Comments (7)
@Yuxin-Yu we have done limited testing, but yes, we can confirm that the DPUCZ IP can be implemented on a Zynq-7000 device with a minor patch to the Vitis AI Library.
vitis-ai-library.zip
from vitis-ai.
Thank you @quentonh . In addition, in order to make dpuczdx8g v4.1 compatible with zynq 7000, I have changed some related parameters, as shown in the red circles in the following two figures. Is that right?
from vitis-ai.
Also, it should be noted that I am using Vitis AI 3.0 version. Are the patch files used the same? @quentonh
from vitis-ai.
@Yuxin-Yu we have not internally verified the patches against Vitis AI 3.0, but the modifications to the Vitis AI Library will be very similar if not identical. The required changes result from differences in the NEON instruction set between the Cortex A53 (ZU+) and Cortex A9 (Zynq-7000). You should do compare between the modified source files in the two releases (3.5 and 3.0) to ensure that the changes apply.
With regard to integration into a hardware platform, I can send you an example. Please drop me an email to quentonh at amd.com.
from vitis-ai.
Hi @quentonh .I don't know how to send email at amd.com.If possible, you can send the example to my email : [email protected] .Thank you.
from vitis-ai.
Hello @quentonh , I successfully ran the resnet50 inference process using zynq7000, but the inference result is incorrect, as shown below:
root@debian:~/Vitis-AI/src/vai_runtime/vart_debug/softmax-runner-cpu/test/resnet_v1_50_softmax_cpu# ./run.sh ~/app/img/bellpeppe-994958.JPEG
score[971] = 0.301356 text: bubble,
score[616] = 0.301356 text: knot,
score[506] = 0.301356 text: coil, spiral, volute, whorl, helix,
score[488] = 0.0247368 text: chain,
score[679] = 0.0150036 text: necklace,
What could be the reason and how should I debug it?
from vitis-ai.
Problem solved, it's because my xmodel file was used incorrectly
from vitis-ai.
Related Issues (20)
- Resource Management and Concurrent Model Execution with Vitis AI on KR260 Board
- Multi-input network model, quantization error occurred
- In the UG1414, the operators supported by PyTorch are not available in the process of training and quantification
- AssertionError
- TypeError: list indices must be integers or slices, not NoneType HOT 1
- AddScalar() does not work for quantization
- Use sigmoid or pow to replace exp
- command 'xdputil' not found
- Duplicate nodes error while converting model from quantized_model.pb to xmodel
- YOLOv5 Quantization
- YOLOv5 quantization HOT 1
- linux hardening: checksecurity 2.0.15 -fail to fetch from: http://ftp.de.debian.org/debian/pool/main/c/checksecurity/checksecurity_2.0.15.tar.gz unable to fetch and compile checksecurity into Petalinux for basic system security checks. HOT 1
- Does AWS F1 support Vitis AI 2.5?
- Resnet-50 test accuracy and loss of accuracy in DPU HOT 6
- Undefined Reference to Symbol HOT 1
- When release 4.0 version? HOT 11
- Cant build: ./docker_build.sh -t gpu -f tf2 HOT 1
- Vck 190 setup problems. HOT 4
- Partial Quantization
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from vitis-ai.