YuE-s1-7B-anneal-en-cot - An open-source music generation model for transforming English lyrics into complete songs.

## Overview of YuE-s1-7B-anneal-en-cot YuE-s1-7B-anneal-en-cot is an open-source music generation model that transforms English lyrics into complete songs, including both vocal and accompaniment tracks. It is designed for music enthusiasts, professional creators, and developers, and supports generating high-quality music up to 5 minutes in length. The model uses Chain of Thought (CoT) conditioning and has 7 billion parameters. ## Overview of YuE-s1-7B-anneal-en-cot The key features of YuE-s1-7B-anneal-en-cot include: - Generation of complete songs with vocal and accompaniment tracks. - Support for English lyrics and multiple music styles. - Ability to generate high-quality music up to 5 minutes in length. - Use of track-decoupled next-token prediction and structural progressive conditioning. - In-context learning for style conversion. - A parameter count of 7 billion, enhancing its generation capabilities. ## Overview of YuE-s1-7B-anneal-en-cot YuE-s1-7B-anneal-en-cot generates music by accepting English lyrics as input and producing complete songs with both vocal and accompaniment tracks. The model uses Chain of Thought (CoT) conditioning to enhance the logical association between lyrics and music. It employs track-decoupled next-token prediction and structural progressive conditioning to ensure coherent and appealing music structures. ## Technical Specifications of YuE-s1-7B-anneal-en-cot The technical specifications of YuE-s1-7B-anneal-en-cot include: - Model size: 7 billion parameters. - Tensor type: BF16. - Language support: English. - Maximum generation length: 5 minutes. - Technical support: Track-decoupled, CoT conditioning. ## Usage of YuE-s1-7B-anneal-en-cot Users can utilize YuE-s1-7B-anneal-en-cot by following these steps: 1. Download the model from Hugging Face. 2. Use the provided infer.py script with appropriate parameters to generate music. 3. Ensure the environment is set up and necessary files are downloaded. 4. Customize prompts in genre.txt and lyrics.txt as needed. 5. Optimize performance by adjusting parameters such as stage2_batch_size. ## Applications of YuE-s1-7B-anneal-en-cot YuE-s1-7B-anneal-en-cot is suitable for various applications, including: - Music creation for enthusiasts and professional creators. - Experimental and commercial music projects. - Style conversion through in-context learning. - Music continuation and generation for creative projects. ## Access Points for YuE-s1-7B-anneal-en-cot Users can access YuE-s1-7B-anneal-en-cot through the following platforms: - Hugging Face: [https://huggingface.co/m-a-p/YuE-s1-7B-anneal-en-cot](https://huggingface.co/m-a-p/YuE-s1-7B-anneal-en-cot) - GitHub: [https://github.com/multimodal-art-projection/YuE](https://github.com/multimodal-art-projection/YuE) - Demo page: [https://map-yue.github.io/](https://map-yue.github.io/) - Research paper: [https://arXiv.org/abs/2503.08638](https://arXiv.org/abs/2503.08638) ### Citation sources: - [YuE-s1-7B-anneal-en-cot](https://huggingface.co/m-a-p/YuE-s1-7B-anneal-en-cot) - Official URL Updated: 2025-03-31