Spaces:
Running on Zero
Running on Zero
| import json | |
| from pathlib import Path | |
| import numpy as np | |
| import tensorflow as tf | |
| from sklearn.metrics import classification_report, confusion_matrix | |
| from src.model import CLASS_NAMES | |
| from src.train import MODEL_DIR, load_split | |
| def main(): | |
| model_path = MODEL_DIR / "human_face_v1.keras" | |
| metadata_path = MODEL_DIR / "human_face_v1.json" | |
| if not model_path.exists(): | |
| raise SystemExit(f"Model file not found: {model_path}") | |
| # 1. Load metadata to get the decision threshold (0.5) | |
| metadata = json.loads(metadata_path.read_text(encoding="utf-8")) | |
| threshold = float(metadata["threshold"]) | |
| print("Loading test dataset...") | |
| # shuffle=False is critical so order of images matches predictions exactly | |
| test_ds = load_split("test", shuffle=False) | |
| print("Loading saved model...") | |
| model = tf.keras.models.load_model(model_path) | |
| print("Running evaluation on test split...") | |
| # Extract the true ground-truth labels (0 or 1) from the test dataset | |
| y_true = np.concatenate([labels.numpy().reshape(-1) for _, labels in test_ds]).astype(int) | |
| # Predict probabilities for all test images | |
| probabilities = model.predict(test_ds, verbose=1).reshape(-1) | |
| # Convert probability scores into 0 (human) or 1 (non_human) using threshold | |
| y_pred = (probabilities >= threshold).astype(int) | |
| # 2. Print evaluation metrics | |
| print("\n" + "=" * 50) | |
| print("Class order:", CLASS_NAMES, "(0 = human, 1 = non_human)") | |
| print("Decision threshold:", threshold) | |
| print("=" * 50) | |
| print("\nConfusion Matrix (Rows = True class, Columns = Predicted class):") | |
| print(confusion_matrix(y_true, y_pred, labels=[0, 1])) | |
| print("\nClassification Report:") | |
| print( | |
| classification_report( | |
| y_true, | |
| y_pred, | |
| labels=[0, 1], | |
| target_names=CLASS_NAMES, | |
| zero_division=0, | |
| digits=4, | |
| ) | |
| ) | |
| if __name__ == "__main__": | |
| main() |