olmo1b_grpo_sweep.sh

for seed in 1 2; do
for lr in 5e-7 7e-7 9e-7; do
python update_command_args.py scripts/train/olmo2/grpo_7b.sh \
    --priority urgent \
    --workspace ai2/olmo-instruct \
    --exp_name 0423_grpo_seed_${seed}_lr_${lr} \
    --model_name_or_path allenai/OLMo-2-0425-1B-DPO \
    --model_revision main \
    --tokenizer_name_or_path allenai/OLMo-2-1124-7B-DPO \
    --save_freq 200 \
    --tokenizer_revision main \
    --image nathanl/open_instruct_auto | uv run bash
done
done


for seed in 1 2; do
for lr in 5e-7 7e-7 9e-7; do
python update_command_args.py scripts/train/olmo2/grpo_7b.sh \
    --cluster ai2/augusta-google-1 \
    --priority urgent \
    --workspace ai2/olmo-instruct \
    --exp_name 0427_grpo_seed_${seed}_lr_${lr} \
    --dataset_mixer_list allenai/RLVR-MATH 1.0 \
    --dataset_mixer_list_splits train \
    --dataset_mixer_eval_list allenai/RLVR-MATH 16 \
    --dataset_mixer_eval_list_splits train \
    --model_name_or_path allenai/OLMo-2-0425-1B-RLVR1 \
    --model_revision main \
    --tokenizer_name_or_path allenai/OLMo-2-1124-7B-DPO \
    --tokenizer_revision main \
    --save_freq 200 \
    --image nathanl/open_instruct_auto
done
done