[WIP] Add e2e test for `tune` api with LLM hyperparameter optimization #1090

Job	Run time
e2e (v1.27.11, long-running-resume,from-volume-resume,median-stop)	15m 34s
e2e (v1.27.11, grid,bayesian-optimization,tpe,multivariate-tpe,cma-es,hyperband)	23m 8s
e2e (v1.27.11, file-metrics-collector,pytorchjob-mnist)	14m 21s
e2e (v1.27.11, median-stop-with-json-format,file-metrics-collector-with-json-format)	15m 40s
e2e (v1.28.7, long-running-resume,from-volume-resume,median-stop)	16m 19s
e2e (v1.28.7, grid,bayesian-optimization,tpe,multivariate-tpe,cma-es,hyperband)	26m 13s
e2e (v1.28.7, file-metrics-collector,pytorchjob-mnist)	18m 26s
e2e (v1.28.7, median-stop-with-json-format,file-metrics-collector-with-json-format)	16m 53s
e2e (v1.29.2, long-running-resume,from-volume-resume,median-stop)	14m 39s
e2e (v1.29.2, grid,bayesian-optimization,tpe,multivariate-tpe,cma-es,hyperband)	24m 18s
e2e (v1.29.2, file-metrics-collector,pytorchjob-mnist)	12m 31s
e2e (v1.29.2, median-stop-with-json-format,file-metrics-collector-with-json-format)	18m 8s
	3h 36m 10s

Provide feedback