wav2vec2-large-xlsr-53-arabic is a fine-tuned wav2vec2 model for Arabic automatic speech recognition. Hosted on Hugging Face, it can be used with the Transformers library for both inference and further training. The open-weight model (Apache 2.0) was trained on Common Voice and Arabic speech corpora.
In the Voice, TTS & speech space, Wav2vec2 Large Xlsr 53 Arabic takes a focused approach. It focuses on performing accurate automatic speech recognition on Arabic audio. Wav2vec2 Large Xlsr 53 Arabic is an open-source project aimed at developers and researchers working with Arabic speech. Wav2vec2 Large Xlsr 53 Arabic is open source under the MIT license. It runs on the web and API.
Behind Wav2vec2 Large Xlsr 53 Arabic is jonatasgrosman, and it first shipped in 2021. Development happens publicly on GitHub with 206 stars. It operates in a well-populated space: PulseGate tracks 10 similar projects. Among its 3 catalogued features are Speech Recognition, Arabic Language, and Fine-tuned Model.
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