simeneide/tts_stortinget_qwen
The simeneide/tts_stortinget_qwen model is a 0.5 billion parameter text-to-speech (TTS) model developed by simeneide. It is based on the Qwen/Qwen2.5-0.5B architecture and has been specifically trained on 2000 hours of Norwegian parliament datasets for three epochs. This model is designed to generate speech from text, outputting wavtokenizer tokens, and is primarily optimized for Norwegian language speech synthesis.
Loading preview...
Overview
The simeneide/tts_stortinget_qwen model is a specialized text-to-speech (TTS) system, built upon the Qwen/Qwen2.5-0.5B architecture. This model has been uniquely fine-tuned for the Norwegian language, leveraging an extensive dataset of 2000 hours of recordings from the Norwegian parliament. The training process involved three epochs, focusing on adapting the base Qwen model to produce high-quality Norwegian speech.
Key Capabilities
- Norwegian Speech Synthesis: Generates natural-sounding speech specifically for the Norwegian language.
- Parliamentary Domain Expertise: Optimized using a large dataset from the Norwegian parliament, making it suitable for formal or political discourse.
- Wavtokenizer Output: Produces wavtokenizer tokens, indicating its integration with specific audio processing pipelines.
Good for
- Norwegian Language Applications: Ideal for any application requiring text-to-speech in Norwegian.
- Formal Speech Generation: Particularly well-suited for synthesizing speech for news, official announcements, or educational content related to Norwegian politics.
- Research and Development: Useful for researchers working on Norwegian language processing or TTS systems, providing a specialized base model.