sora-reasoner/thinking-with-video-pretrain

add .gitignore

Add generated WikiQA video dataset

Update GSM8K video dataset with page-turning feature (7473 samples, 1.8GB)

the whole pretrain

what can i say

initial commit