mlsanityData to 7 Oct 2026

MamTra: A Hybrid Mamba-Transformer Backbone for Speech Synthesis

arXiv 2603.12342 ↗First posted 12 Mar 2026eess.AScited by 1

On the map: State-space sequence models (S4, Mamba) → Speech and audio SSMs

Abstract (arXiv)

Despite the remarkable quality of LLM-based text-to-speech systems, their reliance on autoregressive Transformers leads to quadratic computational complexity, which severely limits practical applications. Linear-time alternatives, notably Mamba, offer a potential remedy; however, they often sacrifice the global context essential for expressive synthesis. In this paper, we propose MamTra, an interleaved Mamba-Transformer framework designed to leverage the advantages of Mamba's efficiency and Transformers' modeling capability. We also introduce novel knowledge transfer strategies to distill insights from a pretrained Transformer into our hybrid architecture, thereby bypassing the prohibitive costs of training from scratch. Systematic experiments identify the optimal hybrid configuration, and demonstrate that MamTra reduces inference VRAM usage by up to 34% without compromising speech fidelity - even trained on only 2% of the original training dataset. Audio samples are available at https://mamtratts.github.io.

Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Placed in Speech and audio SSMs, one of 10 directions in State-space sequence models (S4, Mamba). Builds on 47 of its 51 references that are in the map. First posted on arXiv 12 Mar 2026.

Builds on

47 of 51 references.

Closest ancestors · 4

Same direction, newest first

  1. 2025EDNet: A Versatile Speech Enhancement Framework with Gating Mamba Mechanism and Phase Shift-Invariant Training19 Jun
  2. 2024Speech-Mamba: Long-Context Speech Recognition with Selective State Spaces Models27 Sep
  3. 2024Speech Slytherin: Examining the Performance and Efficiency of Mamba for Speech Separation, Recognition, and Synthesis13 Jul
  4. 2024Mamba in Speech: Towards an Alternative to Self-Attention21 May

Other references · 43

  1. 2017Attention Is All You Needcited by 59,932
  2. 2024The Llama 3 Herd of Modelscited by 14,192
  3. 2023Mamba: Linear-Time Sequence Modeling with Selective State Spacescited by 4,616
  4. 2021Efficiently Modeling Long Sequences with Structured State Spacescited by 2,424
  5. 2020Transformers are RNNs: Fast Autoregressive Transformers with Linear Attentioncited by 1,979
  6. 2023GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpointscited by 1,311
  7. 2024Transformers are SSMs: Generalized Models and Efficient Algorithms Through Structured State Space Dualitycited by 1,283
  8. 2024Qwen2.5-Coder Technical Reportcited by 1,151
  9. 2023H2O: Heavy-Hitter Oracle for Efficient Generative Inference of Large Language Modelscited by 805
  10. 2022Simplified State Space Layers for Sequence Modelingcited by 776
  11. 2025DeepSeek-V3.2: Pushing the Frontier of Open Large Language Modelscited by 755
  12. 2023Neural Codec Language Models are Zero-Shot Text to Speech Synthesizerscited by 737
31 more
  1. 2023LLM-Pruner: On the Structural Pruning of Large Language Modelscited by 643
  2. 2019LibriTTS: A Corpus Derived from LibriSpeech for Text-to-Speechcited by 639
  3. 2022AudioLM: a Language Modeling Approach to Audio Generationcited by 611
  4. 2023Retentive Network: A Successor to Transformer for Large Language Modelscited by 594
  5. 2022Hungry Hungry Hippos: Towards Language Modeling with State Space Modelscited by 495
  6. 2023Gated Linear Attention Transformers with Hardware-Efficient Trainingcited by 490
  7. 2024Gated Delta Networks: Improving Mamba2 with Delta Rulecited by 483
  8. 2023Model Tells You What to Discard: Adaptive KV Cache Compression for LLMscited by 466
  9. 2024Jamba: A Hybrid Transformer-Mamba Language Modelcited by 390
  10. 2024F5-TTS: A Fairytaler that Fakes Fluent and Faithful Speech with Flow Matchingcited by 353
  11. 2024CosyVoice: A Scalable Multilingual Zero-shot Text-to-speech Synthesizer based on Supervised Semantic Tokenscited by 323
  12. 2024Seed-TTS: A Family of High-Quality Versatile Speech Generation Modelscited by 293
  13. 2024CosyVoice 2: Scalable Streaming Speech Synthesis with Large Language Modelscited by 282
  14. 2024Griffin: Mixing Gated Linear Recurrences with Local Attention for Efficient Language Modelscited by 232
  15. 2024Repeat After Me: Transformers are Better than State Space Models at Copyingcited by 208
  16. 2024Samba: Simple Hybrid State Space Models for Efficient Unlimited Context Language Modelingcited by 167
  17. 2024The Mamba in the Llama: Distilling and Accelerating Hybrid Modelscited by 120
  18. 2024Can Mamba Learn How to Learn? A Comparative Study on In-Context Learning Taskscited by 104
  19. 2024Hymba: A Hybrid-head Architecture for Small Language Modelscited by 95
  20. 2025Llasa: Scaling Train-Time and Inference-Time Compute for Llama-based Speech Synthesiscited by 69
  21. 2024Transformers to SSMs: Distilling Quadratic Knowledge to Subquadratic Modelscited by 66
  22. 2024Is Mamba Capable of In-Context Learning?cited by 45
  23. 2025LongMamba: Enhancing Mamba's Long Context Capabilities via Training-Free Receptive Field Enlargementcited by 36
  24. 2025Llamba: Scaling Distilled Recurrent Models for Efficient Language Processingcited by 35
  25. 2024VERSA: A Versatile Evaluation Toolkit for Speech, Audio, and Musiccited by 30
  26. 2024Long-Form Speech Generation with Spoken Language Modelscited by 19
  27. 2024Fast and High-Quality Auto-Regressive Speech Synthesis via Speculative Decodingcited by 17
  28. 2025Hybrid Architectures for Language Models: Systematic Analysis and Design Insightscited by 13
  29. 2024Accelerating Codec-based Speech Synthesis with Multi-Token Prediction and Speculative Decodingcited by 10
  30. 2025DiSTAR: Diffusion over a Scalable Token Autoregressive Representation for Speech Generationcited by 3
  31. 2025SPADE: Structured Pruning and Adaptive Distillation for Efficient LLM-TTScited by 2

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026WAND: Windowed Attention and Knowledge Distillation for Efficient Autoregressive Text-to-Speech Modelscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025TransMamba: A Sequence-Level Hybrid Transformer-Mamba Language Model
  2. 2026MambaVoiceCloning: Efficient and Expressive Text-to-Speech via State-Space Modeling and Diffusion Control
  3. 2025M3-TTS: Multi-modal DiT Alignment & Mel-latent for Zero-shot High-fidelity Speech Synthesis
  4. 2026MambAdapter: Lightweight Mamba-Based Adapters for Parameter-Efficient Transfer Learning in Speech and Audio
  5. 2025SAM: A Mamba-2 State-Space Audio-Language Model
  6. 2024Autoregressive Diffusion Transformer for Text-to-Speech Synthesis
  7. 2026DLLM-TTS: Block Discrete Diffusion Language Model for Text-to-Speech Synthesis
  8. 2025Multimodal Mamba: Decoder-only Multimodal State Space Model via Quadratic to Linear Distillation