[WIP] Add e2e test for `tune` api with LLM hyperparameter optimization #1084

Job	Run time
e2e (v1.27.11, long-running-resume,from-volume-resume,median-stop)	17m 10s
e2e (v1.27.11, grid,bayesian-optimization,tpe,multivariate-tpe,cma-es,hyperband)	21m 25s
e2e (v1.27.11, file-metrics-collector,pytorchjob-mnist)	13m 31s
e2e (v1.27.11, median-stop-with-json-format,file-metrics-collector-with-json-format)	11m 54s
e2e (v1.28.7, long-running-resume,from-volume-resume,median-stop)	16m 41s
e2e (v1.28.7, grid,bayesian-optimization,tpe,multivariate-tpe,cma-es,hyperband)	25m 1s
e2e (v1.28.7, file-metrics-collector,pytorchjob-mnist)	15m 12s
e2e (v1.28.7, median-stop-with-json-format,file-metrics-collector-with-json-format)	14m 15s
e2e (v1.29.2, long-running-resume,from-volume-resume,median-stop)	16m 6s
e2e (v1.29.2, grid,bayesian-optimization,tpe,multivariate-tpe,cma-es,hyperband)	28m 33s
e2e (v1.29.2, file-metrics-collector,pytorchjob-mnist)	16m 29s
e2e (v1.29.2, median-stop-with-json-format,file-metrics-collector-with-json-format)	12m 50s
	3h 29m 7s

Provide feedback