USM-Lite: Quantization and Sparsity Aware Fine-tuning for Speech Recognition with Universal Speech Models
End-to-end automatic speech recognition (ASR) models have seen revolutionary quality gains with the recent development of large-scale universal speech models (USM). However, deploying these massive USMs is extremely expensive due to the enormous memory usage and computational cost. Therefore, model compression is an important research topic to fit USM-based ASR under budget in real-world scenarios. In this study, we propose a USM fine-tuning approach for ASR, with a low-bit quantization and N:M structured sparsity aware paradigm on the model weights, reducing the model complexity from parameter precision and matrix topology perspectives. We conducted extensive experiments with a 2-billion parameter USM on a large-scale voice search dataset to evaluate our proposed method. A series of ablation studies validate the effectiveness of up to int4 quantization and 2:4 sparsity. However, a single compression technique fails to recover the performance well under extreme setups including int2 quantization and 1:4 sparsity. By contrast, our proposed method can compress the model to have 9.4% of the size, at the cost of only 7.3% relative word error rate (WER) regressions. We also provided in-depth analyses on the results and discussions on the limitations and potential solutions, which would be valuable for future studies.
Code (0)
등록된 구현이 없습니다.
Tasks
Automatic Speech RecognitionAutomatic Speech Recognition (ASR)Model CompressionQuantizationspeech-recognitionSpeech RecognitionMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
OWQ: Outlier-Aware Weight Quantization for Efficient Fine-Tuning and Inference of Large Language Models
Large language models (LLMs) with hundreds of billions of parameters require powerful server-grade GPUs for inference, limiting their practical deployment. To address this challenge, we introduce the outlier-aware weight…
parameter-efficient fine-tuningQuantizationTorchAO: PyTorch-Native Training-to-Serving Model Optimization
We present TorchAO, a PyTorch-native model optimization framework leveraging quantization and sparsity to provide an end-to-end, training-to-serving workflow for AI models. TorchAO supports a variety of popular model opt…
Pioneering 4-Bit FP Quantization for Diffusion Models: Mixup-Sign Quantization and Timestep-Aware Fine-Tuning
Model quantization reduces the bit-width of weights and activations, improving memory efficiency and inference speed in diffusion models. However, achieving 4-bit quantization remains challenging. Existing methods, p…
DenoisingQuantizationZeroth-Order Fine-Tuning of LLMs with Extreme Sparsity
Zeroth-order optimization (ZO) is a memory-efficient strategy for fine-tuning Large Language Models using only forward passes. However, the application of ZO fine-tuning in memory-constrained settings such as mobile phon…
GPUQuantizationLoftQ: LoRA-Fine-Tuning-Aware Quantization for Large Language Models
Quantization is an indispensable technique for serving Large Language Models (LLMs) and has recently found its way into LoRA fine-tuning. In this work we focus on the scenario where quantization and LoRA fine-tuning are …
Natural Language UnderstandingQuantizationQuestion AnsweringText Generation