-
Updated
Dec 4, 2020 - C++
hardware-acceleration
Here are 90 public repositories matching this topic...
-
Updated
Mar 30, 2020 - JavaScript
-
Updated
Dec 7, 2020 - C
-
Updated
Aug 27, 2019 - C
-
Updated
Jul 3, 2018 - C++
-
Updated
Oct 14, 2020 - C
-
Updated
Nov 22, 2020 - Java
-
Updated
Dec 6, 2020 - C++
-
Updated
Nov 25, 2020 - Python
-
Updated
Aug 7, 2020 - Shell
-
Updated
Dec 1, 2020 - C
-
Updated
Dec 7, 2020 - Python
-
Updated
Jan 5, 2019 - VHDL
-
Updated
Dec 10, 2019 - Python
-
Updated
Jul 30, 2020 - TypeScript
-
Updated
Dec 7, 2020 - JavaScript
-
Updated
Oct 29, 2020 - C++
-
Updated
Nov 20, 2020 - Verilog
-
Updated
Oct 5, 2017
-
Updated
Aug 22, 2020 - Python
-
Updated
Jul 11, 2018 - Jupyter Notebook
-
Updated
Aug 23, 2020 - C++
-
Updated
Dec 3, 2020 - C
-
Updated
May 2, 2017 - C
-
Updated
May 4, 2020 - Python
Right now it is very easy to deadlock the whole system if a PE e.g. reads from the DDR and writes to the DDR but never deals with the result of either request. Right now such a situation is quite hard to debug. Most often this scenario happens when the DMA engine is active at the same time.
Right now I have to possible solutions:
- Add infrastructure to the AXI ports of a PE that counts the n
-
Updated
May 4, 2020 - VHDL
-
Updated
Dec 7, 2020 - JavaScript
Improve this page
Add a description, image, and links to the hardware-acceleration topic page so that developers can more easily learn about it.
Add this topic to your repo
To associate your repository with the hardware-acceleration topic, visit your repo's landing page and select "manage topics."
Describe the bug
when axis has duplicate value , onnxruntime compute result is all same value ,which is different with expect of tensorflow
Urgency
2020.11.18
System information
Linux Ubuntu 16.04
Expected behavior
When there are duplicate values, the duplicate can be removed. j