Has the model undergone the RL process or just SFT on R1-0558 reasoning trajectory?

#3
by kevinpro - opened

Amazing work and impressive performance!

I am interested in the training process, has the model undergone the RL process or just SFT on R1-0558 reasoning trajectory?

NVIDIA org

Thanks! Just SFT. You can find full details on how we trained it in the documentation here https://nvidia.github.io/NeMo-Skills/releases/openreasoning/

kevinpro changed discussion status to closed

Sign up or log in to comment