Transcripts are difficult to inspect one by one. Also, some training needs to be done is long form and short form, and not knowing the video time length makes it difficult to identify if a transcript are equally representative for the model.
· Sign up or log in to comment
