kneron / host_lib Goto Github PK

Hello,
We have the KL520 board. For starter to check the board's working we followed steps as explained in this screenshot and were successful in performing the face recognition via putty.

However when we wanted to send the YOLO model or run few examples provided in the documentation, we are running into issues. The error screenshot is as follows.

Our main concern is to understand if there are any specific steps to be followed to send the model from computer to KL520 board via putty? If there is any tutorial to do so?
Can you please advise us on this.
Thanks.
Haritha

My environment:

Intel i7H 10th
python 3.8.5

Question:

The function return None type error. when use command "python main.py -t KL520-cam_isi_async_yolo"

Hello, I'm having issues running the example code on my Raspberry PI 4. I currently use Ubuntu 20 and the processor is ARM64. Do you have an appropriate whl file for ARM64 Linux? I believe the one included already is meant for Intel x86-64. Thanks!

Described in Kneron Computer Library P.32
missing file python voc2yolo.py -c coco_classes.txt
could not find /host_lib/misc

Hi there,

 [Situation]
 OS: windows 10
 Libraries: pytorch and keras

 I've trained models by myself. The models are about OCR, which is capable to detect positions of texts and recognizer all texts. It's a python script. Is there any way that I could know the model is accelerated by AI dongle. How could I call AI dongle to improve computation of model.

 Does any step that I should check and follow ? 

Best regards,

Hi,
I've installed all the packages mentioned in README.md and plugged in the dongle, but when I run python main.py -t update_app, I've encountered the error as follow.

How can I solve the problem?
Thanks.
Kaiwen

https://github.com/kneron/host_lib#run-examples-2

Run Examples
Here are the steps you need to run the example. Let's take KL720-cam_isi_pre_post_host_fcos as the example.

Enter python directory.
Run python main.py -t KL720-cam_isi_pre_post_host_fcos.

========================================================

It has issues.......
I find that's in "/host_lib/python/common/pre_post_process/kneron_pre".

I have a question about output number , when I use the public_yolov3_416 in the py , it can get correct c,w,h and output number,
but when I change it to my nef file , the output number always be 0 and channel number was wrong .

My nef file was running correctly in emulator which you provide , the compile file was here