docs: simplify end-to-end diagram labels

docs: add end-to-end architecture diagrams

Update model card: trained + artifacts

Training plots

GRPO LoRA (sft=1000, grpo=200)

GRPO LoRA (sft=200, grpo=80)

Training plots

GRPO LoRA (sft=30, grpo=8)

Initial placeholder model card

initial commit