Closed duxiaodan closed 3 months ago
Hi. It is from the original WebVid (DynamiCrafter). The DCP one is just a proof-of-concept.
Thank you for the response. So just to be sure: if I want to reproduce your results I don't need to get the annotations you provide on HuggingFace, right?
exactly.
Awesome. Thank you for the quick response.
Thank you for the amazing work and the code base. I want to continue training from your ckpts but I'm not sure how are the current ckpts trained. Are they trained on the original WebVid (DynamiCrafter) or your filtered+labeled WebVid (DynamiCrafter_DCP)? Thank you in advance