Can I read PROSODYLM: Advancing Speech Prosody in LLMs on EtoBox?

PROSODYLM: Advancing Speech Prosody in LLMs by pol.buitrago is a document available to read on EtoBox.

What is PROSODYLM: Advancing Speech Prosody in LLMs about?

The paper presents P ROSODY LM, a speech language model designed to enhance prosody processing in speech by utilizing a novel tokenization scheme that incorporates word-level prosody tokens alongside text transcriptions. This approach allows the model to effectively learn the interdependencies between content and prosody, resulting in improved capabilities such as recognizing emotional cues and maintaining prosodic consistency. The findings suggest that P ROSODY LM can achieve diverse prosody processing cap

Author

pol.buitrago

Language

EN