playground/inference/active_cases/run_all.sh

#!/bin/bash
# Orchestrate the full active-cases evaluation pipeline:
#   1. Extract frames for all 238 hard cases
#   2. (Optionally) download missing models
#   3. Evaluate each model sequentially on CUDA_VISIBLE_DEVICES=0
#
# Usage:
#   bash run_all.sh [--skip-extract] [--skip-download] [--models m1,m2,...]
#
# Logs: /mlx/users/jiashuo.fan/playground/inference/active_cases/logs/

set -euo pipefail

SCRIPT_DIR=/mlx/users/jiashuo.fan/playground/inference/active_cases
PYTHON_ABBIE=/mlx/users/jiashuo.fan/miniconda3/envs/abbie/bin/python3
LOG_DIR=$SCRIPT_DIR/logs
FRAMES_DIR=$SCRIPT_DIR/frames_cache
OUTPUT_DIR=/mnt/bn/bohanzhainas1/jiashuo/exp/active_cases_eval

mkdir -p "$LOG_DIR" "$FRAMES_DIR" "$OUTPUT_DIR"

# ── Parse args ────────────────────────────────────────────────────────────────

SKIP_EXTRACT=false
SKIP_DOWNLOAD=false
ONLY_MODELS=""

for arg in "$@"; do
    case $arg in
        --skip-extract)   SKIP_EXTRACT=true   ;;
        --skip-download)  SKIP_DOWNLOAD=true  ;;
        --models=*)       ONLY_MODELS="${arg#*=}" ;;
    esac
done

# ── Step 1: Frame extraction ──────────────────────────────────────────────────

if [ "$SKIP_EXTRACT" = false ]; then
    echo ""
    echo "========================================================"
    echo "STEP 1: Extracting frames for 238 hard cases"
    echo "========================================================"
    LOG_EXTRACT=$LOG_DIR/extract_frames.log
    echo "Log: $LOG_EXTRACT"
    $PYTHON_ABBIE -u $SCRIPT_DIR/extract_frames.py \
        --output-dir "$FRAMES_DIR" \
        2>&1 | tee "$LOG_EXTRACT"
    echo "Frame extraction done."
else
    echo "[SKIP] Frame extraction (--skip-extract)"
fi

# ── Step 2: Download missing models ──────────────────────────────────────────

if [ "$SKIP_DOWNLOAD" = false ]; then
    echo ""
    echo "========================================================"
    echo "STEP 2: Downloading missing models"
    echo "========================================================"
    LOG_DOWNLOAD=$LOG_DIR/download_models.log
    echo "Log: $LOG_DOWNLOAD"
    bash $SCRIPT_DIR/download_models.sh 2>&1 | tee "$LOG_DOWNLOAD"
    echo "Model downloads done."
else
    echo "[SKIP] Model downloads (--skip-download)"
fi

# ── Model list ────────────────────────────────────────────────────────────────
# Format: "model_name:model_type:extra_flags"
# extra_flags can be "--finetuned" for the fine-tuned model

ALL_MODELS=(
    "qwen3vl_finetuned:qwen3vl:--finetuned"
    "qwen3vl_base:qwen3vl:"
    "qwen3vl_instruct:qwen3vl:"
    "qwen25vl_7b:qwen25vl:"
    "internvl3_8b:internvl:"
    "internvl35:internvl:"
    "llava_ov:llava:"
    "llama32_11b_vision:llama32_vision:"
    "phi35_vision:phi3_vision:"
    "minicpm_v25:minicpm_v:"
    "qwen2vl_7b:qwen25vl:"
    "pixtral_12b:pixtral:"
    "smolvlm:generic:"
    "glm4v_9b:generic:"
    "idefics3_8b:generic:"
    "phi4_multimodal:generic:"
    "cogvlm2_19b:generic:"
    "deepseek_vl2_small:generic:"
    "videollama2_7b:generic:"
    "gemma3_4b:generic:"
)

# Filter to requested models if --models= was specified
if [ -n "$ONLY_MODELS" ]; then
    IFS=',' read -ra REQUESTED <<< "$ONLY_MODELS"
    FILTERED=()
    for entry in "${ALL_MODELS[@]}"; do
        model_name="${entry%%:*}"
        for req in "${REQUESTED[@]}"; do
            if [ "$model_name" = "$req" ]; then
                FILTERED+=("$entry")
            fi
        done
    done
    ALL_MODELS=("${FILTERED[@]}")
fi

# ── Step 3: Eval each model ───────────────────────────────────────────────────

echo ""
echo "========================================================"
echo "STEP 3: Evaluating ${#ALL_MODELS[@]} models on 238 hard cases"
echo "========================================================"
echo ""

PASSED=()
FAILED=()

for entry in "${ALL_MODELS[@]}"; do
    model_name="${entry%%:*}"
    rest="${entry#*:}"
    model_type="${rest%%:*}"
    extra_flags="${rest#*:}"

    echo ""
    echo "------------------------------------------------------------"
    echo "Evaluating: $model_name  (type=$model_type)"
    echo "------------------------------------------------------------"

    # Check if already done
    OUT_FILE="$OUTPUT_DIR/${model_name}.json"
    if [ -f "$OUT_FILE" ]; then
        # Check if it has complete results (all 238 samples)
        N_RESULTS=$(python3 -c "
import json
d = json.load(open('$OUT_FILE'))
print(d.get('evaluated', 0))
" 2>/dev/null || echo "0")
        if [ "$N_RESULTS" -ge 230 ]; then
            echo "[SKIP] $model_name already evaluated ($N_RESULTS samples)"
            PASSED+=("$model_name")
            continue
        fi
        echo "[RESUME] $model_name has $N_RESULTS results, continuing..."
    fi

    LOG_FILE="$LOG_DIR/eval_${model_name}.log"
    echo "Log: $LOG_FILE"

    # Build command
    CMD="CUDA_VISIBLE_DEVICES=0 $PYTHON_ABBIE -u $SCRIPT_DIR/eval_active_cases.py \
        --model-name  $model_name \
        --model-type  $model_type \
        --frames-dir  $FRAMES_DIR \
        --output-dir  $OUTPUT_DIR \
        --gpu-id      0"

    if [ -n "$extra_flags" ]; then
        CMD="$CMD $extra_flags"
    fi

    # Run
    START_T=$(date +%s)
    if eval "$CMD" 2>&1 | tee "$LOG_FILE"; then
        END_T=$(date +%s)
        ELAPSED=$((END_T - START_T))
        echo "[OK] $model_name done in ${ELAPSED}s"
        PASSED+=("$model_name")
    else
        END_T=$(date +%s)
        ELAPSED=$((END_T - START_T))
        echo "[FAIL] $model_name failed after ${ELAPSED}s — check $LOG_FILE"
        FAILED+=("$model_name")
    fi
done

# ── Step 4: Print summary table ───────────────────────────────────────────────

echo ""
echo "========================================================"
echo "FINAL SUMMARY"
echo "========================================================"
echo ""

$PYTHON_ABBIE -u $SCRIPT_DIR/compare_results.py \
    --results-dir "$OUTPUT_DIR" \
    2>/dev/null || echo "(compare_results.py not yet available)"

echo ""
echo "Passed: ${PASSED[*]:-none}"
echo "Failed: ${FAILED[*]:-none}"
echo ""
echo "Results at: $OUTPUT_DIR"
echo "Logs at:    $LOG_DIR"