UCSC-VLAA/VLAA-Thinking

[TMLR 25] SFT or RL? An Early Investigation into Training R1-Like Reasoning Large Vision-Language Models