code/slurm_logs/eval_v7_grid.sbatch

#!/bin/bash
#SBATCH --job-name=vl
#SBATCH --partition=gpu-large
#SBATCH --qos=batch-long
#SBATCH --gres=gpu:h200:1
#SBATCH --cpus-per-task=8
#SBATCH --mem=180G
#SBATCH --time=03:00:00
#SBATCH --output=/weka/s225250685/mats-tist/slurm_logs/eval_v7_grid_%j.out

set -u
cd /weka/s225250685/mats-tist
set -a; source /weka/s225250685/mats-tist/.env; set +a
export HF_HOME=/weka/s225250685/Huggingface HF_HUB_CACHE=/weka/s225250685/Huggingface/hub
export PYTHONNOUSERSITE=1 NO_PROXY=localhost,127.0.0.1 no_proxy=localhost,127.0.0.1
export PYTHONPATH=/weka/s225250685/mats-tist TOKENIZERS_PARALLELISM=false
export DB_EXEC_API_DISABLE=1

PY=/weka/s225250685/conda-envs/handbook/bin/python
VLLM=/weka/s225250685/conda-envs/handbook/bin/vllm
SELECTOR_CKPT=alignment-handbook/output/selector-qwen7b-v7-dev-fb
DEV_FILE=eval_results/paper_SFT_VF_passAt8_bird_dev.jsonl

LOG=/weka/s225250685/mats-tist/slurm_logs/eval_v7_grid_${SLURM_JOB_ID}.log
kill_vllm() { pkill -9 -f "vllm serve" 2>/dev/null || true; sleep 5; }
trap kill_vllm EXIT
wait_url() { for i in {1..360}; do curl --noproxy '*' -fs "$1" >/dev/null 2>&1 && return 0; sleep 5; done; return 1; }

$VLLM serve "$SELECTOR_CKPT" \
    --served-model-name selector --port 8103 \
    --dtype bfloat16 --gpu-memory-utilization 0.85 \
    --max-model-len 4096 --enforce-eager --disable-log-requests \
    > "$LOG.serve" 2>&1 &
wait_url http://localhost:8103/v1/models || exit 1

for CONF in "0.5 1.0 0.3" "0.3 0.5 0.3" "0.5 0.5 0.5" "0.1 1.0 0.3" "0.3 1.5 0.3" "0.3 1.0 0.5" "0.0 1.0 0.3" "0.3 1.0 0.0"; do
    read M P F <<<"$CONF"
    TAG=$(echo "m${M}_p${P}_f${F}" | tr '.' '_')
    echo "[$(date)] eval w_maj=$M w_pok=$P w_fix=$F" | tee -a "$LOG"
    $PY scripts/compute_bestofn_v7_fb.py \
        "$DEV_FILE" v7grid_$TAG \
        --selector_host http://localhost:8103 --model_name selector --threads 16 \
        --w_maj $M --w_pok $P --w_fix $F \
        --out_dir eval_results 2>&1 | tee -a "$LOG.eval_$TAG"
done

echo "[$(date)] DONE" | tee -a "$LOG"
kill_vllm