1000% cpu and tensorflow only 300% cpu. But pytorch is much slower. When running 200000 timesteps, the pytorch costs about 1200s but tenso…
-
### Your current environment
PyTorch version: 2.3.1+cu121
Is debug build: False
CUDA used to build PyTorch: 12.1
### How would you like to use vllm
I used this script to start the server:
vllm s…