Dataset Groups Activity Stream Microsoft Research Video Description Corpus (MSVD) The MSVD dataset is a collection of 1970 open domain clips from YouTube, annotated with variable-length captions. BibTex: @dataset{Alvaro_Peris_and_Marc_Bola˜nos_and_Petia_Radeva_and_Francisco_Casacuberta_2025, abstract = {The MSVD dataset is a collection of 1970 open domain clips from YouTube, annotated with variable-length captions.}, author = {Alvaro Peris and Marc Bola˜nos and Petia Radeva and Francisco Casacuberta}, doi = {10.57702/zwcosfzm}, institution = {No Organization}, keyword = {'Captioning', 'Natural Language Processing', 'Video Description'}, month = {jan}, publisher = {TIB}, title = {Microsoft Research Video Description Corpus (MSVD)}, url = {https://service.tib.eu/ldmservice/dataset/microsoft-research-video-description-corpus--msvd-}, year = {2025} }
