Adapting Text LLMs to Speech via Multimodal Depth Up-Scaling
This addresses the challenge of efficiently adapting LLMs for speech tasks while preserving text performance, which is incremental but impactful for multimodal AI applications.
The paper tackled the problem of adapting pre-trained text LLMs to speech without degrading their original text capabilities, proposing Multimodal Depth Upscaling which inserts and trains new layers on speech data; experiments showed it achieves ASR comparable to full fine-tuning while reducing text degradation by over 75% with 60% fewer parameters.
Adapting pre-trained text Large Language Models (LLMs) into Speech Language Models (Speech LMs) via continual pretraining on speech data is promising, but often degrades the original text capabilities. We propose Multimodal Depth Upscaling, an extension of an emerging strategy in continual LLM pre-training, where new transformer layers are inserted into a frozen text LLM and only the added layers are trained on speech data. Experiments with SmolLM2-360M and SmolLM2-1.7B on 48k hours of English Automatic Speech Recognition (ASR) data show that depth up-scaling achieves ASR comparable to full fine-tuning while causing far less text degradation than both full fine-tuning and Low-Rank Adaptation (LoRA). We further show that incorporating E-Branchformer, an architecture designed for speech recognition, as the inserted layers achieves ASR that matches or surpasses full fine-tuning on the larger model while reducing text degradation by over 75% with 60% fewer trainable parameters.