min_tokens_to_keep has to be a positive integer, but it's type is float in mlx_lm/generate.py

[asamiKA]

In mlx_lm/generate.py, the type of --min-tokens-to-keep parameter is set to float.

mlx-examples/llms/mlx_lm/generate.py

Line 80 in 0ca162c

type=float,

    parser.add_argument(
        "--min-tokens-to-keep",
        type=float,  # <- int
        default=DEFAULT_MIN_TOKENS_TO_KEEP,
        help="Minimum tokens to keep for min-p sampling.",
    )

but in mlx_lm/sample_utils.py, this cause error that 'min_tokens_to_keep has to be a positive integer'.

mlx-examples/llms/mlx_lm/sample_utils.py

Line 109 in 0ca162c

if not isinstance(min_tokens_to_keep, int) or (min_tokens_to_keep < 1):

    if not (0 <= min_p <= 1.0):
        raise ValueError(
            f"`min_p` has to be a float in the [0, 1] interval, but is {min_p}"
        )
    if not isinstance(min_tokens_to_keep, int) or (min_tokens_to_keep < 1):
        raise ValueError(
            f"`min_tokens_to_keep` has to be a positive integer, but is {min_tokens_to_keep}"
        )

[awni]

Good catch, that's a typo. Would you like to send a PR to fix it?

[awni]

Actually I just added the fix in #1133