Speech Recognition Wav2Vec Python* Demo¶
This demo demonstrates Automatic Speech Recognition (ASR) with pretrained Wav2Vec model.
How It Works¶
After reading and normalizing audio signal, running a neural network to get character probabilities, and CTC greedy decoding, the demo prints the decoded text.
Preparing to Run¶
The list of models supported by the demo is in
<omz_dir>/demos/speech_recognition_wav2vec_demo/python/models.lst file. This file can be used as a parameter for Model Downloader and Converter to download and, if necessary, convert models to OpenVINO IR format (*.xml + *.bin).
An example of using the Model Downloader:
omz_downloader --list models.lst
An example of using the Model Converter:
omz_converter --list models.lst
Refer to the tables Intel’s Pre-Trained Models Device Support and Public Pre-Trained Models Device Support for the details on models inference support at different devices.
Run the application with
-h option to see help message.
usage: speech_recognition_wav2vec_demo.py [-h] -m MODEL -i INPUT [-d DEVICE] [--vocab VOCAB] [--dynamic_shape] optional arguments: -h, --help Show this help message and exit. -m MODEL, --model MODEL Required. Path to an .xml file with a trained model. -i INPUT, --input INPUT Required. Path to an audio file in WAV PCM 16 kHz mono format. -d DEVICE, --device DEVICE Optional. Specify the target device to infer on, for example: CPU, GPU or HETERO. The demo will look for a suitable OpenVINO Runtime plugin for this device. Default value is CPU. --vocab VOCAB Optional. Path to an .json file with encoding vocabulary. --dynamic_shape Optional. Using dynamic shapes for inputs and outputs of model.
The typical command line is:
python3 speech_recognition_wav2vec_demo.py -m wav2vec2-base.xml -i audio.wav
Only 16-bit, 16 kHz, mono-channel WAVE audio files are supported.
An example audio file can be taken from OpenVINO test data folder.
The application prints the decoded text for the audio file. The demo reports
Latency : total processing time required to process input data (from reading the data to displaying the results).