want to deploy pretrained model on raspberry pi 4 for real time testing

[vs1612]

General Discussion

I also want to run my pre-trained custom model on raspberry pi. I want the low power solution. As per my understanding we can directly run our model on raspberry pi just by copying the model from main machine to raspberry pi then why we need to use docker

[xuhdev]

The same reason that we do so on a laptop. The isolated environment gives ease to maintain and makes it more reliable to run the model.

[vs1612]

I have created one small python script and i am trying to run that on raspberry pi using docker. My image size is around 700Mb so how that is handled on raspberry pi. And for Model training and testing, i have created one virtual environment in conda in my local machine. so that is handled when i have to run using docker.

…

On Wed, Jul 14, 2021 at 7:22 AM Hong Xu ***@***.***> wrote: The same reason that we do so on a laptop. The isolated environment gives ease to maintain and makes it more reliable to run the model. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#26 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AUWVGCTPOGVAV5RNA5G5CGTTXUUJPANCNFSM5AKN5GSA> .