newtype-2038/Qwen3.5-122B-A10B-Newtype-Tuned-v21-grpo
014
docs: rebrand banya -> newtype
v21 GRPO LoRA (v10 (Masked SFT) init + dense reward)
v21 GRPO LoRA (v10 (Masked SFT) init + dense reward)
initial commit
docs: rebrand banya -> newtype
v21 GRPO LoRA (v10 (Masked SFT) init + dense reward)
v21 GRPO LoRA (v10 (Masked SFT) init + dense reward)
initial commit