Focal Loss Implementation

Author: Sayed-Hossein-Hosseini

RhythmAttention_Hybrid_CNN_Transformer_Architecture_for_Arrhythmia_Classification.ipynb

@@ -89,25 +89,15 @@
     },
     {
       "cell_type": "code",
-      "execution_count": 2,
+      "execution_count": 7,
       "metadata": {},
-      "outputs": [
-        {
-          "name": "stderr",
-          "output_type": "stream",
-          "text": [
-            "2026-02-08 17:57:33.534489: I tensorflow/core/util/port.cc:153] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.\n",
-            "2026-02-08 17:57:33.583704: I tensorflow/core/platform/cpu_feature_guard.cc:210] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.\n",
-            "To enable the following instructions: AVX2 AVX512F AVX512_VNNI FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
-            "2026-02-08 17:57:34.748647: I tensorflow/core/util/port.cc:153] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.\n"
-          ]
-        }
-      ],
+      "outputs": [],
       "source": [
         "import pandas as pd\n",
         "import numpy as np\n",
         "import tensorflow as tf\n",
-        "from sklearn.utils import class_weight"
+        "from sklearn.utils import class_weight\n",
+        "from tensorflow.keras import layers, models, backend as K"
       ]
     },
     {
@@ -287,6 +277,40 @@
         "print(f\"Batch data shape: {x_batch.shape}\")\n",
         "print(f\"Batch labels shape: {y_batch.shape}\")"
       ]
+    },
+    {
+      "cell_type": "markdown",
+      "metadata": {},
+      "source": [
+        "## **Focal Loss Implementation**"
+      ]
+    },
+    {
+      "cell_type": "code",
+      "execution_count": 8,
+      "metadata": {},
+      "outputs": [],
+      "source": [
+        "def focal_loss(gamma=2.0, alpha=0.25):\n",
+        "    \"\"\"\n",
+        "    Focuses on hard examples by reducing the weight of easy ones.\n",
+        "    gamma: power factor to scale the loss.\n",
+        "    alpha: balancing factor for classes.\n",
+        "    \"\"\"\n",
+        "    def focal_loss_fixed(y_true, y_pred):\n",
+        "        # Clip values to prevent log(0)\n",
+        "        epsilon = K.epsilon()\n",
+        "        y_pred = K.clip(y_pred, epsilon, 1.0 - epsilon)\n",
+        "        \n",
+        "        # Calculate Cross Entropy\n",
+        "        cross_entropy = -y_true * K.log(y_pred)\n",
+        "        \n",
+        "        # Calculate Focal Loss\n",
+        "        loss = alpha * K.pow(1 - y_pred, gamma) * cross_entropy\n",
+        "        \n",
+        "        return K.mean(K.sum(loss, axis=-1))\n",
+        "    return focal_loss_fixed"
+      ]
     }
   ],
   "metadata": {