We introduce DeSTA2.5-Audio, a general-purpose Large Audio Language Model (LALM) designed for robust auditory perception and instruction-following, without requiring task-specific audio instruction-tuning. Recent LALMs typically augment Large Language Models (LLMs) with auditory capabilities by training on large-scale, manually curated or LLM-synthesized audio-instruction datasets. However, these approaches have often suffered from the catastrophic forgetting of the LLM's original language abilities. To address this, we revisit the data construction pipeline and propose DeSTA, a self-generated cross-modal alignment strategy in which the backbone LLM generates its own training targets. This approach preserves the LLM's native language proficiency while establishing effective audio-text alignment, thereby enabling zero-shot generalization without task-specific tuning. Using DeSTA, we construct DeSTA-AQA5M, a large-scale, task-agnostic dataset containing 5 million training samples derived from 7,000 hours of audio spanning 50 diverse datasets, including speech, environmental sounds, and music. DeSTA2.5-Audio achieves state-of-the-art or competitive performance across a wide range of audio-language benchmarks, including Dynamic-SUPERB, MMAU, SAKURA, Speech-IFEval, and VoiceBench. Comprehensive comparative studies demonstrate that our self-generated strategy outperforms widely adopted data construction and training strategies in both auditory perception and instruction-following capabilities. Our findings underscore the importance of carefully designed data construction in LALM development and offer practical insights for building robust, general-purpose LALMs.
View on arXiv@article{lu2025_2507.02768,
title={ DeSTA2.5-Audio: Toward General-Purpose Large Audio Language Model with Self-Generated Cross-Modal Alignment },
author={ Ke-Han Lu and Zhehuai Chen and Szu-Wei Fu and Chao-Han Huck Yang and Sung-Feng Huang and Chih-Kai Yang and Chee-En Yu and Chun-Wei Chen and Wei-Chih Chen and Chien-yu Huang and Yi-Cheng Lin and Yu-Xiang Lin and Chi-An Fu and Chun-Yi Kuan and Wenze Ren and Xuanjun Chen and Wei-Ping Huang and En-Pei Hu and Tzu-Quan Lin and Yuan-Kuei Wu and Kuan-Po Huang and Hsiao-Ying Huang and Huang-Cheng Chou and Kai-Wei Chang and Cheng-Han Chiang and Boris Ginsburg and Yu-Chiang Frank Wang and Hung-yi Lee },
journal={arXiv preprint arXiv:2507.02768},
year={ 2025 }
}