GigaChat Audio：时间感知的大规模音频语言模型

摘要

时序定位在长录音中对于音频条件的大语言模型而言仍是一项挑战。本文提出了一种具备时间感知能力的音频大语言模型，能够在长达120分钟的输入中，以显式时间戳回答用户问题。我们的方法通过级联流水线生成的大规模合成监督数据，将周期性时间标记与连续音频标记交织在一起。该模型在短时与长时基准测试中均展现出强大的时序定位精度，并支持时间锚定的片段描述与摘要功能。通过大量消融实验，我们深入分析了时间表示方式、标记频率、分词策略以及任务时长混合设计对模型准确性与计算开销的影响。为促进时间感知音频理解的进一步研究，我们已在 https://huggingface.co/ai-sage/GigaChat3.1-Audio-10B-A1.8B 开源模型权重与数据集。

English

Temporal grounding in long recordings remains challenging for audio-conditioned LLMs. We present a time-aware audio LLM that answers questions with explicit timestamps over up to 120 minutes of input. Our approach interleaves periodic time markers with continuous audio tokens using large-scale synthetic supervision from a cascaded pipeline. Our model achieves strong temporal-grounding accuracy on short and long benchmarks and supports time-anchored fragment descriptions and summaries. Extensive ablations examine how time representation, marker frequency, tokenization, and duration-mixture design affect accuracy and computational cost. We release model weights and datasets to support further research on time-aware audio understanding, available at https://huggingface.co/ai-sage/GigaChat3.1-Audio-10B-A1.8B.