File size: 4,136 Bytes
b33e5eb | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 | #!/usr/bin/env bash
# Full SIFQ evaluation pipeline — verify SOTA claim
#
# What this script does:
# Step 1: Run SIFQ inference → Q scores for all test images
# Step 2: Compute Track 1 (ERC), Track 2 (Sensor invariance), Track 4 (Concept grounding)
# Step 3: Print summary table + save plots to eval_results/
#
# SOTA claim requires:
# Track 1: AUC_ERC(SIFQ) < AUC_ERC(NFIQ2) → better matcher utility
# Track 2: mean_KS(SIFQ) < mean_KS(NFIQ2) → sensor-invariant
# cross_sensor_pearson(SIFQ) > pearson(NFIQ2) → consistent cross-sensor
#
# Usage:
# bash scripts/run_eval.sh --checkpoint checkpoints_full/last.pt
# bash scripts/run_eval.sh --checkpoint checkpoints_full/last.pt --nfiq2-scores /tmp/nfiq2.jsonl
# bash scripts/run_eval.sh ... --skip-track1 # fast eval (no MDGT inference)
set -euo pipefail
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"
VENV="$ROOT/../.venv"
[[ -f "$VENV/bin/activate" ]] && source "$VENV/bin/activate"
# ── Defaults ──────────────────────────────────────────────────────────────────
CHECKPOINT="${CHECKPOINT:-$ROOT/checkpoints_full/last.pt}"
MDGT_CKPT="${MDGT_CKPT:-/home/aiserver/works/fingerprint/pad/TRAM-downstream/checkpoint/checkpoints_dinov2_tram/best_eer.pt}"
ROOT_302B="${ROOT_302B:-/home/aiserver/works/fingerprint/dataset/302b/images/baseline}"
ROOT_302D="${ROOT_302D:-/home/aiserver/works/fingerprint/dataset/nist_302d/images/auxiliary}"
SCORES_FILE="${SCORES_FILE:-$ROOT/eval_results/sifq_scores.jsonl}"
OUT_DIR="${OUT_DIR:-$ROOT/eval_results}"
NFIQ2_SCORES="${NFIQ2_SCORES:-}"
MAX_SAMPLES="${MAX_SAMPLES:--1}" # -1 = all images
MAX_PAIRS="${MAX_PAIRS:-5000}"
SKIP_TRACK1=0
SKIP_TRACK4=0
# ── Parse extra args ──────────────────────────────────────────────────────────
while [[ $# -gt 0 ]]; do
case "$1" in
--checkpoint) CHECKPOINT="$2"; shift 2 ;;
--nfiq2-scores) NFIQ2_SCORES="$2"; shift 2 ;;
--out-dir) OUT_DIR="$2"; shift 2 ;;
--max-samples) MAX_SAMPLES="$2"; shift 2 ;;
--skip-track1) SKIP_TRACK1=1; shift ;;
--skip-track4) SKIP_TRACK4=1; shift ;;
*) echo "Unknown arg: $1"; exit 1 ;;
esac
done
mkdir -p "$OUT_DIR"
SCORES_FILE="$OUT_DIR/sifq_scores.jsonl"
echo "============================================================"
echo "SIFQ Evaluation Pipeline"
echo " Checkpoint : $CHECKPOINT"
echo " MDGT : $MDGT_CKPT"
echo " Output dir : $OUT_DIR"
echo "============================================================"
# ── Step 1: Inference ─────────────────────────────────────────────────────────
echo ""
echo "[Step 1/2] Running SIFQ inference..."
python "$SCRIPT_DIR/run_infer.py" \
--checkpoint "$CHECKPOINT" \
--root-302b "$ROOT_302B" \
--root-302d "$ROOT_302D" \
--output "$SCORES_FILE" \
--max-samples "$MAX_SAMPLES" \
--num-workers 2
echo ""
echo "[Step 2/2] Running evaluation (Track 1, 2, 4)..."
EXTRA_ARGS=""
[[ $SKIP_TRACK1 -eq 1 ]] && EXTRA_ARGS="$EXTRA_ARGS --skip-track1"
[[ $SKIP_TRACK4 -eq 1 ]] && EXTRA_ARGS="$EXTRA_ARGS --skip-track4"
[[ -n "$NFIQ2_SCORES" ]] && EXTRA_ARGS="$EXTRA_ARGS --nfiq2-scores $NFIQ2_SCORES"
python "$SCRIPT_DIR/run_eval.py" \
--sifq-scores "$SCORES_FILE" \
--checkpoint "$CHECKPOINT" \
--mdgt-checkpoint "$MDGT_CKPT" \
--out-dir "$OUT_DIR" \
--max-pairs "$MAX_PAIRS" \
$EXTRA_ARGS
echo ""
echo "Done. Results in: $OUT_DIR"
echo " eval_summary.json — numbers for all tracks"
echo " plot_erc.png — Track 1 ERC curve"
echo " plot_sensor_hist.png — Track 2 quality distribution per sensor"
echo " plot_sensor_scatter.png — Track 2 cross-sensor scatter"
echo " plot_crosstalk.png — Track 4 concept grounding heatmap"
|