import type { Language } from './language.js';
export type WhisperModel = 'tiny.en' | 'tiny' | 'base.en' | 'base' | 'small.en' | 'small' | 'medium.en' | 'medium' | 'large' | 'large-v1' | 'large-v2' | 'large-v3' | 'turbo';
export type StringOutputFormat = 'txt' | 'vtt' | 'srt' | 'tsv';
export type OutputFormat = StringOutputFormat | 'json';
export type AllOutputFormats = OutputFormat | 'all';
export interface AudioToTextOptions {
    /** Name of the Whisper model to use (default: small) */
    model?: WhisperModel;
    /** The path to save model files; uses ~/.cache/whisper by default (default: None) */
    model_dir?: string;
    /** Device to use for PyTorch inference (default: cuda) */
    device?: string;
    /** Directory to save the outputs (default: .) */
    output_dir?: string;
    /** Whether to print out the progress and debug messages (default: True) */
    verbose?: boolean;
    /** Whether to perform X->X speech recognition ('transcribe') or X->English translation ('translate') (default: transcribe) */
    task?: 'transcribe' | 'translate';
    /** Language spoken in the audio, specify None to perform language detection (default: None) */
    language?: Language;
    /** Temperature to use for sampling (default: 0) */
    temperature?: number;
    /** Number of candidates when sampling with non-zero temperature (default: 5) */
    best_of?: number;
    /** Number of beams in beam search, only applicable when temperature is zero (default: 5) */
    beam_size?: number;
    /** Optional patience value to use in beam decoding (default: None) */
    patience?: number;
    /** Optional token length penalty coefficient (alpha) (default: None) */
    length_penalty?: number;
    /** Comma-separated list of token ids to suppress during sampling (default: -1) */
    suppress_tokens?: string;
    /** Optional text to provide as a prompt for the first window (default: None) */
    initial_prompt?: string;
    /** If True, provide the previous output of the model as a prompt for the next window (default: True) */
    condition_on_previous_text?: boolean;
    /** Whether to perform inference in fp16; True by default (default: True) */
    fp16?: boolean;
    /** Temperature to increase when falling back when the decoding fails (default: 0.2) */
    temperature_increment_on_fallback?: number;
    /** If the gzip compression ratio is higher than this value, treat the decoding as failed (default: 2.4) */
    compression_ratio_threshold?: number;
    /** If the average log probability is lower than this value, treat the decoding as failed (default: -1.0) */
    logprob_threshold?: number;
    /** If the probability of the token is higher than this value AND the decoding has failed, consider the segment as silence (default: 0.6) */
    no_speech_threshold?: number;
    /** Extract word-level timestamps and refine the results based on them (default: False) */
    word_timestamps?: boolean;
    /** If word_timestamps is True, merge these punctuation symbols with the next word (default: "'“¿([{-) */
    prepend_punctuations?: string;
    /** If word_timestamps is True, merge these punctuation symbols with the previous word (default: "'.。,，!！?？:：”)]}、) */
    append_punctuations?: string;
    /** (Requires --word_timestamps True) Underline each word as it is spoken in srt and vtt (default: False) */
    highlight_words?: boolean;
    /** (Requires --word_timestamps True) The maximum number of characters in a line before breaking the line (default: None) */
    max_line_width?: number;
    /** (Requires --word_timestamps True) The maximum number of lines in a segment (default: None) */
    max_line_count?: number;
    /** The maximum number of characters per line */
    max_words_per_line?: number;
    /** Number of threads used by torch for CPU inference; supercedes MKL_NUM_THREADS/OMP_NUM_THREADS (default: 0) */
    threads?: number;
}
//# sourceMappingURL=options.d.ts.map