Regenerated.md and .ipynb with the latest changes of .py

Author: LakshmiKalaKadali

examples/vision/focal_modulation_network.py

@@ -78,7 +78,7 @@
 import os
 
 # Set backend before importing keras
-os.environ["KERAS_BACKEND"] = "jax"  # Or "torch" or "tensorflow"
+os.environ["KERAS_BACKEND"] = "tensorflow"  # Or "torch" or "tensorflow"
 
 import numpy as np
 import keras

examples/vision/ipynb/focal_modulation_network.ipynb

@@ -102,13 +102,12 @@
     "import os\n",
     "\n",
     "# Set backend before importing keras\n",
-    "os.environ[\"KERAS_BACKEND\"] = \"jax\"  # Or \"torch\" or \"tensorflow\"\n",
+    "os.environ[\"KERAS_BACKEND\"] = \"tensorflow\"  # Or \"torch\" or \"tensorflow\"\n",
     "\n",
     "import numpy as np\n",
     "import keras\n",
     "from keras import layers\n",
     "from keras import ops\n",
-    "from typing import Optional\n",
     "from matplotlib import pyplot as plt\n",
     "from random import randint\n",
     "\n",
@@ -155,6 +154,7 @@
    },
    "source": [
     "## Data Loading with PyDataset\n",
+    "\n",
     "Keras 3 introduces PyDataset as a standardized way to handle data.\n",
     "It works identically across all backends and avoids the \"Symbolic Tensor\" issues often found\n",
     "when using tf.data with JAX or PyTorch."
@@ -363,12 +363,7 @@
    "outputs": [],
    "source": [
     "\n",
-    "def MLP(\n",
-    "    in_features: int,\n",
-    "    hidden_features: Optional[int] = None,\n",
-    "    out_features: Optional[int] = None,\n",
-    "    mlp_drop_rate: float = 0.0,\n",
-    "):\n",
+    "def MLP(in_features, hidden_features=None, out_features=None, mlp_drop_rate=0.0):\n",
     "    hidden_features = hidden_features or in_features\n",
     "    out_features = out_features or in_features\n",
     "    return keras.Sequential(\n",
@@ -468,7 +463,7 @@
     "| Equation 5: Linear projection of `Z^0` (Source: Aritra and Ritwik) |\n",
     "\n",
     "`Z^0` is then passed on to a series of Depth-Wise (DWConv) Conv and\n",
-    "[GeLU](hhttps://keras.io/api/layers/activations/#gelu-function) layers. The\n",
+    "[GeLU](https://keras.io/api/layers/activations/#gelu-function) layers. The\n",
     "authors term each block of DWConv and GeLU as levels denoted by `l`. In **Figure 6** we\n",
     "have two levels. Mathematically this is represented as:\n",
     "\n",