run_rollouts_eval_3_3_5_multiseed.sh

seeds=(247 378 478 589 756)
for seed in ${seeds[@]}
do
    python dialogue-progression/rollouts_eval.py \
        --n-candidates=3 \
        --n-rollouts-per-candidate=3 \
        --n-utterances-per-rollout=5 \
        --outputpath=rollouts_self_play/3_3_5/$seed \
        --agent-modelpath=LACAI/DialoGPT-large-PFG \
        --progression-modelpath=models/progression/persuasion_for_good/supervised_large_adapted \
        --model-random-state=$seed \
        --resume
done