diff options
| author | YurenHao0426 <blackhao0426@gmail.com> | 2026-01-27 09:57:37 -0600 |
|---|---|---|
| committer | YurenHao0426 <blackhao0426@gmail.com> | 2026-01-27 09:57:37 -0600 |
| commit | dc801c07cf38b0c495686463e6ca6f871a64440e (patch) | |
| tree | 599f03114775921dbc472403c701f4a3a8ea188a /collaborativeagents/slurm/run_experiments_collab_baselines.sh | |
| parent | e43b3f8aa36c198b95c1e46bea2eaf3893b13dc3 (diff) | |
Add collaborativeagents module and update gitignore
- Add collaborativeagents subproject with adapters, agents, and evaluation modules
- Update .gitignore to exclude large binary files (.whl, .tar), wandb logs, and results
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Diffstat (limited to 'collaborativeagents/slurm/run_experiments_collab_baselines.sh')
| -rwxr-xr-x | collaborativeagents/slurm/run_experiments_collab_baselines.sh | 59 |
1 files changed, 59 insertions, 0 deletions
diff --git a/collaborativeagents/slurm/run_experiments_collab_baselines.sh b/collaborativeagents/slurm/run_experiments_collab_baselines.sh new file mode 100755 index 0000000..bbf53fc --- /dev/null +++ b/collaborativeagents/slurm/run_experiments_collab_baselines.sh @@ -0,0 +1,59 @@ +#!/bin/bash +#SBATCH --job-name=run_collab_baselines +#SBATCH --account=bfqt-delta-gpu +#SBATCH --partition=gpuA100x4 +#SBATCH --nodes=1 +#SBATCH --ntasks=1 +#SBATCH --cpus-per-task=16 +#SBATCH --gres=gpu:nvidia_a100:4 +#SBATCH --mem=200G +#SBATCH --time=48:00:00 +#SBATCH --output=logs/run_collab_baselines_%j.out +#SBATCH --error=logs/run_collab_baselines_%j.err + +# Run CollaborativeAgents baselines on 4x A100 80GB +# - contextual: Full history in context (summarize on overflow) +# - reflection: CollaborativeAgents' agent_notes approach +# - reflection_grpo: Reflection + GRPO training (with_proper_scaffolding) + +set -e + +cd /projects/bfqt/users/yurenh2/ml-projects/personalization-user-model + +# Create logs and results directories +mkdir -p collaborativeagents/slurm/logs +mkdir -p collaborativeagents/results + +echo "Starting CollaborativeAgents baselines at $(date)" +echo "Job ID: $SLURM_JOB_ID" +echo "Node: $SLURMD_NODENAME" +echo "GPUs: $CUDA_VISIBLE_DEVICES" + +# Activate environment +source /u/yurenh2/miniforge3/etc/profile.d/conda.sh +conda activate eval + +# Check GPU availability +nvidia-smi + +# Redirect HF cache to project space (avoid home quota issues) +export HF_HOME=/projects/bfqt/users/yurenh2/.cache/huggingface +mkdir -p $HF_HOME + +# Add project to path +export PYTHONPATH="${PWD}/src:${PWD}/collaborativeagents:${PYTHONPATH}" + +# Run experiments +cd collaborativeagents/scripts + +# Run the 3 CollaborativeAgents baselines +echo "Running contextual, reflection, reflection_grpo baselines..." +python run_experiments.py \ + --methods contextual,reflection,reflection_grpo \ + --datasets math-500 \ + --n-profiles 20 \ + --n-sessions 5 \ + --profile-path ../data/complex_profiles_v2/profiles_100.jsonl \ + --output-dir ../results/collab_baselines_$(date +%Y%m%d_%H%M%S) + +echo "CollaborativeAgents baselines completed at $(date)" |
