dicer2000/fordemo_visiontransformerattention3.ipynb

## fordemo_visiontransformerattention3.ipynb
{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "gpuType": "T4",
      "authorship_tag": "ABX9TyPqzAEMJZDEAPsrO+SPrVgY",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/gist/dicer2000/b4210a364243d24fad2ee8dd824b6474/fordemo_visiontransformerattention3.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Transformer Demo 3\n",
        "(c)2024 Brett Huffman"
      ],
      "metadata": {
        "id": "uuGbAfQJarpC"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install vit-keras\n",
        "!pip install tensorflow_addons"
      ],
      "metadata": {
        "id": "OIq5Hht0OR6z"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from vit_keras import vit, utils, visualize, layers\n",
        "import tensorflow as tf"
      ],
      "metadata": {
        "id": "YBx9IckXPd88"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "tazNa6fBOEWi"
      },
      "outputs": [],
      "source": [
        "\n",
        "\n",
        "image_size = 224\n",
        "classes = utils.get_imagenet_classes()\n",
        "model = vit.vit_b16(\n",
        "    image_size=image_size,\n",
        "    activation='sigmoid',\n",
        "    pretrained=True,\n",
        "    include_top=True,\n",
        "    pretrained_top=True\n",
        ")\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def read_image(file_name):\n",
        "  image = tf.io.read_file(file_name)\n",
        "  image = tf.io.decode_jpeg(image, channels=3)\n",
        "  image = tf.image.convert_image_dtype(image, tf.float32)\n",
        "  image = tf.image.resize_with_pad(image, target_height=224, target_width=224)\n",
        "  img_numpy = image.numpy()\n",
        "  return img_numpy"
      ],
      "metadata": {
        "id": "etIrl5q2a8QI"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def read_image(file_name):\n",
        "  image = tf.io.read_file(file_name)\n",
        "  image = tf.io.decode_jpeg(image, channels=3)\n",
        "  image = tf.image.convert_image_dtype(image, tf.float32)\n",
        "  image = tf.image.resize_with_pad(image, target_height=224, target_width=224)\n",
        "  return image"
      ],
      "metadata": {
        "id": "UikxAK582IHX"
      },
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def read_image3(file_path):\n",
        "    # Read the image file\n",
        "    img_raw = tf.io.read_file(file_path)\n",
        "\n",
        "    # Decode the image\n",
        "    try:\n",
        "        img_tensor = tf.image.decode_jpeg(img_raw, channels=3)\n",
        "    except tf.errors.InvalidArgumentError:\n",
        "        print(f\"Error decoding image: {file_path}\")\n",
        "        return None\n",
        "\n",
        "    # Resize the image to the expected size\n",
        "    img_tensor = tf.image.resize(img_tensor, [224, 224])\n",
        "\n",
        "    # Convert to numpy ndarray\n",
        "    img_numpy = img_tensor.numpy()\n",
        "\n",
        "    return img_numpy"
      ],
      "metadata": {
        "id": "_KEyeH4A6qCn"
      },
      "execution_count": 140,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "img_url = {\n",
        "    'el': 'https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEht-h9VBfRnHAeBYV9PrCrb4yzi5MJkkvcIrePisFF583RtFjmrGIlRDeJnNg06pQokt7OTusGHwqr-qZWphU2NN59LsIK-2ssQjFHGlGWf0OlQq9uzW8GP1EveQUQHtmf0rOMl3x6RgX2_BB5hYCxnnhIKtPx3Iw_sCC0p2s6g5B7Mop-btPhKW8bH/s1000/OriginalBullElephant_good.jpeg',\n",
        "    'Giant Panda': 'http://storage.googleapis.com/download.tensorflow.org/example_images/Giant_Panda_2.jpeg',\n",
        "}\n",
        "\n",
        "img_paths = {name: tf.keras.utils.get_file(name, url) for (name, url) in img_url.items()}\n",
        "img_name_tensors = {name: read_image3(img_path) for (name, img_path) in img_paths.items()}"
      ],
      "metadata": {
        "id": "pZ2a0HRqas5C"
      },
      "execution_count": 145,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# For Testing\n",
        "#url = 'http://storage.googleapis.com/download.tensorflow.org/example_images/Giant_Panda_2.jpeg'\n",
        "url2 = 'https://blogger.googleusercontent.com/img/b/R29vZ2xl/AVvXsEht-h9VBfRnHAeBYV9PrCrb4yzi5MJkkvcIrePisFF583RtFjmrGIlRDeJnNg06pQokt7OTusGHwqr-qZWphU2NN59LsIK-2ssQjFHGlGWf0OlQq9uzW8GP1EveQUQHtmf0rOMl3x6RgX2_BB5hYCxnnhIKtPx3Iw_sCC0p2s6g5B7Mop-btPhKW8bH/s1000/OriginalBullElephant_good.jpeg'\n",
        "img = read_image3(tf.keras.utils.get_file('panda', url2))"
      ],
      "metadata": {
        "id": "0XgI0IIk5Wcb"
      },
      "execution_count": 141,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plt.imshow(img/255.0)\n",
        "plt.axis('off')  # To turn off the axis\n",
        "plt.show()\n",
        "\n",
        "idx, cls = top_k_predictions(img, image_size)\n"
      ],
      "metadata": {
        "id": "EjEhRURS9TUZ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def top_k_predictions(incoming, image_size, k=3):\n",
        "\n",
        "  # Make a copy\n",
        "  incoming_copy = np.copy(incoming)\n",
        "\n",
        "  # Add a dimension for the batch\n",
        "  in_exp = np.expand_dims(incoming_copy, axis=0)\n",
        "  X = vit.preprocess_inputs(in_exp)\n",
        "\n",
        "  # Get model Inference\n",
        "  Y = model.predict(X)\n",
        "\n",
        "  # Print top k classes\n",
        "  for idx, l in enumerate(Y[0].argsort()[-k:][::-1]): #Top k classes\n",
        "    print(idx+1, classes[l])\n",
        "\n",
        "  predicted_class_index = Y[0].argmax()\n",
        "  predicted_class = classes[predicted_class_index]\n",
        "\n",
        "  return predicted_class_index, predicted_class"
      ],
      "metadata": {
        "id": "_ebenZtjdGRs"
      },
      "execution_count": 163,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "for n, (name, img_tens) in enumerate(img_name_tensors.items()):\n",
        "  plt.imshow(img_tens/255.0)\n",
        "  plt.title(name, fontweight='bold')\n",
        "  plt.axis('off')\n",
        "  plt.show()\n",
        "\n",
        "  target_class_idx, predicted_class = top_k_predictions(img_tens, image_size)\n",
        "  print('Target Class: ', target_class_idx, predicted_class)"
      ],
      "metadata": {
        "id": "6UCCpTOKc437"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Build an Integrated Gradients implementation\n"
      ],
      "metadata": {
        "id": "yXt1nAmKAy61"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Using Integrated Gradients\n",
        "\n",
        "def integrated_gradients(model, baseline, input_image, target_class_idx, steps=5):\n",
        "    # Create a series of linearly interpolated images from the baseline to the input image\n",
        "    interpolated_images_np = np.array([(baseline + (input_image - baseline) * i / steps) for i in range(0, steps+1)])\n",
        "\n",
        "    # Convert the interpolated images to a TensorFlow tensor\n",
        "    interpolated_images = tf.convert_to_tensor(interpolated_images_np, dtype=tf.float32)\n",
        "\n",
        "\n",
        "    # Create a function to compute gradients\n",
        "    with tf.GradientTape(watch_accessed_variables=False) as tape:\n",
        "\n",
        "        tape.watch(interpolated_images)\n",
        "        preds = model(interpolated_images)\n",
        "        target_class_scores = preds[:, target_class_idx]\n",
        "\n",
        "\n",
        "    gradients = tape.gradient(target_class_scores, interpolated_images)\n",
        "\n",
        "    # Integrate the gradients\n",
        "    avg_gradients = np.mean(gradients, axis=0)\n",
        "\n",
        "    # Compute the attribution scores\n",
        "    attributions = (input_image - baseline) * avg_gradients\n",
        "\n",
        "    return interpolated_images, attributions"
      ],
      "metadata": {
        "id": "7hMTdFWtPEVv"
      },
      "execution_count": 226,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Baseline Black\n",
        "baseline = np.zeros_like(img) # Black"
      ],
      "metadata": {
        "id": "vwym9PPvPacS"
      },
      "execution_count": 227,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Define baseline image and calculate Integrated Gradient\n",
        "m_steps = 5\n",
        "image_to_analyze = img_name_tensors['Giant Panda']\n",
        "alphas = tf.linspace(start=0.0, stop=1.0, num=m_steps+1) # Generate m_steps intervals\n",
        "img_gradients, img_attributions = integrated_gradients(model, baseline, image_to_analyze, target_class_idx, steps=m_steps)\n"
      ],
      "metadata": {
        "id": "26IQ72suPsDQ"
      },
      "execution_count": 231,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "for n, int_image in enumerate(img_gradients):\n",
        "\n",
        "  plt.imshow(int_image/255.0)\n",
        "  plt.axis('off')  # To turn off the axis\n",
        "  plt.show()"
      ],
      "metadata": {
        "id": "Gsl84MwGLQB2"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plt.imshow(img_attributions*50.0)\n",
        "plt.axis('off')  # To turn off the axis\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "a2BPO9AsJuSF"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "pred = model(img_gradients)\n",
        "pred_proba = tf.nn.softmax(pred, axis=-1)[:, 387]"
      ],
      "metadata": {
        "id": "0B2QpVkyNJac"
      },
      "execution_count": 208,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(pred_proba)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "J7HxklcqNo5H",
        "outputId": "0cd07250-6ff3-4b4d-83e9-915ed261bfcf"
      },
      "execution_count": 209,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "tf.Tensor([0.00073145 0.00116896 0.00109917 0.00109573 0.00108756 0.00108285], shape=(6,), dtype=float32)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#@title\n",
        "plt.figure(figsize=(10, 4))\n",
        "ax1 = plt.subplot(1, 2, 1)\n",
        "ax1.plot(alphas, pred_proba*500)\n",
        "ax1.set_title('Target class predicted probability over alpha')\n",
        "ax1.set_ylabel('model p(target class)')\n",
        "ax1.set_xlabel('alpha')\n",
        "ax1.set_ylim([0, 1])\n",
        "\n",
        "ax2 = plt.subplot(1, 2, 2)\n",
        "# Average across interpolation steps\n",
        "average_grads = tf.reduce_mean(img_gradients, axis=[1, 2, 3])\n",
        "# Normalize gradients to 0 to 1 scale. E.g. (x - min(x))/(max(x)-min(x))\n",
        "average_grads_norm = (average_grads-tf.math.reduce_min(average_grads))/(tf.math.reduce_max(average_grads)-tf.reduce_min(average_grads))\n",
        "ax2.plot(alphas, average_grads_norm)\n",
        "ax2.set_title('Average pixel gradients (normalized) over alpha')\n",
        "ax2.set_ylabel('Average pixel gradients')\n",
        "ax2.set_xlabel('alpha')\n",
        "ax2.set_ylim([0, 1]);\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 410
        },
        "id": "Zk3RCBsGNKUF",
        "outputId": "5c9bf21a-bf54-45aa-e426-60aaef51ad64"
      },
      "execution_count": 220,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1000x400 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#@title\n",
        "plt.figure(figsize=(20, 4))\n",
        "ax1 = plt.subplot(1, 2, 1)\n",
        "ax1.plot(alphas, pred_proba)\n",
        "ax1.set_title('Target class predicted probability over alpha')\n",
        "ax1.set_ylabel('model p(target class)')\n",
        "ax1.set_xlabel('alpha')\n",
        "ax1.set_ylim([0, 1])\n",
        "\n",
        "# Average across interpolation steps\n",
        "average_grads = tf.reduce_mean(img_gradients, axis=[1, 2, 3])\n",
        "# Normalize gradients to 0 to 1 scale. E.g. (x - min(x))/(max(x)-min(x))\n",
        "average_grads_norm = (average_grads-tf.math.reduce_min(average_grads))/(tf.math.reduce_max(average_grads)-tf.reduce_min(average_grads))\n",
        "ax1.plot(alphas, average_grads_norm)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 427
        },
        "id": "XGM1ESKfPEW5",
        "outputId": "e250943a-846e-4632-a978-bac52fb04066"
      },
      "execution_count": 215,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7df9d181d990>]"
            ]
          },
          "metadata": {},
          "execution_count": 215
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 2000x400 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    }
  ]
}