TinyMusician: On-Device Music Generation with Knowledge Distillation and Mixed Precision Quantization

• URL: http://arxiv.org/abs/2509.00914v1
• Date: Sun, 31 Aug 2025 15:57:13 GMT
• Title: TinyMusician: On-Device Music Generation with Knowledge Distillation and Mixed Precision Quantization
• Abstract summary: We present TinyMusician, a lightweight music generation model distilled from MusicGen.<n>TinyMusician integrates two innovations: Stage-mixed Bidirectional and Skewed KL-Divergence.<n>Results demonstrate that TinyMusician retains 93% of the MusicGen-Small performance with 55% less model size.
• Score: 2.633261256867499
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: The success of the generative model has gained unprecedented attention in the music generation area. Transformer-based architectures have set new benchmarks for model performance. However, their practical adoption is hindered by some critical challenges: the demand for massive computational resources and inference time, due to their large number of parameters. These obstacles make them infeasible to deploy on edge devices, such as smartphones and wearables, with limited computational resources. In this work, we present TinyMusician, a lightweight music generation model distilled from MusicGen (a State-of-the-art music generation model). TinyMusician integrates two innovations: (i) Stage-mixed Bidirectional and Skewed KL-Divergence and (ii) Adaptive Mixed-Precision Quantization. The experimental results demonstrate that TinyMusician retains 93% of the MusicGen-Small performance with 55% less model size. TinyMusician is the first mobile-deployable music generation model that eliminates cloud dependency while maintaining high audio fidelity and efficient resource usage

Related papers

• SqueezeComposer: Temporal Speed-up is A Simple Trick for Long-form Music Composing [35.732692220471606]
  We assume that AI models can understand and generate time-accelerated (speeded-up) audio at rates such as 2x, 4x, or even 8x.<n>By first generating a high-speed version of the music, we greatly reduce the temporal length and resource requirements.<n>We instantiate this idea in SqueezeComposer, a framework that employs diffusion models for generation in the accelerated domain and refinement in the restored domain.
  arXiv  Detail & Related papers  (2026-03-22T06:00:41Z)
• Muse: Towards Reproducible Long-Form Song Generation with Fine-Grained Style Control [66.46754271097555]
  We release a fully open-source system for long-form song generation with fine-grained style conditioning.<n>The dataset consists of 116k fully licensed synthetic songs with automatically generated lyrics and style descriptions.<n>We train Muse via single-stage supervised finetuning of a Qwen-based language model extended with discrete audio tokens.
  arXiv  Detail & Related papers  (2026-01-07T14:40:48Z)
• Moonbeam: A MIDI Foundation Model Using Both Absolute and Relative Music Attributes [9.283206048560322]
  Moonbeam is a transformer-based foundation model for symbolic music.<n>It is pretrained on a large and diverse collection of MIDI data totaling 81.6K hours of music and 18 billion tokens.<n>We open-source the code, pretrained model, and generated samples on Github.
  arXiv  Detail & Related papers  (2025-05-21T14:17:25Z)
• InspireMusic: Integrating Super Resolution and Large Language Model for High-Fidelity Long-Form Music Generation [43.690876909464336]
  We introduce InspireMusic, a framework integrated super resolution and large language model for high-fidelity long-form music generation.<n>A unified framework generates high-fidelity music, songs, and audio, which incorporates an autoregressive transformer with a super-resolution flow-matching model.<n>Our model differs from previous approaches, as we utilize an audio tokenizer with one codebook that contains richer semantic information.
  arXiv  Detail & Related papers  (2025-02-28T09:58:25Z)
• MusicFlow: Cascaded Flow Matching for Text Guided Music Generation [53.63948108922333]
  MusicFlow is a cascaded text-to-music generation model based on flow matching. We leverage masked prediction as the training objective, enabling the model to generalize to other tasks such as music infilling and continuation.
  arXiv  Detail & Related papers  (2024-10-27T15:35:41Z)
• UniMuMo: Unified Text, Music and Motion Generation [57.72514622935806]
  We introduce UniMuMo, a unified multimodal model capable of taking arbitrary text, music, and motion data as input conditions to generate outputs across all three modalities. By converting music, motion, and text into token-based representation, our model bridges these modalities through a unified encoder-decoder transformer architecture.
  arXiv  Detail & Related papers  (2024-10-06T16:04:05Z)
• Instruct-MusicGen: Unlocking Text-to-Music Editing for Music Language Models via Instruction Tuning [24.6866990804501]
  Instruct-MusicGen is a novel approach that finetunes a pretrained MusicGen model to efficiently follow editing instructions.<n>Remarkably, Instruct-MusicGen only introduces 8% new parameters to the original MusicGen model and only trains for 5K steps.
  arXiv  Detail & Related papers  (2024-05-28T17:27:20Z)
• Arrange, Inpaint, and Refine: Steerable Long-term Music Audio Generation and Editing via Content-based Controls [6.176747724853209]
  Large Language Models (LLMs) have shown promise in generating high-quality music, but their focus on autoregressive generation limits their utility in music editing tasks. We propose a novel approach leveraging a parameter-efficient heterogeneous adapter combined with a masking training scheme. Our method integrates frame-level content-based controls, facilitating track-conditioned music refinement and score-conditioned music arrangement.
  arXiv  Detail & Related papers  (2024-02-14T19:00:01Z)
• Simple and Controllable Music Generation [94.61958781346176]
  MusicGen is a single Language Model (LM) that operates over several streams of compressed discrete music representation, i.e., tokens. Unlike prior work, MusicGen is comprised of a single-stage transformer LM together with efficient token interleaving patterns.
  arXiv  Detail & Related papers  (2023-06-08T15:31:05Z)
• Exploring the Efficacy of Pre-trained Checkpoints in Text-to-Music Generation Task [86.72661027591394]
  We generate complete and semantically consistent symbolic music scores from text descriptions. We explore the efficacy of using publicly available checkpoints for natural language processing in the task of text-to-music generation. Our experimental results show that the improvement from using pre-trained checkpoints is statistically significant in terms of BLEU score and edit distance similarity.
  arXiv  Detail & Related papers  (2022-11-21T07:19:17Z)
• Museformer: Transformer with Fine- and Coarse-Grained Attention for Music Generation [138.74751744348274]
  We propose Museformer, a Transformer with a novel fine- and coarse-grained attention for music generation. Specifically, with the fine-grained attention, a token of a specific bar directly attends to all the tokens of the bars that are most relevant to music structures. With the coarse-grained attention, a token only attends to the summarization of the other bars rather than each token of them so as to reduce the computational cost.
  arXiv  Detail & Related papers  (2022-10-19T07:31:56Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.