Is it possible to send OpenCV frames to runpod?
Hi,
I am trying to send frames to runpod for inference. I am currently using serverless endpoints (but open to warm or 24/7 containers as well!). Basically, in opencv, you would get the frames within the video loop. I will be sending those frames to runpod for inference.
I am wondering if this is possible. In my test.json, I have the example of the image path (the full b64 file). I tried initializing the serverless pods with two image_paths: one, an example b64 one (made up), and the second, the full b64 image path. Both failed.
My goal is to send frames in real time to runpod.
---
In python, this is what would normally happen:
I am trying to replace face_rec with:
---
basically wondering if it's possible to send OpenCV frames (as image paths) to runpod, get the AI inference, and then receive it in my application.
I am trying to send frames to runpod for inference. I am currently using serverless endpoints (but open to warm or 24/7 containers as well!). Basically, in opencv, you would get the frames within the video loop. I will be sending those frames to runpod for inference.
I am wondering if this is possible. In my test.json, I have the example of the image path (the full b64 file). I tried initializing the serverless pods with two image_paths: one, an example b64 one (made up), and the second, the full b64 image path. Both failed.
My goal is to send frames in real time to runpod.
---
In python, this is what would normally happen:
I am trying to replace face_rec with:
---
basically wondering if it's possible to send OpenCV frames (as image paths) to runpod, get the AI inference, and then receive it in my application.
