whisper Documentation

File: README.md Copy Chapter

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 Copy Chapter

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 Copy Chapter

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 Copy Chapter

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.

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.

Command-line usage Copy Chapter

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 Copy Chapter

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 Copy Chapter

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 Copy Chapter

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

File: data/README.md Copy Chapter

This directory supplements the paper with more details on how we prepared the data for evaluation, to help replicate our experiments.

Short-form English-only datasets Copy Chapter

LibriSpeech Copy Section

We used the test-clean and test-other splits from the LibriSpeech ASR corpus.

TED-LIUM 3 Copy Section

We used the test split of TED-LIUM Release 3, using the segmented manual transcripts included in the release.

Common Voice 5.1 Copy Section

We downloaded the English subset of Common Voice Corpus 5.1 from the official website

Artie Copy Section

We used the Artie bias corpus. This is a subset of the Common Voice dataset.

CallHome & Switchboard Copy Section

We used the two corpora from LDC2002S09 and LDC2002T43 and followed the eval2000_data_prep.sh script for preprocessing. The wav.scp files can be converted to WAV files with the following bash commands:

mkdir -p wav
while read name cmd; do
    echo $name
    echo ${cmd/\|/} wav/$name.wav | bash
done < wav.scp

WSJ Copy Section

We used LDC93S6B and LDC94S13B and followed the s5 recipe to preprocess the dataset.

CORAAL Copy Section

We used the 231 interviews from CORAAL (v. 2021.07) and used the segmentations from the FairSpeech project.

CHiME-6 Copy Section

We downloaded the CHiME-5 dataset and followed the stage 0 of the s5_track1 recipe to create the CHiME-6 dataset which fixes synchronization. We then used the binaural recordings (*_P??.wav) and the corresponding transcripts.

AMI-IHM, AMI-SDM1 Copy Section

We preprocessed the AMI Corpus by following the stage 0 and 2 of the s5b recipe.

Long-form English-only datasets Copy Chapter

TED-LIUM 3 Copy Section

To create a long-form transcription dataset from the TED-LIUM3 dataset, we sliced the audio between the beginning of the first labeled segment and the end of the last labeled segment of each talk, and we used the concatenated text as the label. Below are the timestamps used for slicing each of the 11 TED talks in the test split.

Meanwhile Copy Section

This dataset consists of 64 segments from The Late Show with Stephen Colbert. The YouTube video ID, start and end timestamps, and the labels can be found in meanwhile.json. The labels are collected from the closed-caption data for each video and corrected with manual inspection.

Rev16 Copy Section

We use a subset of 16 files from the 30 podcast episodes in Rev.AI's Podcast Transcription Benchmark, after finding that there are multiple cases where a significant portion of the audio and the labels did not match, mostly on the parts introducing the sponsors. We selected 16 episodes that do not have this error, whose "file number" are:

3 4 9 10 11 14 17 18 20 21 23 24 26 27 29 32

Kincaid46 Copy Section

This dataset consists of 46 audio files and the corresponding transcripts compiled in the blog article Which automatic transcription service is the most accurate - 2018 by Jason Kincaid. We used the 46 audio files and reference transcripts from the Airtable widget in the article.

For the human transcription benchmark in the paper, we use a subset of 25 examples from this data, whose "Ref ID" are:

2 4 5 8 9 10 12 13 14 16 19 21 23 25 26 28 29 30 33 35 36 37 42 43 45

Earnings-21, Earnings-22 Copy Section

For these datasets, we used the files available in the speech-datasets repository, as of their 202206 version.

CORAAL Copy Section

We used the 231 interviews from CORAAL (v. 2021.07) and used the full-length interview files and transcripts.

Multilingual datasets Copy Chapter

Multilingual LibriSpeech Copy Section

We used the test splits from each language in the Multilingual LibriSpeech (MLS) corpus.

Fleurs Copy Section

We collected audio files and transcripts using the implementation available as HuggingFace datasets. To use as a translation dataset, we matched the numerical utterance IDs to find the corresponding transcript in English.

VoxPopuli Copy Section

We used the get_asr_data.py script from the official repository to collect the ASR data in 14 languages.

Common Voice 9 Copy Section

We downloaded the Common Voice Corpus 9 from the official website

CoVOST 2 Copy Section

We collected the X into English data collected using the official repository.