Thanks for the great work. An attempt was made to reproduce the training dataset based on the provided Markdown files. However, after filtering, only 80 usable samples were obtained, less than 156 video pairs. To facilitate a fair comparison and improve reproducibility, it would be helpful if the authors could also release the training dataset.
Thanks for the great work. An attempt was made to reproduce the training dataset based on the provided Markdown files. However, after filtering, only 80 usable samples were obtained, less than 156 video pairs. To facilitate a fair comparison and improve reproducibility, it would be helpful if the authors could also release the training dataset.