Open DHUAVY opened 1 year ago
Hi, sorry for the late reply. You need to construct the pkl file based on the annotation data of MSRVTT-1kA dataset on your own. The pkl file format is a list of dictionaries [{}, {}, {}], where each dictionary represents the annotation information for a video. It includes two keys, 'filename' and 'text'. 'filename' is the real path of the video, and 'text' is the corresponding text. Hope this can help you :)
Hi, I want to train the model using the MSR-VTT dataset. And it tells me that I need a pkl file but I can only find the mp4 and txt files. So how can I tranfer them to or maybe to find the pkl file.