Japanese HuBERT Base is an open-weight speech representation model trained on approximately 19,000 hours of Japanese speech from ReazonSpeech v1. Developers can use it locally with PyTorch and Safetensors for speech recognition and related audio-processing workflows.
Japanese Hubert Base is a Foundation models & chat project. It focuses on building Japanese speech recognition and audio-processing systems without training a speech representation model from scratch. Japanese Hubert Base is an open-source project aimed at machine learning developers and speech researchers. Japanese Hubert Base is open source under the MIT license. It ships for the web, and it can be self-hosted.
Behind Japanese Hubert Base is rinna Co., Ltd., and it first shipped in 2017. The GitHub repository has been archived. Among its 6 catalogued features are speech representation, japanese speech model, and pyTorch support.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do