We have to run 12 instances of FRCNN model in parallel on the GPU. We are trying to determine which graphics card to but for this purpose. We need very fast inferencing (50-100 ms) in our application.
Graphics cards of choice:
a) Nvidia Quadro P5000 16 gb
b) Nvidia Quadro RTX 4000 8gb
c)Nvidia Quadro RTX 5000 16gb
The system which we are using is a workstation.
Please feel free to suggest any other if the above mentioned cards are in sufficient to process 12 FRCNN models in parallel.