whisper 命令

常用示例

说明

Whisper is OpenAI's automatic speech recognition (ASR) system. It transcribes audio in many languages and can translate to English. Model sizes trade accuracy for speed: tiny runs fastest, large is most accurate. The turbo model (default) offers a good balance, running ~8x faster than large with minor quality loss. The .en suffix (tiny.en, base.en) denotes English-only models, slightly better for English. The turbo model is not trained for translation tasks. Language detection is automatic but can be hinted. For non-English audio, specifying the language improves accuracy. Translation mode transcribes any language to English text. Output formats include plain text, subtitles (SRT, VTT), and JSON with timing data. Word-level timestamps enable karaoke-style highlighting. Processing uses GPU (CUDA) when available, significantly faster than CPU. The --fp16 flag enables half-precision math on compatible GPUs. Audio preprocessing handles various formats via FFmpeg. Long files are processed in segments with context maintained across segments.

参数

--model _SIZE_

Model size: tiny, base, small, medium, large, turbo (default: turbo). English-only variants: tiny.en, base.en, small.en, medium.en.

--language _LANG_

Language code (en, de, fr, etc.) or auto.

--task _TASK_

Task: transcribe or translate.

--output_format _FORMAT_

Output format: txt, vtt, srt, tsv, json, all.

--output_dir _DIR_

Output directory.

--device _DEVICE_

Device: cpu, cuda.

--fp16 / --no-fp16

Use float16 (GPU) or float32.

--temperature _TEMP_

Sampling temperature.

--best_of _NUM_

Number of candidates.

--beam_size _NUM_

Beam search size.

--word_timestamps _BOOL_

Include word-level timestamps.

--condition_on_previous_text _BOOL_

Use previous output as context.

--verbose _BOOL_

Show progress and transcription.

--threads _NUM_

CPU threads.

--model_dir _DIR_

Directory to save and load models (default: ~/.cache/whisper).

--initial_prompt _TEXT_

Optional text to provide as prompt for the first window.

--clip_timestamps _TIMESTAMPS_

Comma-separated start/end timestamps to process specific audio segments.