-
### 🐛 Describe the bug
For the `mobilenet_v3_large` model from the torch-vision repository, the fp32 accuracy achieves `Acc@1 74.054 and Acc@5 91.340` on the ImageNet validation dataset. Through PT2 …
-
I have implemented task assignment internally for FB.
Related issue: #1309
Implementation:
* Optional setting to enable task assignment
* The map shows icons to show which tasks are assign…
-
### 🐛 Describe the bug
when I set the `dropout_p=0.0`, the result is different. But `dropout_p=-1`, the result is same. Maybe the op scaled_dot_product_attention has some bug. Please fix it, thank…
-
### 🐛 Describe the bug
Hi, here is a reproduction case which encounters severe performance issue:
```
import time
import torch
dtype = torch.bfloat16
sizes = (64, 128, 513, 513)
t = tor…
-
## Issue description
I've been observing an issue with the `torch.cuda.memory_reserved("cuda:0")` function in PyTorch. Despite having a model training on the GPU, the function `torch.cuda.memory_…
-
**What am I doing**
I am generating texture mipmaps using a simple compute shader. The texture loading process is asynchronous and happens in a separate thread, which loads the texture from file, c…
-
# Bug Report
**YDB Python SDK version:**
ydb 3.8.1
**Environment**
Ubuntu 22.04.4 LTS
Python 3.12.2
processor : 0
vendor_id : GenuineIntel
cpu family : 6
model :…
-
This issue seems to occur when the wheel was built on another system.
Steps to reproduce:
- Arch Linux
- CPU: i7 2600
- Conda env with Python 3.8 or 3.10
- `pip install apsi==0.1.2`
- run exam…
-
### 🐛 Describe the bug
`torch.atan(inf)` returns NaN when dtype is complex128, but pi/2 on other dtypes such as float64. Please check the code below.
```python
input = torch.tensor([np.inf], dtype=…
dlibk updated
5 months ago
-
### Your current environment
```text
The output of `python collect_env.py`
```
### 🐛 Describe the bug
1.flashinfer + chunked prefill outputs garbage
2.missing support of copy / swap blocks wit…