issai/WavCaps_Kazakh_Russian

该数据集是WavCaps测试集的哈萨克语和俄语机器翻译版本，来自AudioBench基准测试套件，用于评估音频语言模型对环境声音和音频场景的音频字幕生成能力。原始WavCaps基准测试旨在探究音频语言模型是否能对非语音音频内容（如环境声音、音乐、动物叫声、机械噪音和背景场景）生成自由形式的自然语言描述。此版本将文本指令和参考字幕本地化为哈萨克语和俄语，使得可以在相同的音频刺激下评估多语言音频语言模型。音频内容未变，只有文本字段（instruction, answer）进行了机器翻译。

This dataset is a machine-translated Kazakh and Russian adaptation of the WavCaps test set from the AudioBench benchmark suite, designed for evaluating audio captioning over general environmental sounds and audio scenes in Audio-Language Models. The original WavCaps benchmark probes whether audio-language models can produce free-form natural-language descriptions of non-speech audio content — including environmental sounds, music, animal vocalizations, mechanical noises, and ambient scenes. This release localizes the textual instructions and reference captions into Kazakh and Russian, enabling evaluation of multilingual audio-language models on the same audio stimuli. Audio is unchanged. Only the textual fields (`instruction`, `answer`) were machine-translated.