[WIP] Add e2e test for `tune` api with LLM hyperparameter optimization #1069

Job	Run time
e2e (v1.27.11, long-running-resume,from-volume-resume,median-stop)	14m 30s
e2e (v1.27.11, grid,bayesian-optimization,tpe,multivariate-tpe,cma-es,hyperband)	18m 40s
e2e (v1.27.11, file-metrics-collector,pytorchjob-mnist)	13m 14s
e2e (v1.27.11, median-stop-with-json-format,file-metrics-collector-with-json-format)	14m 28s
e2e (v1.28.7, long-running-resume,from-volume-resume,median-stop)	20m 10s
e2e (v1.28.7, grid,bayesian-optimization,tpe,multivariate-tpe,cma-es,hyperband)	20m 0s
e2e (v1.28.7, file-metrics-collector,pytorchjob-mnist)	13m 37s
e2e (v1.28.7, median-stop-with-json-format,file-metrics-collector-with-json-format)	13m 57s
e2e (v1.29.2, long-running-resume,from-volume-resume,median-stop)	15m 14s
e2e (v1.29.2, grid,bayesian-optimization,tpe,multivariate-tpe,cma-es,hyperband)	20m 32s
e2e (v1.29.2, file-metrics-collector,pytorchjob-mnist)	12m 48s
e2e (v1.29.2, median-stop-with-json-format,file-metrics-collector-with-json-format)	12m 23s
	3h 9m 33s

Provide feedback