whisper README

The project's README from github.com/openai/whisper. As one markdown file: README.md.

Whisper

[Blog] [Paper] [Model card] [Colab example]

Whisper is a general-purpose speech recognition model. It is trained on a large dataset of diverse audio and is also a multitasking model that can perform multilingual speech recognition, speech translation, and language identification.

Approach

Approach

A Transformer sequence-to-sequence model is trained on various speech processing tasks, including multilingual speech recognition, speech translation, spoken language identification, and voice activity detection. These tasks are jointly represented as a sequence of tokens to be predicted by the decoder, allowing a single model to replace many stages of a traditional speech-processing pipeline. The multitask training format uses a set of special tokens that serve as task specifiers or classification targets.

Setup

We used Python 3.9.9 and PyTorch 1.10.1 to train and test our models, but the codebase is expected to be compatible with Python 3.8-3.11 and recent PyTorch versions. The codebase also depends on a few Python packages, most notably OpenAI's tiktoken for their fast tokenizer implementation. You can download and install (or update to) the latest release of Whisper with the following command:

pip install -U openai-whisper

Alternatively, the following command will pull and install the latest commit from this repository, along with its Python dependencies:

pip install git+https://github.com/openai/whisper.git

To update the package to the latest version of this repository, please run:

pip install --upgrade --no-deps --force-reinstall git+https://github.com/openai/whisper.git

It also requires the command-line tool ffmpeg to be installed on your system, which is available from most package managers:

# on Ubuntu or Debian
sudo apt update && sudo apt install ffmpeg

# on Arch Linux
sudo pacman -S ffmpeg

# on MacOS using Homebrew (https://brew.sh/)
brew install ffmpeg

# on Windows using Chocolatey (https://chocolatey.org/)
choco install ffmpeg

# on Windows using Scoop (https://scoop.sh/)
scoop install ffmpeg

You may need rust installed as well, in case tiktoken does not provide a pre-built wheel for your platform. If you see installation errors during the pip install command above, please follow the Getting started page to install Rust development environment. Additionally, you may need to configure the PATH environment variable, e.g. export PATH="$HOME/.cargo/bin:$PATH". If the installation fails with No module named 'setuptools_rust', you need to install setuptools_rust, e.g. by running:

pip install setuptools-rust

Available models and languages

There are six model sizes, four with English-only versions, offering speed and accuracy tradeoffs. Below are the names of the available models and their approximate memory requirements and inference speed relative to the large model. The relative speeds below are measured by transcribing English speech on a A100, and the real-world speed may vary significantly depending on many factors including the language, the speaking speed, and the available hardware.

Size Parameters English-only model Multilingual model Required VRAM Relative speed
tiny 39 M tiny.en tiny ~1 GB ~10x
base 74 M base.en base ~1 GB ~7x
small 244 M small.en small ~2 GB ~4x
medium 769 M medium.en medium ~5 GB ~2x
large 1550 M N/A large ~10 GB 1x
turbo 809 M N/A turbo ~6 GB ~8x

The .en models for English-only applications tend to perform better, especially for the tiny.en and base.en models. We observed that the difference becomes less significant for the small.en and medium.en models. Additionally, the turbo model is an optimized version of large-v3 that offers faster transcription speed with a minimal degradation in accuracy.

Whisper's performance varies widely depending on the language. The figure below shows a performance breakdown of large-v3 and large-v2 models by language, using WERs (word error rates) or CER (character error rates, shown in Italic) evaluated on the Common Voice 15 and Fleurs datasets. Additional WER/CER metrics corresponding to the other models and datasets can be found in Appendix D.1, D.2, and D.4 of the paper, as well as the BLEU (Bilingual Evaluation Understudy) scores for translation in Appendix D.3.

WER breakdown by language

Command-line usage

The following command will transcribe speech in audio files, using the turbo model:

whisper audio.flac audio.mp3 audio.wav --model turbo

The default setting (which selects the turbo model) works well for transcribing English. However, the turbo model is not trained for translation tasks. If you need to translate non-English speech into English, use one of the multilingual models (tiny, base, small, medium, large) instead of turbo.

For example, to transcribe an audio file containing non-English speech, you can specify the language:

whisper japanese.wav --language Japanese

To translate speech into English, use:

whisper japanese.wav --model medium --language Japanese --task translate

Note: The turbo model will return the original language even if --task translate is specified. Use medium or large for the best translation results.

Run the following to view all available options:

whisper --help

See tokenizer.py for the list of all available languages.

Python usage

Transcription can also be performed within Python: 

import whisper

model = whisper.load_model("turbo")
result = model.transcribe("audio.mp3")
print(result["text"])

Internally, the transcribe() method reads the entire file and processes the audio with a sliding 30-second window, performing autoregressive sequence-to-sequence predictions on each window.

Below is an example usage of whisper.detect_language() and whisper.decode() which provide lower-level access to the model.

import whisper

model = whisper.load_model("turbo")

# load audio and pad/trim it to fit 30 seconds
audio = whisper.load_audio("audio.mp3")
audio = whisper.pad_or_trim(audio)

# make log-Mel spectrogram and move to the same device as the model
mel = whisper.log_mel_spectrogram(audio, n_mels=model.dims.n_mels).to(model.device)

# detect the spoken language
_, probs = model.detect_language(mel)
print(f"Detected language: {max(probs, key=probs.get)}")

# decode the audio
options = whisper.DecodingOptions()
result = whisper.decode(model, mel, options)

# print the recognized text
print(result.text)

More examples

Please use the 🙌 Show and tell category in Discussions for sharing more example usages of Whisper and third-party extensions such as web demos, integrations with other tools, ports for different platforms, etc.

License

Whisper's code and model weights are released under the MIT License. See LICENSE for further details.

whisper (openai/whisper)

Searching Database…

110.8K ·GitHub
Whisper
Что это такое?Whisper — это модель распознавания речи общего назначения, обученная на большом и разнообразном наборе данных с неточной разметкой. Она умеет выполнять многоязычную транскрибацию, переводить речь на английский и определять язык, на котором говорят.
Это самостоятельное приложение или библиотека, которую я использую в коде?Это библиотека на Python со встроенной утилитой командной строки: её устанавливают через pip, а затем либо запускают команду 'whisper' для аудиофайла, либо вызывают из своего Python-кода.
Как установить и запустить?Установите командой 'pip install -U openai-whisper' (или 'pip install git+...' для последнего коммита), затем расшифруйте файл, выполнив 'whisper audio.mp3' в командной строке; по умолчанию используется модель turbo.
Что нужно установить до запуска?Нужны Python 3.8-3.11 и PyTorch, а также утилита командной строки ffmpeg, установленная в системе. Если для tiktoken нет готовой сборки (wheel) под вашу платформу, может понадобиться набор инструментов Rust.
Самая быстрая модель умеет и переводить?Нет — модель turbo по умолчанию оптимизирована для скорости, но не обучена переводу и будет возвращать текст на исходном языке, даже если вы запросите задачу перевода; для перевода используйте многоязычные модели medium или large.
Сколько видеопамяти (VRAM) мне нужно, чтобы запустить модель?Всё зависит от размера модели: примерно 1 ГБ для tiny и base, 2 ГБ для small, 5 ГБ для medium, до 10 ГБ для large и 6 ГБ для turbo; скорость измерялась на GPU A100.
Какие размеры моделей доступны?Есть шесть размеров: tiny (39 млн параметров), base (74 млн), small (244 млн), medium (769 млн), large (1550 млн) и turbo (809 млн); каждый из них даёт свой баланс между скоростью и точностью. Модели от tiny до medium также есть в вариантах только для английского, с суффиксом '.en'.
Зачем выбирать модель только для английского (.en)?Модели .en обычно работают лучше своих многоязычных аналогов при использовании только английского языка, а самый заметный выигрыш виден у размеров tiny.en и base.en.
Чем отличается модель turbo?Turbo — это оптимизированная версия large-v3: она расшифровывает быстрее при незначительном падении точности, но не поддерживает перевод, который есть у других многоязычных моделей.
Как обрабатываются длинные аудиофайлы?Метод transcribe() читает весь файл и обрабатывает его скользящим окном в 30 секунд, выполняя авторегрессионные предсказания «последовательность в последовательность» для каждого окна.
What are similar projects?faster-whisper, WhisperLiveKit, buzz, pyvideotrans, whisperX, FunASR, DeepSpeech, VibeVoice
Which topics is it in?asr, transcription, whisper, pretrained-models, speech-to-text, speech-recognition, translation, transformer, audio, artificial-intelligence, pytorch, openai
Creator
openai
Size
12.5 MB
Stars
110.8K
Updated
CommitAug 31Releasev20250625Jun 26
Age
4y
Commits · 90 Days