Model card: expand training data section

Add syncreward-stage1 (Stage-1 PE-AV encoder, fp32 safetensors)

Add syncreward-stage2 (Stage-2 reward model, fp32 safetensors)

Add model card and Apache-2.0 license

initial commit