-
_Downstream PyTorch issue:_
https://github.com/pytorch/pytorch/issues/133780
I'm trying to do attention on a batch-of-zero, because my program uses a static graph and I rely on zero-batching (in…
-
### Your current environment
The output of `python collect_env.py`
```text
Collecting environment information...
PyTorch version: 2.4.0+cu121
Is debug build: False
CUDA used to build PyTor…
-
### 🐛 Describe the bug
I found the scaled_dot_product_attention() can't implemented the backwark() . I
RuntimeError: derivative for aten::_scaled_dot_product_flash_attention_backward is not impleme…
-
I am very pleased to see your interesting paper and hope it will be accepted by a top conference soon! I would like to study it in depth but have encountered the following problems. I apologize for t…
-
### Describe the issue
Hi IPEX team,
I have an application where I want to serve multiple models concurrently, and I want to share weights across concurrent instances. I normally do this with `tor…
-
### Export Microsoft/Phi-3-small-8k-instruct ONNX model on CPU (Ubuntu 22.04.4 LTS)
As per [suggestion](https://github.com/microsoft/onnxruntime-genai/pull/710#issue-2415518051), I referred to [ONN…
-
Hello, this is an excellent job. When I was reading your paper, I had a question: You do graph convolution network on the feature space and it feels similar to channel attention. Is there any differen…
-
Dear developers,
I hope this message finds you well. Firstly, I would like to express my appreciation for your excellent work on the Soot-FlowDroid module. It has been instrumental in my recent ana…
-
您好!我对在使用社交网络和兴趣网络更新用户表示过程中,注意力分数的计算有些疑问。
首先,
从以上代码可以看出 gama^(k+1)_(a1) =1/2* self.consumed_items_attention,gama^(k+1)_(a2) =1/2* self.social_neighbors_attention。gama^(k+1)_(a1)和gama^(k+1)_(a2)…
-
Can you share more details on the technique for repo level concatenation part?