Benchmark dataset for abstractive multi-video captioning. - View it on GitHub