1 files changed, 78 insertions, 0 deletions
diff --git a/collaborativeagents/scripts/quick_rag_debug.sbatch b/collaborativeagents/scripts/quick_rag_debug.sbatch
new file mode 100644
index 0000000..efc4c31
--- /dev/null
+++ b/collaborativeagents/scripts/quick_rag_debug.sbatch
@@ -0,0 +1,78 @@
+#!/bin/bash
+#SBATCH --job-name=rag_debug
+#SBATCH --account=bfqt-delta-gpu
+#SBATCH --partition=gpuH200x8
+#SBATCH --nodes=1
+#SBATCH --ntasks=1
+#SBATCH --cpus-per-task=16
+#SBATCH --gres=gpu:4
+#SBATCH --mem=200G
+#SBATCH --time=00:40:00
+#SBATCH --output=/projects/bfqt/users/yurenh2/ml-projects/personalization-user-model/rag_debug-%j.out
+#SBATCH --error=/projects/bfqt/users/yurenh2/ml-projects/personalization-user-model/rag_debug-%j.err
+
+# Quick debug test: 2 profiles, 5 sessions - should complete in ~20 min
+
+cd /projects/bfqt/users/yurenh2/ml-projects/personalization-user-model/collaborativeagents
+source /u/yurenh2/miniforge3/etc/profile.d/conda.sh
+conda activate eval
+export HF_HOME=/projects/bfqt/users/yurenh2/hf_cache/huggingface
+export PYTHONPATH="/projects/bfqt/users/yurenh2/ml-projects/personalization-user-model/src:$PYTHONPATH"
+
+PROFILE_PATH="/projects/bfqt/users/yurenh2/ml-projects/personalization-user-model/collaborativeagents/data/complex_profiles_v2/profiles_200.jsonl"
+AGENT_MODEL="/projects/bfqt/users/yurenh2/ml-projects/personalization-user-model/models/llama-3.1-8b-instruct"
+USER_MODEL="meta-llama/Llama-3.1-70B-Instruct"
+
+echo "=== Quick RAG Debug Test ==="
+echo "2 profiles, 5 sessions - checking if extraction/storage works"
+date
+
+# Clear empty store
+> /projects/bfqt/users/yurenh2/ml-projects/personalization-user-model/data/corpora/empty_store/memory_cards.jsonl
+rm -f /projects/bfqt/users/yurenh2/ml-projects/personalization-user-model/data/corpora/empty_store/memory_embeddings.npy
+
+# Start vLLM servers
+CUDA_VISIBLE_DEVICES=0,1 python -m vllm.entrypoints.openai.api_server \
+    --model $USER_MODEL \
+    --port 8004 --tensor-parallel-size 2 --gpu-memory-utilization 0.90 \
+    --max-model-len 16384 --dtype bfloat16 --download-dir $HF_HOME &
+
+CUDA_VISIBLE_DEVICES=2,3 python -m vllm.entrypoints.openai.api_server \
+    --model $AGENT_MODEL \
+    --port 8003 --tensor-parallel-size 2 --gpu-memory-utilization 0.45 \
+    --max-model-len 16384 --dtype bfloat16 &
+
+echo "Waiting for vLLM servers..."
+for i in {1..200}; do
+    if curl -s http://localhost:8004/health > /dev/null 2>&1; then
+        echo "User simulator ready after $((i*5))s"
+        break
+    fi
+    sleep 5
+done
+for i in {1..60}; do
+    if curl -s http://localhost:8003/health > /dev/null 2>&1; then
+        echo "Agent ready after $((i*5))s"
+        break
+    fi
+    sleep 5
+done
+sleep 5
+
+OUTPUT_DIR="../results/rag_debug_$(date +%Y%m%d_%H%M%S)"
+
+# Only test RAG to see debug output
+echo "============================================"
+echo "Testing RAG with debug output"
+echo "============================================"
+
+python scripts/run_experiments.py --methods rag \
+    --datasets math-hard --n-profiles 2 --n-sessions 5 --max-turns 10 \
+    --use-vllm --no-batch-processing --parallel-profiles 2 \
+    --output-dir $OUTPUT_DIR --profile-path $PROFILE_PATH
+
+echo "Memory cards in file: $(wc -l < /projects/bfqt/users/yurenh2/ml-projects/personalization-user-model/data/corpora/empty_store/memory_cards.jsonl)"
+echo "=== Done ==="
+date
+
+pkill -f "vllm.entrypoints" 2>/dev/null || true