server
server copied to clipboard
Does triton inference server support customers custom feature but do not need to modify the origin code, like some plugin feature?
Is your feature request related to a problem? Please describe. I have some special feature want to apply in triton and I don't want to modify the origin code directly, it will cause that I need maintain a local triton version, and need to adapt incoming new feature everytime
Describe the solution you'd like Add a plugin feature, we can add some preprocess and postprocess about the triton server run
Describe alternatives you've considered the worst option is that we may add a new gateway service in front of the triton
Additional context No other context