Closed shipengai closed 10 months ago
As paper says,“Instead, it leverages a more compact 0.5M pre-training caption data following llava”。 It means that “In the first pretrain stage, the train dataset is only blip_laion_cc_sbu_558k.json?”
Yes, this is amount of data used for initial alignment.
As paper says,“Instead, it leverages a more compact 0.5M pre-training caption data following llava”。 It means that “In the first pretrain stage, the train dataset is only blip_laion_cc_sbu_558k.json?”