Saved in:
Bibliographic Details
Main Authors: Liao, Yujie, Geng, Xuelong, Xue, Hongfei, Wang, Shuiyuan, Xie, Lei
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.10862
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910049373257728
author Liao, Yujie
Geng, Xuelong
Xue, Hongfei
Wang, Shuiyuan
Xie, Lei
author_facet Liao, Yujie
Geng, Xuelong
Xue, Hongfei
Wang, Shuiyuan
Xie, Lei
contents Recent advancements in Speech Large Language Models have significantly enhanced multi-dimensional speech understanding. However, the majority of high-performance frameworks are predominantly optimized for GPU centric ecosystems and proprietary backbones, creating a significant gap for deployment on non-CUDA computing infrastructures. In this paper, we present OSUM-Pangu, a fully open-source speech understanding foundation model developed on a completely non-CUDA software and hardware stack. By integrating an audio encoder with the openPangu-7B LLM backbone, we successfully implement the entire training and inference pipeline on the Ascend NPU platform. To facilitate efficient task alignment under non-CUDA resource constraints, we adopt a practical training process that sequentially bridges speech perception and user intent recognition. Experimental results demonstrate that OSUM-Pangu achieves task accuracy comparable to mainstream GPU-based models while maintaining robust natural language interaction capabilities. Our work provides a reproducible, non-CUDA baseline for the open-source speech community, promoting the independent evolution of multimodal intelligence.
format Preprint
id arxiv_https___arxiv_org_abs_2603_10862
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle OSUM-Pangu: An Open-Source Multidimension Speech Understanding Foundation Model Built upon OpenPangu on Ascend NPUs
Liao, Yujie
Geng, Xuelong
Xue, Hongfei
Wang, Shuiyuan
Xie, Lei
Sound
Recent advancements in Speech Large Language Models have significantly enhanced multi-dimensional speech understanding. However, the majority of high-performance frameworks are predominantly optimized for GPU centric ecosystems and proprietary backbones, creating a significant gap for deployment on non-CUDA computing infrastructures. In this paper, we present OSUM-Pangu, a fully open-source speech understanding foundation model developed on a completely non-CUDA software and hardware stack. By integrating an audio encoder with the openPangu-7B LLM backbone, we successfully implement the entire training and inference pipeline on the Ascend NPU platform. To facilitate efficient task alignment under non-CUDA resource constraints, we adopt a practical training process that sequentially bridges speech perception and user intent recognition. Experimental results demonstrate that OSUM-Pangu achieves task accuracy comparable to mainstream GPU-based models while maintaining robust natural language interaction capabilities. Our work provides a reproducible, non-CUDA baseline for the open-source speech community, promoting the independent evolution of multimodal intelligence.
title OSUM-Pangu: An Open-Source Multidimension Speech Understanding Foundation Model Built upon OpenPangu on Ascend NPUs
topic Sound
url https://arxiv.org/abs/2603.10862