Need help deploying a workflow

🤝 Community Help
1

I have created a workflow for a video and would like to be able to run it on a live video where it saves the outputs to a file how can I do this?

2

Hi @Sebastien_Berrang !

Are you processing your live stream on your self-hosted solution (i.e. jetson) or are you processing video with use of Roboflow cloud?

Thanks!

Grzegorz

3

I don’t know whichever easier and can be done on a free account.

4

If you are looking to process within your free account you can simply pip install inference-gpu(if you have CUDA-capable GPU, otherwise inference) and start the pipeline with workflow and you can handle saving results of workflow in sink

There is also slightly more advanced route if you are looking to have a standalone service, you can run it via our latest Docker image, if you have CUDA-capable GPU:

  • docker run roboflow/roboflow-inference-server-gpu:latest

otherwise use CPU:

  • docker run roboflow/roboflow-inference-server-cpu:latest

and interface with it using inference-sdk

If you don’t have docker, you can install and run inference natively:

Hope this helps!

5

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.