Skip to content

Latest commit

 

History

History
98 lines (72 loc) · 2.71 KB

README.md

File metadata and controls

98 lines (72 loc) · 2.71 KB

WeNet Python Binding

This is a python binding of WeNet.

WeNet is a production first and production ready end-to-end speech recognition toolkit.

The best things of the binding are:

  1. Multiple languages supports, including English, Chinese. Other languages are in development.
  2. Non-streaming and streaming API
  3. N-best, contextual biasing, and timestamp supports, which are very important for speech productions.
  4. Alignment support. You can get phone level alignments this tool, on developing.

Install

Python 3.6+ is required.

pip3 install wenet

Usage

Non-streaming Usage

import sys
import wenet

model_dir = sys.argv[1]
wav_file = sys.argv[2]
decoder = wenet.Decoder(model_dir)
ans = decoder.decode_wav(wav_file)
print(ans)
# call decoder.reset() if you want to do the next decoding

model_dir is the Runtime Model directory of WeNet, it contains:

  • final.zip: runtime TorchScript ASR model.
  • words.txt: symbol table for decoding.
  • TLG.fst: optional, it means decoding with LM when TLG.fst is given.
  • units.txt: optional, e2e model units which is for generating unit level timestamp when decoded with LM.

Please refer https://github.com/wenet-e2e/wenet/blob/main/docs/pretrained_models.md for the details of Runtime Model.

You can also specify the following parameter in wenet.Decoder

  • lang (str): The language you used, chs for Chinese, and en for English.
  • nbest (int): Output the top-n best result.
  • enable_timestamp (bool): Whether to enable the word level timestamp.
  • context (List[str]): a list of context biasing words.
  • context_score (float): context bonus score.

For example:

decoder = wenet.Decoder(model_dir,
                        lang='chs',
                        nbest=5,
                        enable_timestamp=True,
                        context=['不忘初心', '牢记使命'],
                        context_score=3.0)

Streaming Usage

import sys
import wave
import wenet

model_dir = sys.argv[1]
test_wav = sys.argv[2]

with wave.open(test_wav, 'rb') as fin:
    assert fin.getnchannels() == 1
    wav = fin.readframes(fin.getnframes())

decoder = wenet.Decoder(model_dir)
# We suppose the wav is 16k, 16bits, and decode every 0.5 seconds
interval = int(0.5 * 16000) * 2
for i in range(0, len(wav), interval):
    last = False if i + interval < len(wav) else True
    chunk_wav = wav[i: min(i + interval, len(wav))]
    ans = decoder.decode(chunk_wav, last)
    print(ans)

You can use the same parameters as we introduced above to control the behavior of wenet.Decoder

Build on Your Local Machine

git clone [email protected]:wenet-e2e/wenet.git
cd wenet/runtime/binding/python
python setup.py install