Trying to run opencv DNN demos on imx8mq-evk

[vsuneja63]

Following eIQ machine learning software development environment for imx application processor. Working on OpenCV object detection in imx8. Attaching the same document.

I am trying to run opencv DNN demos on imx8mq-evk target. Running "python download_models.py" to downloads the NN models, configuration files, and input images for some OpenCV examples. Looks like download_models.py have so many models, i am running out of space. Document doesn't mention anything about how much rootfs space required to download it. It consumed 7 GB in imx8 & still lot more models are pending. Can you please clarify space requirement? As space is limited in imx board, how feasible it is to download all the models. Can we limit our models download as per requirement rather than just downloading everything? 

[Bio_TICFSL]

Hello Vipul,

The model has 1GB, however if you are running out of space, we recommend to burn you sdcard image with a 16GB card, otherwise don´t load the models.

Regards

[vsuneja63]

Hi,

Thanks!

Would like to hear from you on this. I am running dnn classification demo on imx8 target but didn't get output. What could be the reason or am i missing some dependent package?

Thanks & Regards,

Vipul Kumar

[vsuneja63]

Hi,

Thanks!

I used 32GB sd card & download all models, it took approximately 14 GB of space.

I am running image classification example through command given below. Its neither poping up output image nor saving it. What can I do to save the output image?  

./example_dnn_classification --input=dog416.png --zoo=models.yml squeezenet

It's not throwing any error.

Thanks & Regards,

Vipul Kumar