MMR1 Enhancing Multimodal Reasoning with Variance-Aware Sampling and Open Resources

→ View original post on X — @_akhaliq