![]() detect_language ( mel ) print ( f "Detected language: " ) # decode the audio options = whisper. device ) # detect the spoken language _, probs = model. pad_or_trim ( audio ) # make log-Mel spectrogram and move to the same device as the model mel = whisper. load_audio ( "audio.mp3" ) audio = whisper. load_model ( "base" ) # load audio and pad/trim it to fit 30 seconds audio = whisper. 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.īelow is an example usage of tect_language() and code() which provide lower-level access to the model. transcribe ( "audio.mp3" ) print ( result ) Transcription can also be performed within Python: import whisper model = whisper. See tokenizer.py for the list of all available languages. Run the following to view all available options: whisper -help To transcribe an audio file containing non-English speech, you can specify the language using the -language option: whisper japanese.wav -language JapaneseĪdding -task translate will translate the speech into English: whisper japanese.wav -language Japanese -task translate The default setting (which selects the small model) works well for transcribing English. The following command will transcribe speech in audio files, using the medium model: whisper audio.flac audio.mp3 audio.wav -model medium 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. 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. Whisper's performance varies widely depending on the language. We observed that the difference becomes less significant for the small.en and medium.en models. en models for English-only applications tend to perform better, especially for the tiny.en and base.en models. ![]() Below are the names of the available models and their approximate memory requirements and inference speed relative to the large model actual speed may vary depending on many factors including the available hardware. There are five model sizes, four with English-only versions, offering speed and accuracy tradeoffs. If the installation fails with No module named 'setuptools_rust', you need to install setuptools_rust, e.g. Additionally, you may need to configure the PATH environment variable, e.g. If you see installation errors during the pip install command above, please follow the Getting started page to install Rust development environment. You may need rust installed as well, in case tiktoken does not provide a pre-built wheel for your platform. Sudo apt update & sudo apt install ffmpeg 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 ![]() To update the package to the latest version of this repository, please run: pip install -upgrade -no-deps -force-reinstall git+ You can download and install (or update to) the latest release of Whisper with the following command: pip install -U openai-whisperĪlternatively, the following command will pull and install the latest commit from this repository, along with its Python dependencies: pip install git+ ![]() The codebase also depends on a few Python packages, most notably OpenAI's tiktoken for their fast tokenizer implementation. 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 multitask training format uses a set of special tokens that serve as task specifiers or classification targets. 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. ApproachĪ 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. 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. Whisper is a general-purpose speech recognition model. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |