Speech Recognition & Audio Models

Pre-trained multilingual automatic speech recognition models built on wav2vec 2.0 and transformer architectures, enabling speech-to-text across diverse languages including Hungarian, Finnish, Persian, Chinese, Arabic, and Greek. The cluster primarily consists of fine-tuned model repositories that apply the XLSR-53 (cross-lingual speech representations) framework to language-specific datasets, leveraging PyTorch and the Hugging Face transformers library for production speech recognition applications.

Libraries, models, and frameworks for automatic speech recognition (ASR) and audio processing, with a focus on PyTorch-based implementations. The cluster centers on end-to-end neural architectures like CTC, RNN-T, and TDT for converting speech to text, with several pre-trained Parakeet model variants at different scales. Developers here will find model implementations, training pipelines, and inference tools for building speech-to-text systems.

Deep learning approaches to speech enhancement, synthesis, and real-time audio processing. The cluster centers on neural codec models and end-to-end speech systems built with PyTorch, including tools for training speech enhancement networks, full-duplex audio interaction, and audio compression. Repositories span from low-level audio DSP to high-level conversational AI applications, with a focus on practical speech quality improvements and efficient model deployment.