Upload folder using huggingface_hub
Browse files
app.py
CHANGED
|
@@ -7,14 +7,13 @@ os.environ["TF_USE_LEGACY_KERAS"] = "1"
|
|
| 7 |
import numpy as np
|
| 8 |
import tensorflow as tf
|
| 9 |
from flask import Flask, request, jsonify
|
| 10 |
-
from tensorflow.keras.models import load_model
|
| 11 |
from model import build_model
|
| 12 |
|
| 13 |
pneumonia_detector_api = Flask(__name__)
|
| 14 |
|
| 15 |
IMG_SIZE = 224
|
| 16 |
-
#MODEL_PATH = "pneumonia_resnet_model.keras"
|
| 17 |
MODEL_PATH = "pneumonia_resnet.weights.h5"
|
|
|
|
| 18 |
print("TensorFlow Version:", tf.__version__)
|
| 19 |
|
| 20 |
# -----------------------------
|
|
@@ -24,12 +23,14 @@ def load_trained_model():
|
|
| 24 |
|
| 25 |
try:
|
| 26 |
|
| 27 |
-
print("
|
| 28 |
-
print("Loading model...")
|
| 29 |
|
| 30 |
model = build_model()
|
|
|
|
|
|
|
|
|
|
| 31 |
model.load_weights(
|
| 32 |
-
|
| 33 |
skip_mismatch=True
|
| 34 |
)
|
| 35 |
|
|
@@ -42,6 +43,7 @@ def load_trained_model():
|
|
| 42 |
print("Model loading failed:", e)
|
| 43 |
raise RuntimeError("Failed to load model")
|
| 44 |
|
|
|
|
| 45 |
model = load_trained_model()
|
| 46 |
|
| 47 |
|
|
@@ -58,16 +60,23 @@ def home():
|
|
| 58 |
# -----------------------------
|
| 59 |
def preprocess_image(img_array):
|
| 60 |
|
| 61 |
-
img =
|
| 62 |
|
| 63 |
-
|
|
|
|
|
|
|
| 64 |
|
| 65 |
-
|
|
|
|
| 66 |
|
| 67 |
-
#
|
|
|
|
|
|
|
|
|
|
| 68 |
if img.shape[-1] == 1:
|
| 69 |
-
img =
|
| 70 |
|
|
|
|
| 71 |
img = np.expand_dims(img, axis=0)
|
| 72 |
|
| 73 |
return img
|
|
@@ -86,16 +95,18 @@ def predict():
|
|
| 86 |
|
| 87 |
file = request.files["file"]
|
| 88 |
|
| 89 |
-
#
|
| 90 |
img_array = np.load(file)
|
| 91 |
|
| 92 |
processed_img = preprocess_image(img_array)
|
| 93 |
|
|
|
|
|
|
|
| 94 |
prediction = model.predict(processed_img)
|
| 95 |
|
| 96 |
probability = float(prediction[0][0])
|
| 97 |
|
| 98 |
-
label = "Pneumonia" if probability > 0.5 else "Normal"
|
| 99 |
|
| 100 |
return jsonify({
|
| 101 |
"Predicted_Class": label,
|
|
@@ -104,6 +115,8 @@ def predict():
|
|
| 104 |
|
| 105 |
except Exception as e:
|
| 106 |
|
|
|
|
|
|
|
| 107 |
return jsonify({
|
| 108 |
"error": str(e)
|
| 109 |
}), 500
|
|
@@ -117,5 +130,5 @@ if __name__ == "__main__":
|
|
| 117 |
pneumonia_detector_api.run(
|
| 118 |
host="0.0.0.0",
|
| 119 |
port=7860,
|
| 120 |
-
debug=
|
| 121 |
)
|
|
|
|
| 7 |
import numpy as np
|
| 8 |
import tensorflow as tf
|
| 9 |
from flask import Flask, request, jsonify
|
|
|
|
| 10 |
from model import build_model
|
| 11 |
|
| 12 |
pneumonia_detector_api = Flask(__name__)
|
| 13 |
|
| 14 |
IMG_SIZE = 224
|
|
|
|
| 15 |
MODEL_PATH = "pneumonia_resnet.weights.h5"
|
| 16 |
+
|
| 17 |
print("TensorFlow Version:", tf.__version__)
|
| 18 |
|
| 19 |
# -----------------------------
|
|
|
|
| 23 |
|
| 24 |
try:
|
| 25 |
|
| 26 |
+
print("Building model architecture...")
|
|
|
|
| 27 |
|
| 28 |
model = build_model()
|
| 29 |
+
|
| 30 |
+
print("Loading weights...")
|
| 31 |
+
|
| 32 |
model.load_weights(
|
| 33 |
+
MODEL_PATH,
|
| 34 |
skip_mismatch=True
|
| 35 |
)
|
| 36 |
|
|
|
|
| 43 |
print("Model loading failed:", e)
|
| 44 |
raise RuntimeError("Failed to load model")
|
| 45 |
|
| 46 |
+
|
| 47 |
model = load_trained_model()
|
| 48 |
|
| 49 |
|
|
|
|
| 60 |
# -----------------------------
|
| 61 |
def preprocess_image(img_array):
|
| 62 |
|
| 63 |
+
img = np.array(img_array)
|
| 64 |
|
| 65 |
+
# Ensure image has channel dimension
|
| 66 |
+
if len(img.shape) == 2:
|
| 67 |
+
img = np.expand_dims(img, axis=-1)
|
| 68 |
|
| 69 |
+
# Resize image
|
| 70 |
+
img = tf.image.resize(img, (IMG_SIZE, IMG_SIZE)).numpy()
|
| 71 |
|
| 72 |
+
# Normalize
|
| 73 |
+
img = img.astype("float32") / 255.0
|
| 74 |
+
|
| 75 |
+
# Convert grayscale → RGB
|
| 76 |
if img.shape[-1] == 1:
|
| 77 |
+
img = np.repeat(img, 3, axis=-1)
|
| 78 |
|
| 79 |
+
# Add batch dimension
|
| 80 |
img = np.expand_dims(img, axis=0)
|
| 81 |
|
| 82 |
return img
|
|
|
|
| 95 |
|
| 96 |
file = request.files["file"]
|
| 97 |
|
| 98 |
+
# Load .npy file
|
| 99 |
img_array = np.load(file)
|
| 100 |
|
| 101 |
processed_img = preprocess_image(img_array)
|
| 102 |
|
| 103 |
+
print("Input shape to model:", processed_img.shape)
|
| 104 |
+
|
| 105 |
prediction = model.predict(processed_img)
|
| 106 |
|
| 107 |
probability = float(prediction[0][0])
|
| 108 |
|
| 109 |
+
label = "Pneumonia" if probability >= 0.5 else "Normal"
|
| 110 |
|
| 111 |
return jsonify({
|
| 112 |
"Predicted_Class": label,
|
|
|
|
| 115 |
|
| 116 |
except Exception as e:
|
| 117 |
|
| 118 |
+
print("Prediction error:", str(e))
|
| 119 |
+
|
| 120 |
return jsonify({
|
| 121 |
"error": str(e)
|
| 122 |
}), 500
|
|
|
|
| 130 |
pneumonia_detector_api.run(
|
| 131 |
host="0.0.0.0",
|
| 132 |
port=7860,
|
| 133 |
+
debug=False
|
| 134 |
)
|