This diagram illustrates the workflow of a
A user submits a
ProflingExperimentvia a RPC or front-end UI interface, specifying the ML model, tunable configuration parameters, optimization objectives, and sampling budgets.
Within the sampling budget, Morphling iteratively communicates with the algorithm server to get the next configuration for sampling.
Then Morphling starts a
Trialto evaluate that sampling.
Trial, a model serving inference instance
Deploymentis launched, and its “readiness” is reported to trigger a client-side RPS stress-test
After the client
Jobcompletes, the measured peak RPS is stored in the
Trialfinishes, and the result is sent to the
ProflingExperimentcompletes when the sampling budget is reached.
The sequence diagram of the ProflingExperiment workfolow is shown as follows: