NotaGen: Advancing Musicality in Symbolic Music Generation with Large Language Model Training Paradigms
This work addresses the challenge of improving musical aesthetics in symbolic music generation for applications in music composition and AI creativity, representing an incremental advancement by adapting existing LLM techniques to a specific domain.
The paper tackles the problem of generating high-quality classical sheet music by introducing NotaGen, a model that uses large language model training paradigms including a novel reinforcement learning method called CLaMP-DPO, resulting in outperforming baseline models in subjective A/B tests against human compositions.
We introduce NotaGen, a symbolic music generation model aiming to explore the potential of producing high-quality classical sheet music. Inspired by the success of Large Language Models (LLMs), NotaGen adopts pre-training, fine-tuning, and reinforcement learning paradigms (henceforth referred to as the LLM training paradigms). It is pre-trained on 1.6M pieces of music in ABC notation, and then fine-tuned on approximately 9K high-quality classical compositions conditioned on "period-composer-instrumentation" prompts. For reinforcement learning, we propose the CLaMP-DPO method, which further enhances generation quality and controllability without requiring human annotations or predefined rewards. Our experiments demonstrate the efficacy of CLaMP-DPO in symbolic music generation models with different architectures and encoding schemes. Furthermore, subjective A/B tests show that NotaGen outperforms baseline models against human compositions, greatly advancing musical aesthetics in symbolic music generation.