Will we open-source the training data?
#1
by
Kaguya-19
- opened
Hello! As a new team, achieving such surprising results on challenging tasks with a small parameter model in just one month， MiroThinker v0.2 is truly remarkable. Will we, as we did with MiroThinker v0.1, open-source the corresponding training dataset? We greatly appreciate your contributions to the open-source community and academic research.