{"ID":2865793,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2509.20802","arxiv_id":"2509.20802","title":"SPADE: Structured Pruning and Adaptive Distillation for Efficient LLM-TTS","abstract":"The goal of this paper is to introduce SPADE, a framework for Structured Pruning and Adaptive Distillation for Efficient Large Language Model-based text-to-speech (LLM-TTS). Recent LLM-TTS systems achieve strong controllability and zero-shot generalization, but their large parameter counts and high latency limit real-world deployment. SPADE addresses this by combining (i) a pruning step guided by a word-error-rate-based layer importance index to remove non-essential Transformer layers, with (ii) multi-level knowledge distillation to restore autoregressive coherence. On zero-shot benchmarks, SPADE preserves near-parity perceptual quality while halving Transformer depth, reducing VRAM usage by up to 20%, and achieving up to 1.7x faster real-time factor with less than 5% of the original training data. These results show that compact LLM-TTS models can maintain naturalness and speaker similarity while enabling practical real-time speech generation. Audio samples are available at https://mm.kaist.ac.kr/projects/SPADE/.","short_abstract":"The goal of this paper is to introduce SPADE, a framework for Structured Pruning and Adaptive Distillation for Efficient Large Language Model-based text-to-speech (LLM-TTS). Recent LLM-TTS systems achieve strong controllability and zero-shot generalization, but their large parameter counts and high latency limit real-w...","url_abs":"https://arxiv.org/abs/2509.20802","url_pdf":"https://arxiv.org/pdf/2509.20802v3","authors":"[\"Tan Dat Nguyen\",\"Jaehun Kim\",\"Ji-Hoon Kim\",\"Shukjae Choi\",\"Youshin Lim\",\"Joon Son Chung\"]","published":"2025-09-25T06:32:12Z","proceeding":"eess.AS","tasks":"[\"eess.AS\",\"cs.SD\"]","methods":"[\"Transformer\",\"Large Language Model\",\"Language Model\"]","project_urls":"[\"https://mm.kaist.ac.kr/projects/SPADE/\"]","has_code":false,"code_links":[{"ID":609306,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_id":2865793,"paper_url":"https://arxiv.org/abs/2509.20802","paper_title":"SPADE: Structured Pruning and Adaptive Distillation for Efficient LLM-TTS","repo_url":"https://github.com/nerfies/nerfies.github.io","is_official":false,"mentioned_in_paper":false,"mentioned_in_github":true,"github_stars":0}]}