# Speech-FT: Merging Pre-trained And Fine-Tuned Speech Representation Models For Cross-Task Generalization

arXiv: https://arxiv.org/abs/2502.12672 · HTML: https://mlsanity.com/paper/2502.12672/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 57 of its 77 references that are in the map.
- First posted on arXiv 18 Feb 2025.

## Builds on (most cited first, 30 of 57)

- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [Representation Learning with Contrastive Predictive Coding](https://mlsanity.com/paper/1807.03748/) (2018)
- [Overcoming catastrophic forgetting in neural networks](https://mlsanity.com/paper/1612.00796/) (2016)
- [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://mlsanity.com/paper/2006.11477/) (2020)
- [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://mlsanity.com/paper/2106.07447/) (2021)
- [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://mlsanity.com/paper/2110.13900/) (2021)
- [Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time](https://mlsanity.com/paper/2203.05482/) (2022)
- [Common Voice: A Massively-Multilingual Speech Corpus](https://mlsanity.com/paper/1912.06670/) (2019)
- [Editing Models with Task Arithmetic](https://mlsanity.com/paper/2212.04089/) (2022)
- [TIES-Merging: Resolving Interference When Merging Models](https://mlsanity.com/paper/2306.01708/) (2023)
- [Robust fine-tuning of zero-shot models](https://mlsanity.com/paper/2109.01903/) (2021)
- [DoRA: Weight-Decomposed Low-Rank Adaptation](https://mlsanity.com/paper/2402.09353/) (2024)
- [Fine-Tuning can Distort Pretrained Features and Underperform Out-of-Distribution](https://mlsanity.com/paper/2202.10054/) (2022)
- [SUPERB: Speech processing Universal PERformance Benchmark](https://mlsanity.com/paper/2105.01051/) (2021)
- [MLS: A Large-Scale Multilingual Dataset for Speech Research](https://mlsanity.com/paper/2012.03411/) (2020)
- [Layer-wise Analysis of a Self-supervised Speech Representation Model](https://mlsanity.com/paper/2107.04734/) (2021)
- [Self-Supervised Speech Representation Learning: A Review](https://mlsanity.com/paper/2205.10643/) (2022)
- [An Unsupervised Autoregressive Model for Speech Representation Learning](https://mlsanity.com/paper/1904.03240/) (2019)
- [Mitigating the Alignment Tax of RLHF](https://mlsanity.com/paper/2309.06256/) (2023)
- [Recall and Learn: Fine-tuning Deep Pretrained Language Models with Less Forgetting](https://mlsanity.com/paper/2004.12651/) (2020)
- [Self-Training for End-to-End Speech Recognition](https://mlsanity.com/paper/1909.09116/) (2019)
- [A Fine-tuned Wav2vec 2.0/HuBERT Benchmark For Speech Emotion Recognition, Speaker Verification and Spoken Language Understanding](https://mlsanity.com/paper/2111.02735/) (2021)
- [DeCoAR 2.0: Deep Contextualized Acoustic Representations with Vector Quantization](https://mlsanity.com/paper/2012.06659/) (2020)
- [SUPERB-SG: Enhanced Speech processing Universal PERformance Benchmark for Semantic and Generative Capabilities](https://mlsanity.com/paper/2203.06849/) (2022)
- [ML-SUPERB: Multilingual Speech Universal PERformance Benchmark](https://mlsanity.com/paper/2305.10615/) (2023)
- [MagMax: Leveraging Model Merging for Seamless Continual Learning](https://mlsanity.com/paper/2407.06322/) (2024)
- [MetaGPT: Merging Large Language Models Using Model Exclusive Task Arithmetic](https://mlsanity.com/paper/2406.11385/) (2024)
- [Going Beyond Linear Mode Connectivity: The Layerwise Linear Feature Connectivity](https://mlsanity.com/paper/2307.08286/) (2023)
- [Exploring Wav2vec 2.0 fine-tuning for improved speech emotion recognition](https://mlsanity.com/paper/2110.06309/) (2021)
- [A Large-Scale Evaluation of Speech Foundation Models](https://mlsanity.com/paper/2404.09385/) (2024)

## Built on by (most cited first, 2 of 2)

- [A correlation-permutation approach for speech-music encoders model merging](https://mlsanity.com/paper/2506.11403/) (2025)
- [Pseudo2Real: Task Arithmetic for Pseudo-Label Correction in Automatic Speech Recognition](https://mlsanity.com/paper/2510.08047/) (2025)

## Abstract (arXiv)

Fine-tuning speech representation models can enhance performance on specific tasks but often compromises their cross-task generalization ability. This degradation is often caused by excessive changes in the representations, making it difficult to retain information learned during pre-training. Existing approaches, such as regularizing weight changes during fine-tuning, may fail to maintain sufficiently high feature similarity with the pre-trained model, and thus could possibly lose cross-task generalization. To address this issue, we propose Speech-FT, a novel two-stage fine-tuning framework designed to maintain cross-task generalization while benefiting from fine-tuning. Speech-FT first applies fine-tuning specifically designed to reduce representational drift, followed by weight-space interpolation with the pre-trained model to restore cross-task generalization. Extensive experiments on HuBERT, wav2vec 2.0, DeCoAR 2.0, and WavLM Base+ demonstrate that Speech-FT consistently improves performance across a wide range of supervised, unsupervised, and multitask fine-tuning scenarios. Moreover, Speech-FT achieves superior cross-task generalization compared to fine-tuning baselines that explicitly constrain weight changes, such as weight-space regularization and LoRA fine-tuning. Our analysis reveals that Speech-FT maintains higher feature similarity to the pre-trained model compared to alternative strategies, despite allowing larger weight-space updates. Notably, Speech-FT achieves significant improvements on the SUPERB benchmark. For example, when fine-tuning HuBERT on automatic speech recognition, Speech-FT is able to reduce phone error rate from 5.17% to 3.94%, lower word error rate from 6.38% to 5.75%, and increase speaker identification accuracy from 81.86% to 84.11%. Speech-FT provides a simple yet powerful solution for further refining speech representation models after pre-training.