Question

以下是我的代码的一部分：

if __name__ == "__main__":

    # Inputs for the main function
    parser = argparse.ArgumentParser(
        description="A script that emulates the competition's scoring process, "
        "the hider (from hider.py) is run against the seeker (from seeker.py)."
    )
    parser.add_argument(
        "-d",
        "--data_path",
        metavar="PATH",
        default="./data/amsterdam_subset_1A.csv",  # !!! I changed this to 1% data file, originally is full data
        type=str,
        help="Data file path (Amsterdam dataset). Defaults to './data/train_longitudinal_data.csv'.",
    )
    parser.add_argument(
        "-o",
        "--output_dir",
        metavar="PATH",
        default="./output",
        type=str,
        help="Output directory. Defaults to './output'.",
    )
    parser.add_argument(
        "-m", "--max_seq_len", metavar="INT", default=100, type=int, help="Max sequence length limit. Defaults to 100."
    )
    parser.add_argument(
        "-t", "--train_frac", default=0.5, metavar="FLOAT", type=float, help="Training set fraction. Defaults to 0.5."
    )
    parser.add_argument(
        "-f",
        "--feature_prediction_no",
        metavar="INT",
        default= int(5),
        type=int,
        help="Number of features in the subset of features used to run feature prediction "
        "(part of hider evaluation). Defaults to 5.",
    )
    parser.add_argument("-s", "--seed", metavar="INT", default=0, type=int, help="Random seed. Defaults to 0.")
    parser.add_argument(
        "-g",
        "--debug_data",
        metavar="INT",
        default=0,
        type=int,
        help="Set this to a non-0 value to use a 'debug' subset of the dataset instead of the whole dataset "
        "(useful for speedy debugging), only the first --debug_data many rows of the data file will be loaded. "
        "Defaults to 0.",
    )
    parser.add_argument(
        "--skip_fp", action="store_true", default=False, help="Skip feature prediction step of hider evaluation if set."
    )
    parser.add_argument(
        "--skip_osa",
        action="store_true",
        default=False,
        help="Skip one-step-ahead prediction step of hider evaluation if set.",
    )
    parser.add_argument(
        "--eval_verbose",
        action="store_true",
        default=False,
        help="If set, the underlying training in hider evaluation stages will be shown verbosely "
        "(training epoch etc.).",
    )
    parser.add_argument(
        "--timegan_opt",
        choices = ['Adam', 'dpgrad', 'noisyadam'],
        default = 'Adam',
        type = str)
    parser.add_argument(
        "--algorithm",
        choices = ["add_noise", "timegan"],
        default = "timegan",
        type = str)
    
    parsed_args, unknown = parser.parse_known_args()
    #parsed_args = parser.parse_args()

    # Call main function
    main(parsed_args)

我的问题是当我尝试在jupyter笔记本中运行它时，它给了我这个错误：

Traceback (most recent call last):
  File "C:\Users\zyr13\anaconda3\lib\argparse.py", line 2409, in _get_value
    result = type_func(arg_string)
ValueError: invalid literal for int() with base 10: 'C:\\Users\\zyr13\\AppData\\Roaming\\jupyter\\runtime\\kernel-2789c574-de13-44e2-aadc-3d7a8a319564.json'

似乎有一些奇怪的目录传递给参数“ feature_prediction_no”。如您所见，默认值为5，所以我希望应该传递5，但是结果清楚地表明这不会发生。我一直在寻找this线程，并且思考可能是类似的问题，但是它没有给出足够清晰的说明。我正在为我的特殊情况寻找答案，并提供一些解释，发生了什么事？

jupyter笔记本中使用parse_args的问题

0 个答案: