[GSoC] Add e2e test for `tune` api with LLM hyperparameter optimization #1258

Job	Run time
e2e (v1.29.2, long-running-resume,from-volume-resume,median-stop)	16m 41s
e2e (v1.29.2, grid,bayesian-optimization,tpe,multivariate-tpe,cma-es,hyperband)	20m 14s
e2e (v1.29.2, file-metrics-collector,pytorchjob-mnist)	13m 32s
e2e (v1.29.2, median-stop-with-json-format,file-metrics-collector-with-json-format)	12m 44s
e2e (v1.30.7, long-running-resume,from-volume-resume,median-stop)	15m 6s
e2e (v1.30.7, grid,bayesian-optimization,tpe,multivariate-tpe,cma-es,hyperband)	22m 26s
e2e (v1.30.7, file-metrics-collector,pytorchjob-mnist)	12m 8s
e2e (v1.30.7, median-stop-with-json-format,file-metrics-collector-with-json-format)	12m 47s
e2e (v1.31.3, long-running-resume,from-volume-resume,median-stop)	14m 19s
e2e (v1.31.3, grid,bayesian-optimization,tpe,multivariate-tpe,cma-es,hyperband)	20m 12s
e2e (v1.31.3, file-metrics-collector,pytorchjob-mnist)	14m 26s
e2e (v1.31.3, median-stop-with-json-format,file-metrics-collector-with-json-format)	12m 20s
	3h 6m 55s

Provide feedback