scripts/launcher_distributed.py [324:373]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
    )
    parser.add_argument("--use_downloaded_model", type=bool, default=False)

    args = parser.parse_known_args()

    return args


def print_env_vars():

    print("***** Printing enviroment variables *****")
    print(f"Master Addr: {MASTER_ADDR}")
    print(f"Mater Port: {MASTER_PORT}")
    print(f"Total number of GPUs (WORLD SIZE): {WORLD_SIZE}")
    print(f"The (global) rank of the current process: {GLOBAL_RANK}")
    print(f"Local node rank: {LOCAL_RANK}")
    print(f"Number of GPUs per node: {NUM_GPUS_PER_NODE}")
    print(f"Number of nodes: {NUM_NODES}")
    print(f"Use Downloaded Model: {args.use_downloaded_model}")
    print(f"Type of use_downloaded_model: {type(args.use_downloaded_model)}")
    print(f"Action: {args.tune_action}")
    check_pytorch_version()


def completion_status():
    print("***** Finished Task *****")

    list_model_dir = f"ls -ltr {args.model_dir}"
    run_command(list_model_dir)

    list_quantized_model_dir = f"ls -ltr {args.model_dir}/quantized"
    run_command(list_quantized_model_dir)


def training_function():

    print_env_vars()

    # Step 1: Map values to functions
    function_map = {
        "fine-tune": finetune_model,
        "run-eval": run_eval,
        "run-quant": run_quant,
    }

    # Step 2: Iterate through the array and call the corresponding functions
    for value in args.tune_action.split(","):
        if value in function_map:
            print(f"function_key: {value}")
            try:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -



scripts/launcher_distributed_kd.py [329:378]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
    )
    parser.add_argument("--use_downloaded_model", type=bool, default=False)

    args = parser.parse_known_args()

    return args


def print_env_vars():

    print("***** Printing enviroment variables *****")
    print(f"Master Addr: {MASTER_ADDR}")
    print(f"Mater Port: {MASTER_PORT}")
    print(f"Total number of GPUs (WORLD SIZE): {WORLD_SIZE}")
    print(f"The (global) rank of the current process: {GLOBAL_RANK}")
    print(f"Local node rank: {LOCAL_RANK}")
    print(f"Number of GPUs per node: {NUM_GPUS_PER_NODE}")
    print(f"Number of nodes: {NUM_NODES}")
    print(f"Use Downloaded Model: {args.use_downloaded_model}")
    print(f"Type of use_downloaded_model: {type(args.use_downloaded_model)}")
    print(f"Action: {args.tune_action}")
    check_pytorch_version()


def completion_status():
    print("***** Finished Task *****")

    list_model_dir = f"ls -ltr {args.model_dir}"
    run_command(list_model_dir)

    list_quantized_model_dir = f"ls -ltr {args.model_dir}/quantized"
    run_command(list_quantized_model_dir)


def training_function():

    print_env_vars()

    # Step 1: Map values to functions
    function_map = {
        "fine-tune": finetune_model,
        "run-eval": run_eval,
        "run-quant": run_quant,
    }

    # Step 2: Iterate through the array and call the corresponding functions
    for value in args.tune_action.split(","):
        if value in function_map:
            print(f"function_key: {value}")
            try:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -