
Sign up to save your podcasts
Or


Paper: https://arxiv.org/pdf/2501.17161
Paper: https://github.com/deepseek-ai/Janus/blob/main/janus_pro_tech_report.pdf
Paper: https://arxiv.org/pdf/2412.09764
Paper: https://scontent-dfw5-1.xx.fbcdn.net/...
Technical Report: https://arxiv.org/pdf/2412.19437
Paper: https://arxiv.org/pdf/2309.16588
Paper: https://arxiv.org/pdf/2412.09871v1.pdf
This research paper introduces CosyVoice 2, an improved streaming speech synthesis model. Building upon its predecessor, CosyVoice 2 utilizes advancements in large language models (LLMs) and incorporates optimizations like finite scalar quantization and a chunk-aware causal flow matching model. The result is a system achieving near human-parity naturalness with minimal latency in streaming mode, supporting multiple languages and offering fine-grained control over speech characteristics. The paper details the model's architecture, training data, and experimental results, demonstrating its superior performance compared to existing models. Limitations and future research directions are also discussed.
Blog: https://openai.com/12-days/
Paper: https://arxiv.org/pdf/2412.14093
From the publisher's feed