How to setup own environment for Building Real-Time Video AI Applications course

AI & Data Science Computer Vision & Image Processing
1

Hello ,

I have enjoyed Building Real-Time Video AI Applications course and would like to replicate these labs to my own dev environment. I’m considering using runpod.ai GPU environments. How can I go about this.

Which Docker container can I import to runpod?. Are there any GitHub references to these labs and datasets?. Which GPUs would be suitable to replicate these lab or similar ? Which GPU firmware or cuda version would I need to consider? Which other vcpu, ram, vram and storage would be suitable?
1 Like
3

Hi @duncanngacha.gitau,

Pleased you enjoyed the course!

This course uses TAO Toolkit and DeepStream - the environment uses the TAO 5.0 container (nvcr.io/nvidia/tao/tao-toolkit:5.0.0-tf1.15.5) and installs DeepStream (deepstream-6.3_6.3.0-1_amd64.deb) on top of it. that’s what we would recommend for recreating the environment.

Best,

Sophie

4

Hello Sophie,

Thanks for the feedback.

Is there a git hub reference to clone the dataset or do I download and replicate manuall from the from the lab?

Regards,
Duncan

6

Hi Duncan,

There isn’t a public source to download from. You can download from our lab environment. Please keep in mind that our service agreement states that the dataset can only be used for academic purposes.

Best,

Sophie