Faster Whisper transcription with CTranslate2

faster-whisper is a reimplementation of OpenAI's Whisper model using CTranslate2, which is a fast inference engine for Transformer models.

This implementation is up to 4 times faster than openai/whisper for the same accuracy while using less memory. The efficiency can be further improved with 8-bit quantization on both CPU and GPU.

Benchmark

Whisper

For reference, here's the time and memory usage that are required to transcribe 13 minutes of audio using different implementations:

• openai/whisper@6dea21fd
• whisper.cpp@3b010f9
• faster-whisper@cce6b53e

Large-v2 model on GPU

Implementation	Precision	Beam size	Time	Max. GPU memory	Max. CPU memory
openai/whisper	fp16	5	4m30s	11325MB	9439MB
faster-whisper	fp16	5	54s	4755MB	3244MB
faster-whisper	int8	5	59s	3091MB	3117MB

Executed with CUDA 11.7.1 on a NVIDIA Tesla V100S.

Small model on CPU

Implementation	Precision	Beam size	Time	Max. memory
openai/whisper	fp32	5	10m31s	3101MB
whisper.cpp	fp32	5	17m42s	1581MB
whisper.cpp	fp16	5	12m39s	873MB
faster-whisper	fp32	5	2m44s	1675MB
faster-whisper	int8	5	2m04s	995MB

Executed with 8 threads on a Intel(R) Xeon(R) Gold 6226R.

Distil-whisper

Implementation	Precision	Beam size	Time	Gigaspeech WER
distil-whisper/distil-large-v2	fp16	4	-	10.36
faster-distil-large-v2	fp16	5	-	10.28
distil-whisper/distil-medium.en	fp16	4	-	11.21
faster-distil-medium.en	fp16	5	-	11.21

Executed with CUDA 11.4 on a NVIDIA 3090.

testing details (click to expand)

For distil-whisper/distil-large-v2, the WER is tested with code sample from link. for faster-distil-whisper, the WER is tested with setting:

from faster_whisper import WhisperModel

model_size = "distil-large-v2"
# model_size = "distil-medium.en"
# Run on GPU with FP16
model = WhisperModel(model_size, device="cuda", compute_type="float16")
segments, info = model.transcribe("audio.mp3", beam_size=5, language="en")

Requirements

• Python 3.8 or greater

Unlike openai-whisper, FFmpeg does not need to be installed on the system. The audio is decoded with the Python library PyAV which bundles the FFmpeg libraries in its package.

GPU

GPU execution requires the following NVIDIA libraries to be installed:

• cuBLAS for CUDA 11
• cuDNN 8 for CUDA 11

There are multiple ways to install these libraries. The recommended way is described in the official NVIDIA documentation, but we also suggest other installation methods below.

Other installation methods (click to expand)

Use Docker

The libraries are installed in this official NVIDIA Docker image: nvidia/cuda:11.8.0-cudnn8-runtime-ubuntu22.04.

Install with pip (Linux only)

On Linux these libraries can be installed with pip. Note that LD_LIBRARY_PATH must be set before launching Python.

pip install nvidia-cublas-cu11 nvidia-cudnn-cu11

export LD_LIBRARY_PATH=`python3 -c 'import os; import nvidia.cublas.lib; import nvidia.cudnn.lib; print(os.path.dirname(nvidia.cublas.lib.__file__) + ":" + os.path.dirname(nvidia.cudnn.lib.__file__))'`

Download the libraries from Purfview's repository (Windows & Linux)

Purfview's whisper-standalone-win provides the required NVIDIA libraries for Windows & Linux in a single archive. Decompress the archive and place the libraries in a directory included in the PATH.

Installation

The module can be installed from PyPI:

pip install faster-whisper

Other installation methods (click to expand)

Install the master branch

pip install --force-reinstall "faster-whisper @ https://github.com/guillaumekln/faster-whisper/archive/refs/heads/master.tar.gz"

Install a specific commit

pip install --force-reinstall "faster-whisper @ https://github.com/guillaumekln/faster-whisper/archive/a4f1cc8f11433e454c3934442b5e1a4ed5e865c3.tar.gz"

Usage

Faster-whisper

from faster_whisper import WhisperModel

model_size = "large-v3"

# Run on GPU with FP16
model = WhisperModel(model_size, device="cuda", compute_type="float16")

# or run on GPU with INT8
# model = WhisperModel(model_size, device="cuda", compute_type="int8_float16")
# or run on CPU with INT8
# model = WhisperModel(model_size, device="cpu", compute_type="int8")

segments, info = model.transcribe("audio.mp3", beam_size=5)

print("Detected language '%s' with probability %f" % (info.language, info.language_probability))

for segment in segments:
    print("[%.2fs -> %.2fs] %s" % (segment.start, segment.end, segment.text))

Warning: segments is a generator so the transcription only starts when you iterate over it. The transcription can be run to completion by gathering the segments in a list or a for loop:

segments, _ = model.transcribe("audio.mp3")
segments = list(segments)  # The transcription will actually run here.

Faster-distil-whisper

For usage of faster-ditil-whisper, please refer to: SYSTRAN#533

model_size = "distil-large-v2"
# model_size = "distil-medium.en"
model = WhisperModel(model_size, device="cuda", compute_type="float16")
segments, info = model.transcribe("audio.mp3", beam_size=5, 
    language="en", max_new_tokens=128, condition_on_previous_text=False)

NOTE: emprically, condition_on_previous_text=True will degrade the performance of faster-distil-whisper for long audio. Degradation on the first chunk was observed with initial_prompt too.

Word-level timestamps

segments, _ = model.transcribe("audio.mp3", word_timestamps=True)

for segment in segments:
    for word in segment.words:
        print("[%.2fs -> %.2fs] %s" % (word.start, word.end, word.word))

VAD filter

The library integrates the Silero VAD model to filter out parts of the audio without speech:

segments, _ = model.transcribe("audio.mp3", vad_filter=True)

The default behavior is conservative and only removes silence longer than 2 seconds. See the available VAD parameters and default values in the source code. They can be customized with the dictionary argument vad_parameters:

segments, _ = model.transcribe(
    "audio.mp3",
    vad_filter=True,
    vad_parameters=dict(min_silence_duration_ms=500),
)

Logging

The library logging level can be configured like this:

import logging

logging.basicConfig()
logging.getLogger("faster_whisper").setLevel(logging.DEBUG)

Going further

See more model and transcription options in the WhisperModel class implementation.

Community integrations

Here is a non exhaustive list of open-source projects using faster-whisper. Feel free to add your project to the list!

• whisper-ctranslate2 is a command line client based on faster-whisper and compatible with the original client from openai/whisper.
• whisper-diarize is a speaker diarization tool that is based on faster-whisper and NVIDIA NeMo.
• whisper-standalone-win Standalone CLI executables of faster-whisper for Windows, Linux & macOS.
• asr-sd-pipeline provides a scalable, modular, end to end multi-speaker speech to text solution implemented using AzureML pipelines.
• Open-Lyrics is a Python library that transcribes voice files using faster-whisper, and translates/polishes the resulting text into .lrc files in the desired language using OpenAI-GPT.
• wscribe is a flexible transcript generation tool supporting faster-whisper, it can export word level transcript and the exported transcript then can be edited with wscribe-editor
• aTrain is a graphical user interface implementation of faster-whisper developed at the BANDAS-Center at the University of Graz for transcription and diarization in Windows (Windows Store App) and Linux.
• Whisper-Streaming implements real-time mode for offline Whisper-like speech-to-text models with faster-whisper as the most recommended back-end. It implements a streaming policy with self-adaptive latency based on the actual source complexity, and demonstrates the state of the art.
• WhisperLive is a nearly-live implementation of OpenAI's Whisper which uses faster-whisper as the backend to transcribe audio in real-time.

Model conversion

When loading a model from its size such as WhisperModel("large-v3"), the correspondig CTranslate2 model is automatically downloaded from the Hugging Face Hub.

We also provide a script to convert any Whisper models compatible with the Transformers library. They could be the original OpenAI models or user fine-tuned models.

For example the command below converts the original "large-v3" Whisper model and saves the weights in FP16:

pip install transformers[torch]>=4.23

ct2-transformers-converter --model openai/whisper-large-v3 --output_dir whisper-large-v3-ct2
--copy_files tokenizer.json preprocessor_config.json --quantization float16

• The option --model accepts a model name on the Hub or a path to a model directory.
• If the option --copy_files tokenizer.json is not used, the tokenizer configuration is automatically downloaded when the model is loaded later.

Models can also be converted from the code. See the conversion API.

Load a converted model

1. Directly load the model from a local directory:

model = faster_whisper.WhisperModel("whisper-large-v3-ct2")

2. Upload your model to the Hugging Face Hub and load it from its name:

model = faster_whisper.WhisperModel("username/whisper-large-v3-ct2")

Comparing performance against other implementations

If you are comparing the performance against other Whisper implementations, you should make sure to run the comparison with similar settings. In particular:

• Verify that the same transcription options are used, especially the same beam size. For example in openai/whisper, model.transcribe uses a default beam size of 1 but here we use a default beam size of 5.
• When running on CPU, make sure to set the same number of threads. Many frameworks will read the environment variable OMP_NUM_THREADS, which can be set when running your script:

OMP_NUM_THREADS=4 python3 my_script.py