Open christophschuhmann opened 2 years ago
Pick a dataset you would like to convert to our standard training format. Format specs see here:
https://docs.google.com/document/d/1ArSsmV9SXKOkZKGc8Bmaja-83yw4R6nZupW7iEB9yGw/edit
If you post a new dataset project, put a prefix into the name that specifies the modalities like "IMAGE-TEXT", "AUDIO-TEXT", "VQA (IMAGE-TEXT)", "VIDEO-AUDIO", ...
Pick a dataset you would like to convert to our standard training format. Format specs see here:
https://docs.google.com/document/d/1ArSsmV9SXKOkZKGc8Bmaja-83yw4R6nZupW7iEB9yGw/edit
If you post a new dataset project, put a prefix into the name that specifies the modalities like "IMAGE-TEXT", "AUDIO-TEXT", "VQA (IMAGE-TEXT)", "VIDEO-AUDIO", ...