[WIP] Add e2e test for `tune` api with LLM hyperparameter optimization #1096

Job	Run time
e2e (v1.27.11, long-running-resume,from-volume-resume,median-stop)	27m 12s
e2e (v1.27.11, grid,bayesian-optimization,tpe,multivariate-tpe,cma-es,hyperband)	15m 1s
e2e (v1.27.11, file-metrics-collector,pytorchjob-mnist)	14m 7s
e2e (v1.27.11, median-stop-with-json-format,file-metrics-collector-with-json-format)	14m 20s
e2e (v1.28.7, long-running-resume,from-volume-resume,median-stop)	14m 25s
e2e (v1.28.7, grid,bayesian-optimization,tpe,multivariate-tpe,cma-es,hyperband)	13m 2s
e2e (v1.28.7, file-metrics-collector,pytorchjob-mnist)	11m 47s
e2e (v1.28.7, median-stop-with-json-format,file-metrics-collector-with-json-format)	14m 25s
e2e (v1.29.2, long-running-resume,from-volume-resume,median-stop)	15m 48s
e2e (v1.29.2, grid,bayesian-optimization,tpe,multivariate-tpe,cma-es,hyperband)	31m 17s
e2e (v1.29.2, file-metrics-collector,pytorchjob-mnist)	14m 16s
e2e (v1.29.2, median-stop-with-json-format,file-metrics-collector-with-json-format)	14m 42s
	3h 20m 22s

Provide feedback