- Wav2letter@anywhere was developed by by Facebook AI Research (FAIR)
- Research has predicted that the joint speech and voice recognition market would be worth $31.82 billion in 2025
According to a report by Mobileidworld, Facebook has announced that it will make its wav2letter@anywhere online speech recognition framework easily available as an open source platform. It was developed by Facebook AI Research (FAIR). It claimed that it had created the fastest open source automatic speech recognition (ASR) platform in the market.
The report said that eight FAIR researchers in a recent paper said that the system has almost three times the throughput of a well-tuned hybrid ASR baseline. It also has a lower latency and a better word error rate.
Time-depth separable convolutional neural network
The report said that wav2letter@anywhere framework is based on the wav2letter and wav2letter++ neural net language models. It utilises time-depth separable (TDS) convolutional neural network (CNN) technology instead of the recurrent neural network (RNN) technology.
Speech recognition has become very common in the recent past. The report said that research has predicted that the joint speech and voice recognition market would be worth $31.82 billion in 2025.