[GSoC] Update `tune` API for LLM hyperparameters optimization #825

Job	Run time
e2e (v1.27.11, long-running-resume,from-volume-resume,median-stop)	14m 28s
e2e (v1.27.11, grid,bayesian-optimization,tpe,multivariate-tpe,cma-es,hyperband)	18m 10s
e2e (v1.27.11, file-metrics-collector,pytorchjob-mnist)	13m 34s
e2e (v1.27.11, median-stop-with-json-format,file-metrics-collector-with-json-format)	13m 30s
e2e (v1.28.7, long-running-resume,from-volume-resume,median-stop)	14m 32s
e2e (v1.28.7, grid,bayesian-optimization,tpe,multivariate-tpe,cma-es,hyperband)	20m 32s
e2e (v1.28.7, file-metrics-collector,pytorchjob-mnist)	13m 21s
e2e (v1.28.7, median-stop-with-json-format,file-metrics-collector-with-json-format)	13m 45s
e2e (v1.29.2, long-running-resume,from-volume-resume,median-stop)	16m 14s
e2e (v1.29.2, grid,bayesian-optimization,tpe,multivariate-tpe,cma-es,hyperband)	21m 32s
e2e (v1.29.2, file-metrics-collector,pytorchjob-mnist)	14m 4s
e2e (v1.29.2, median-stop-with-json-format,file-metrics-collector-with-json-format)	13m 11s
	3h 6m 53s

Provide feedback