YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
Gemma-3-4B-it 路 TTRL 路 OpenR1 路 mmupt(Gemma variant)
Recipe: mmupt, Gemma variant (beta 0.01, K 10, T 1.0, cap 1024, lr 1e-6, warmup 0, weight_decay 0.01, max_grad_norm 1.0, bnpo, scale_rewards group, 12 prompts/step = EB 120), same as big-tier Gemma-12B column. TTRL = self-labeling majority vote. best/ = best-by-val (MathVista-150) step 220; endpoint/ = checkpoint-640 (1 epoch); training/ = train.log, trainer_state, best_metric.
Eval protocol for the paper tables: v2 (T=0, 16k, boxed prompt, rule + Qwen2.5-32B judge), endpoint selection. Local source: /weka/scratch/jhu/dssg2026-ext-rghani1/yyang331/mllm-repro-out/mllm-co-grpo-dp/openr1_gemma3_4b_ttrl_mmuptG Trained 2026-09-01/02 on JHU a100.