huggingface/transformers · error · ValueError

norm must be one of None or "slaney"

Error message

norm must be one of None or "slaney"

What it means

Thrown by `mel_filter_bank` when the `norm` argument is neither None nor the string "slaney". Normalization controls whether triangular mel weights are divided by the mel-band width (Slaney-style area normalization); no other normalization scheme is implemented, so anything else is rejected before the filter bank matrix is built.

Source

Thrown at src/transformers/audio_utils.py:692

            Lowest frequency of interest in Hz.
        max_frequency (`float`):
            Highest frequency of interest in Hz. This should not exceed `sampling_rate / 2`.
        sampling_rate (`int`):
            Sample rate of the audio waveform.
        norm (`str`, *optional*):
            If `"slaney"`, divide the triangular mel weights by the width of the mel band (area normalization).
        mel_scale (`str`, *optional*, defaults to `"htk"`):
            The mel frequency scale to use, `"htk"`, `"kaldi"` or `"slaney"`.
        triangularize_in_mel_space (`bool`, *optional*, defaults to `False`):
            If this option is enabled, the triangular filter is applied in mel space rather than frequency space. This
            should be set to `true` in order to get the same results as `torchaudio` when computing mel filters.

    Returns:
        `np.ndarray` of shape (`num_frequency_bins`, `num_mel_filters`): Triangular filter bank matrix. This is a
        projection matrix to go from a spectrogram to a mel spectrogram.
    """
    if norm is not None and norm != "slaney":
        raise ValueError('norm must be one of None or "slaney"')

    if num_frequency_bins < 2:
        raise ValueError(f"Require num_frequency_bins: {num_frequency_bins} >= 2")

    if min_frequency > max_frequency:
        raise ValueError(f"Require min_frequency: {min_frequency} <= max_frequency: {max_frequency}")

    # center points of the triangular mel filters
    mel_min = hertz_to_mel(min_frequency, mel_scale=mel_scale)
    mel_max = hertz_to_mel(max_frequency, mel_scale=mel_scale)
    mel_freqs = np.linspace(mel_min, mel_max, num_mel_filters + 2)
    filter_freqs = mel_to_hertz(mel_freqs, mel_scale=mel_scale)

    if triangularize_in_mel_space:
        # frequencies of FFT bins in Hz, but filters triangularized in mel space
        fft_bin_width = sampling_rate / ((num_frequency_bins - 1) * 2)
        fft_freqs = hertz_to_mel(fft_bin_width * np.arange(num_frequency_bins), mel_scale=mel_scale)
        filter_freqs = mel_freqs

View on GitHub (pinned to a597f97485)

Solutions

1. Set norm=None for no normalization or norm="slaney" for area normalization
2. If translating librosa code, map librosa's norm=1 to "slaney" and norm=None to None
3. Check preprocessor_config.json / kwargs overrides for invalid norm values

Example fix

// before
mel_filters = mel_filter_bank(num_frequency_bins=257, num_mel_filters=80, sampling_rate=16000, norm=1)  # ValueError

// after
mel_filters = mel_filter_bank(num_frequency_bins=257, num_mel_filters=80, sampling_rate=16000, norm="slaney")

if norm not in (None, "slaney"):
    raise ValueError(f"norm must be None or 'slaney', got {norm!r}")
mel = mel_filter_bank(num_frequency_bins=257, num_mel_filters=80, sampling_rate=16000, norm=norm)

def is_valid_mel_norm(n) -> bool:
    return n is None or n == "slaney"

try:
    mel = mel_filter_bank(..., norm=norm)
except ValueError as e:
    if 'norm must be one of' in str(e):
        norm = None  # or "slaney" depending on desired behavior
        mel = mel_filter_bank(..., norm=norm)
    else:
        raise

When it happens

Trigger: Calling `mel_filter_bank(..., norm=...)` with values like "none", "l2", 1, True, or "Slaney" (wrong casing). Often reached indirectly through a feature extractor that computes mel filters with a norm setting read from a config.

Common situations: Porting code from librosa (which uses norm=1 or norm="slaney") without translating the value; passing a boolean where None was intended; typos or casing in preprocessor configs.

Related errors

• mel_scale should be one of "htk", "slaney" or "kaldi".
• Require num_frequency_bins: {num_frequency_bins} >= 2
• Require min_frequency: {min_frequency} <= max_frequency: {ma
• You have provided `mel_filters` but `power` is `None`. Mel s
• Cannot use log_mel option '{log_mel}' with power {power}

AI-assisted analysis of huggingface/transformers@a597f97485 (2026-08-14). Data as JSON: /api/errors/49b584b4e2a76f5e. Report an issue: GitHub.