paper-with-me

홈 › Papers

USM-Lite: Quantization and Sparsity Aware Fine-tuning for Speech Recognition with Universal Speech Models

2023-12-13 · Shaojin Ding, David Qiu, David Rim, Yanzhang He, Oleg Rybakov, Bo Li, Rohit Prabhavalkar, Weiran Wang, Tara N. Sainath, Zhonglin Han, Jian Li, Amir Yazdanbakhsh, Shivani Agrawal

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.

📄 PDF Abstract BibTeX arXiv:2312.08553

Code (0)

등록된 구현이 없습니다.

Tasks

Automatic Speech RecognitionAutomatic Speech Recognition (ASR)Model CompressionQuantizationspeech-recognitionSpeech Recognition

Methods 이 논문이 사용한 방법론

AWARE We propose to theoretically and empirically examine the effect of incorporating weighting schemes into walk-aggregating GNNs. To this end, we propose a simple, interpretable, and…

Similar Papers 제목 키워드 기반

OWQ: Outlier-Aware Weight Quantization for Efficient Fine-Tuning and Inference of Large Language Models

2023-06-04 · Changhun Lee, Jungyu Jin, Taesu Kim, HyungJun Kim 외

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-tuningQuantization

TorchAO: PyTorch-Native Training-to-Serving Model Optimization

2025-07-21 · Andrew Or, Apurva Jain, Daniel Vega-Myhre, Jesse Cai 외 arxiv

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

2025-01-01 · CVPR 2025 1 · Maosen Zhao, Pengtao Chen, Chong Yu, Yan Wen 외

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…

DenoisingQuantization

Zeroth-Order Fine-Tuning of LLMs with Extreme Sparsity

2024-06-05 · Wentao Guo, Jikai Long, Yimeng Zeng, Zirui Liu 외

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…

GPUQuantization

LoftQ: LoRA-Fine-Tuning-Aware Quantization for Large Language Models

2023-10-12 · Yixiao Li, Yifan Yu, Chen Liang, Pengcheng He 외

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