search

OFA-Sys / AIR-Bench

AIR-Bench: Benchmarking Large Audio-Language Models via Generative Comprehension
46 stars 1 forks source link

数据集 #4

Open Jasper-sudo-Sun opened 1 week ago

Jasper-sudo-Sun commented 1 week ago

你好，我在复现qwen2-audio关于air-bench的结果，想问下Speech 和 mixed audio为什么有好多的数据集只有一个结果啊

qyang1021 commented 1 week ago

最终的结果是每个数据源（例如Speech｜Mixed Audio）的平均值；你可以参考cal_score脚本