Debugging issue: triton-inference-server/server#5951
This repository attempts to debug issues with using the decoupled exec in a python bls with a triton server (ver. 23.04). The behavior is that the res.send hangs after a certain number of requests.
Update: Jul 6
Unable to reproduce the error, moved to triton version 23.06 and haven't experienced the error 🤔