Merge pull request #113 from Aydinhamedi/main

sync

Model_T&T.ipynb

@@ -6,7 +6,7 @@
    "source": [
     "# keras/TF model\n",
     "<pre>\n",
-    " Copyright (c) 2023 Aydin Hamedi\n",
+    " Copyright (c) 2024 Aydin Hamedi\n",
     " \n",
     " This software is released under the MIT License.\n",
     " https://opensource.org/licenses/MIT\n",
@@ -19405,7 +19405,7 @@
     "# Garbage Collection (memory)\n",
     "gc.collect()\n",
     "\n",
-    "Extra_EXT = '_T_BL' # _T or _T_BL\n",
+    "Extra_EXT = '_T' # _T or _T_BL\n",
     "Train_data_test = False\n",
     "if SAVE_TYPE == 'TF':\n",
     "    # Load the pre-trained model\n",
@@ -19419,39 +19419,21 @@
     "\n",
     "# Make predictions on validation data\n",
     "val_predictions = model.predict(x_val)\n",
-    "val_predictions = np.argmax(val_predictions, axis=1)\n",
-    "\n",
-    "# Make predictions on Train data\n",
-    "if Train_data_test:\n",
-    "    Train_predictions = model.predict(x_train)\n",
-    "    Train_predictions = np.argmax(Train_predictions, axis=1)\n",
     "\n",
     "# Make predictions on test data\n",
     "test_predictions = model.predict(x_test)\n",
-    "test_predictions = np.argmax(test_predictions, axis=1)\n",
-    "\n",
-    "# Convert y_val and y_test from one-hot encoder to their original form\n",
-    "y_val_original = np.argmax(y_val, axis=1)\n",
-    "y_test_original = np.argmax(y_test, axis=1)\n",
-    "if Train_data_test:\n",
-    "    y_train_original = np.argmax(y_train, axis=1)\n",
-    "\n",
-    "# Calculate accuracy on validation data\n",
-    "val_accuracy = accuracy_score(y_val_original, val_predictions)\n",
-    "\n",
-    "# Calculate accuracy on Train data\n",
-    "if Train_data_test:\n",
-    "    Train_accuracy = accuracy_score(y_val_original, Train_predictions)\n",
-    "\n",
-    "# Calculate accuracy on test data\n",
-    "test_accuracy = accuracy_score(y_test_original, test_predictions)\n",
     "\n",
     "# Print acc\n",
-    "if Train_data_test:\n",
-    "    print(f'The accuracy of the model on Train data is {Train_accuracy:.2%}({Train_accuracy:.5%})')\n",
-    "print(f'The accuracy of the model on validation data is {val_accuracy:.2%}({val_accuracy:.5%})')\n",
-    "print(f'The accuracy of the model on test data is {test_accuracy:.2%}({test_accuracy:.5%})')\n",
+    "print('Val data acc:')\n",
+    "evaluate_model_full(y_val, val_predictions)\n",
+    "print('Test data acc:')\n",
+    "evaluate_model_full(y_test, test_predictions)\n",
     "\n",
+    "# format data\n",
+    "val_predictions = np.argmax(val_predictions, axis=1)\n",
+    "test_predictions = np.argmax(test_predictions, axis=1)\n",
+    "y_val_original = np.argmax(y_val, axis=1)\n",
+    "y_test_original = np.argmax(y_test, axis=1)\n",
     "# Visualize the predictions on validation data as a grid of squares\n",
     "plt.figure(figsize=(12, 6))\n",
     "for i in range(10):\n",
@@ -19512,8 +19494,6 @@
     "\n",
     "# Define the range of test data sizes to use\n",
     "data_sizes = range(1, len(x_test), 4)  \n",
-    "# Calculate the probability of a wrong prediction based on test accuracy\n",
-    "prob_wrong = 1 - test_accuracy\n",
     "\n",
     "# Create a list to store the number of incorrect predictions for each test data size\n",
     "incorrect_predictions = []\n",
@@ -19549,51 +19529,7 @@
     "plt.yticks(np.arange(0, max(incorrect_predictions) + 5, 3))\n",
     "\n",
     "plt.title('Number of Incorrect Predictions vs. Number of Data Points')\n",
-    "plt.show()\n",
-    "# Deprecated⚠️------------------------------>>>\n",
-    "# prob_L = 0.9995\n",
-    "# # Define the range of test data sizes to use\n",
-    "# data_sizes = range(1, len(x_test), 1) \n",
-    "\n",
-    "# # Calculate the probability of a wrong prediction based on test accuracy\n",
-    "# prob_wrong = 1 - test_accuracy\n",
-    "\n",
-    "# # Create a list to store the probability of getting at least one wrong answer for each test data size\n",
-    "# probabilities = []\n",
-    "\n",
-    "# # Calculate the probability of getting at least one wrong answer for each data size\n",
-    "# for size in data_sizes:\n",
-    "#     # Calculate the cumulative distribution function (CDF) of the binomial distribution at 0\n",
-    "#     cdf = binom.cdf(0, size, prob_wrong)\n",
-    "#     # Subtract the CDF from 1 to get the probability of getting at least one wrong answer\n",
-    "#     prob = 1 - cdf\n",
-    "#     probabilities.append(prob)\n",
-    "\n",
-    "# # Find the index of the first data point that has a probability greater than prob_L%\n",
-    "# index = next((i for i, p in enumerate(probabilities) if p > prob_L), len(probabilities))\n",
-    "\n",
-    "# # Limit the x-axis to the first data point that has a probability greater than prob_L%\n",
-    "# data_sizes = data_sizes[:index+1]\n",
-    "# probabilities = probabilities[:index+1]\n",
-    "\n",
-    "# # Plot the probability vs. the number of data points\n",
-    "# plt.figure(figsize=(10, 6))\n",
-    "# plt.plot(data_sizes, probabilities)\n",
-    "# plt.xlabel('Number of Data Points')\n",
-    "# plt.ylabel('Probability')\n",
-    "\n",
-    "# # Add gridlines for the x and y axes\n",
-    "# plt.grid(True)\n",
-    "\n",
-    "# # Change the tick spacing for the x and y axes\n",
-    "# plt.xticks(np.arange(min(data_sizes), max(data_sizes)+1, 5 + 10))\n",
-    "# plt.yticks(np.arange(0, max(probabilities)+0.1, 5 / 100))\n",
-    "\n",
-    "# plt.ylim(top=1.01)\n",
-    "\n",
-    "# plt.title('Probability of Getting at Least One Wrong Answer vs. Number of Data Points')\n",
-    "# plt.show()\n",
-    "# Deprecated⚠️------------------------------<<<"
+    "plt.show()"
    ]
   }
  ],

README.md

@@ -1,4 +1,4 @@
-# Pneumonia Detection AI 🤖 - (🚧Alpha-b🚧)
+# Pneumonia Detection AI 🤖
 
 <img src="https://img.shields.io/badge/Python-FFD43B?style=for-the-badge&logo=python&logoColor=blue"/> <img src="https://img.shields.io/badge/Jupyter-F37626.svg?&style=for-the-badge&logo=Jupyter&logoColor=white"/> <img src="https://img.shields.io/badge/TensorFlow-FF6F00?style=for-the-badge&logo=tensorflow&logoColor=white"/> <img src="https://img.shields.io/badge/Keras-FF0000?style=for-the-badge&logo=keras&logoColor=white"/> <img src="https://img.shields.io/badge/OpenCV-27338e?style=for-the-badge&logo=OpenCV&logoColor=white"/>
 

Utils/Other.py

@@ -49,19 +49,24 @@ def P_warning(msg):
     """
     print_Color_V2(f'<light_red>Warning: <yellow>{msg}')
 
-def evaluate_model_full(model, x_test, y_test):
+def evaluate_model_full(y_test, model_pred, model=None, x_test=None):
     """Evaluates a machine learning model on a test set.
 
     Args:
-        model: The model to evaluate.
-        x_test: The test data features.
-        y_test: The test data labels.
+        x_test: Test set features.
+        y_test: Test set labels.  
+        model_pred: Model predictions.
+        model: The model object.
 
     Returns:
-        None. Prints a table with accuracy, precision, recall and F1 score.
+        None. Prints a table with accuracy, precision, recall and 
+        F1 score.
     """
     # Get the model predictions
-    y_pred = model.predict(x_test)
+    if model_pred is None:
+        y_pred = model.predict(x_test)
+    else:
+        y_pred = model_pred
 
     # Convert one-hot encoded predictions and labels to label encoded form
     y_pred_bin = np.argmax(y_pred, axis=1)