ML-Explainability/NLP/gradient_explanations_BERT.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "e_Hedb6SoFFT"
      },
      "source": [
        "# Explaining `DistilBERT` with integrated gradients\n",
        "\n",
        "<a href=\"https://colab.research.google.com/drive/1Q-Tk27YFR3N5elO9XJvdfrhAUWqzBL3M\" target=\"_blank\">\n",
        "  <img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\">\n",
        "</a>\n",
        "\n",
        "Return to the [castle](https://github.com/Nkluge-correa/TeenyTinyCastle).\n",
        "\n",
        "BERT ([Bidirectional Encoder Representations from Transformers](https://paperswithcode.com/method/bert)) models have gained significant prominence in natural language processing tasks due to their ability to capture context and relationships within language. However, the complexity of these models poses challenges in terms of interpretability.\n",
        "\n",
        "The intricate architecture of BERT, with its multiple layers and attention mechanisms, makes it difficult for humans to comprehend the decision-making process behind its predictions. Hence, XAI researchers (EXplaineble AI) strive to bridge this gap by offering techniques to make BERT models more transparent and interpretable.\n",
        "\n",
        "XAI techniques, when applied to BERT models, can involve things like:\n",
        "\n",
        "- Visualizing attention maps to highlight important words or phrases.\n",
        "- Extracting relevant features that contribute to predictions.\n",
        "\n",
        "<img src=\"https://production-media.paperswithcode.com/methods/new_BERT_Overall.jpg\" alt=\"bert-model\" width=\"800\"/>\n",
        "\n",
        "[Source](https://arxiv.org/abs/1810.04805v2).\n",
        "\n",
        "In this tutorial, we will try to understand the outputs of a BERT model by using a technique called [`integrated Gradients`](https://www.tensorflow.org/tutorials/interpretability/integrated_gradients). This method allows us to get some insight into the inner workings of models via an empirical analysis (input $x$ generates output $y$), using its gradient to determine what **influences** the individual parts of an input (**like words in a sentence or pixels in an image**) have on the output of a model.\n",
        "\n",
        "> **Note: This method was proposed in Sundararajan et al., in \"_[Axiomatic Attribution for Deep Networks](https://arxiv.org/abs/1703.01365)_\". To learn more about it, check out other tutorials [[1](https://github.com/Nkluge-correa/TeenyTinyCastle/blob/master/ML-Explainability/NLP/integrated_gradients_in_keras_nlp.ipynb)] [[2](https://colab.research.google.com/github/tensorflow/docs/blob/master/site/en/tutorials/interpretability/integrated_gradients.ipynb)].**\n",
        "\n",
        "The object of our explainability analysis will be the [DistilBERT model we created in this tutorial](https://github.com/Nkluge-correa/TeenyTinyCastle/blob/master/ML-Explainability/NLP/model_maker.ipynb). Hence, let us downloaded from the Hub! 🤗"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 35,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KKj9rVjJNqw0",
        "outputId": "64922c58-0a5e-4810-eb47-18e1dd2c2913"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\n",
            "Review: 'Is to complicated and boring.'\n",
            "(Label: NEGATIVE | Confidence: 94.99%)\n",
            "\n",
            "Review: 'Is nice to see philosophers doing machine learning.'\n",
            "(Label: POSITIVE | Confidence: 99.69%)\n"
          ]
        }
      ],
      "source": [
        "from transformers import TFAutoModelForSequenceClassification, AutoTokenizer\n",
        "from transformers import TextClassificationPipeline\n",
        "\n",
        "# Load the model and tokenizer\n",
        "model = TFAutoModelForSequenceClassification.from_pretrained(\"AiresPucrs/distilbert-base-cased-sentiment-classifier\")\n",
        "tokenizer = AutoTokenizer.from_pretrained(\"AiresPucrs/distilbert-base-cased-sentiment-classifier\")\n",
        "\n",
        "# Create a text classification pipeline\n",
        "pipeline = TextClassificationPipeline(model=model, tokenizer=tokenizer)\n",
        "\n",
        "# Classify some samples\n",
        "texts = [\n",
        "    'Is to complicated and boring.',\n",
        "    'Is nice to see philosophers doing machine learning.',\n",
        "  ]\n",
        "\n",
        "for text in texts:\n",
        "  preds = pipeline(text)\n",
        "\n",
        "  print(f\"\"\"\\nReview: '{text}'\\n(Label: {preds[0]['label']} | Confidence: {preds[0]['score'] * 100:.2f}%)\"\"\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LZdaGBk1Q0Ss"
      },
      "source": [
        "You can also do classifications by passing inputs straight to the model and processing the **logits** yourself. **Logits** provide the raw scores or predictions from a model. At the same time, the **softmax** function is applied to these logits to convert them into probabilities, making it easier to interpret the model's confidence in each class."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 57,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GjA6ZZC4K6BI",
        "outputId": "90d111fc-62a4-4065-daef-85666017310a"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Logits:  tf.Tensor([[-3.9134598  3.0466256]], shape=(1, 2), dtype=float32)\n",
            "Class:  POSITIVE\n",
            "Confidence:  0.99905187\n"
          ]
        }
      ],
      "source": [
        "import tensorflow as tf\n",
        "\n",
        "text = \"This was a masterpiece. Not completely faithful to the books, \\\n",
        "but enthralling from beginning to end. Might be my favorite of the three.\"\n",
        "\n",
        "inputs = tokenizer(text, return_tensors=\"tf\")\n",
        "\n",
        "# Pass it through the model and get the logits\n",
        "logits = model(**inputs).logits\n",
        "print(\"Logits: \", logits)\n",
        "\n",
        "# Get the index of the maximum argument (the class)\n",
        "prediction = tf.math.argmax(logits, axis=-1)\n",
        "print(\"Class: \", model.config.id2label[int(prediction)])\n",
        "\n",
        "# Get the logit value after softmax (the confidence)\n",
        "softmax_output = tf.nn.softmax(logits).numpy()[0][int(prediction)]\n",
        "print(\"Confidence: \", softmax_output)\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PWaeQCdFoFFX"
      },
      "source": [
        "Now, let us try to understand these results.\n",
        "\n",
        "The `integrated gradients` is a powerful technique for computing feature attributions within a deep learning model. This approach leverages the gradient of the model's output (prediction) concerning the input features, providing a comprehensive understanding of the contribution of each feature. Notably, integrated gradients apply to a broad range of deep learning models, encompassing NLP, CV, and models that work with tabular data.\n",
        "\n",
        "In the context of a text classification model, the method proves invaluable for interpreting predictions. Specifically, integrated gradients allow us to derive attribution scores for each input token, shedding light on the significance of individual terms in influencing the final prediction. By analyzing these attribution scores, we gain insights into the pivotal tokens in shaping the model's output.\n",
        "\n",
        "To implement this methodology, we have devised a straightforward function. **This function inputs a piece of text, the target model, and a tokenizer. The function output comprises a list of gradient attributions, the corresponding attributed tokens, and the final output class**. This information equips practitioners with a nuanced understanding of the model's decision-making process, facilitating interpretability and providing a basis for informed analysis in various natural language processing scenarios."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 77,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 208
        },
        "id": "2TBBVFvjSM3Q",
        "outputId": "a1d91d0b-c5a5-4951-ed7a-8ea01b51f4e0"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "<b>Text</b>: <i>This was a masterpiece. Not completely faithful to the books, but enthralling from beginning to end. Might be my favorite of the three.</i>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "text/html": [
              "<b>Prediction class</b>: <i>POSITIVE</i>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "text/html": [
              "<b>Gradient scores</b>: <br>{'[CLS]': 0.21509483456611633, 'This': 0.3559677302837372, 'was': 0.3312913477420807, 'a': 0.17503726482391357, 'masterpiece': 0.6650427579879761, '.': 0.1221601665019989, 'Not': 1.0, 'completely': 0.48374468088150024, 'faithful': 0.7101697325706482, 'to': 0.09554636478424072, 'the': 0.10963360220193863, 'books': 0.26106077432632446, ',': 0.19107495248317719, 'but': 0.35722872614860535, 'en': 0.2396479845046997, '##th': 0.150899276137352, '##ral': 0.2721118628978729, '##ling': 0.19241978228092194, 'from': 0.1545965075492859, 'beginning': 0.22666917741298676, 'end': 0.2057007998228073, 'Might': 0.2811778783798218, 'be': 0.16716516017913818, 'my': 0.20016631484031677, 'favorite': 0.49090877175331116, 'of': 0.2002367377281189, 'three': 0.3074657618999481, '[SEP]': 0.1290356069803238}"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "from IPython.display import HTML\n",
        "\n",
        "def get_gradients(text, model, tokenizer):\n",
        "    \"\"\"\n",
        "    Compute the gradients of a given text with respect to the model's parameters.\n",
        "\n",
        "    Args:\n",
        "        text (str): The input text for which gradients need to be computed.\n",
        "        model: the model that generates the predictions.\n",
        "        tokenizer: The tokenizer object used for tokenizing the text.\n",
        "\n",
        "    Returns:\n",
        "        gradients (list): A list of gradient tensors, each corresponding to a\n",
        "        token in the sequence.\n",
        "        token_words (list): A list of tokens corresponding to the input text.\n",
        "        prediction_label (str): The predicted label for the input text.\n",
        "    \"\"\"\n",
        "    # Get the embedding weights of the model (which is a distilbert model)\n",
        "    embedding_matrix = model.distilbert.embeddings.weights[0]\n",
        "\n",
        "    # Get the size of the vocabulary (length of the embedding matrix)\n",
        "    vocab_size = embedding_matrix.shape[0]\n",
        "\n",
        "    # Encode the text\n",
        "    encoded_tokens = tokenizer(text, return_tensors=\"tf\")\n",
        "\n",
        "    # Get a list of corresponding tokens\n",
        "    token_ids = list(encoded_tokens[\"input_ids\"].numpy()[0])\n",
        "\n",
        "    # Convert token ids to a one hot vector of length vocab_size\n",
        "    token_ids_tensor = tf.constant([token_ids], dtype='int32')\n",
        "    token_ids_tensor_one_hot = tf.one_hot(token_ids_tensor, vocab_size)\n",
        "\n",
        "    # Compute gradients in respect to input embeddings\n",
        "    with tf.GradientTape(watch_accessed_variables=False) as tape:\n",
        "        # Input variable\n",
        "        tape.watch(token_ids_tensor_one_hot)\n",
        "\n",
        "        # Forward pass through the models embedding layer\n",
        "        inputs_embeds = tf.matmul(token_ids_tensor_one_hot,embedding_matrix)\n",
        "\n",
        "        # Forward pass through the rest of the model and get logits\n",
        "        logits = model({\"inputs_embeds\": inputs_embeds, \"attention_mask\": encoded_tokens[\"attention_mask\"] } ).logits\n",
        "\n",
        "        # Get the prediction\n",
        "        prediction_class = tf.argmax(logits, axis=1).numpy()[0]\n",
        "\n",
        "        # Get the logit of the most likely class\n",
        "        target_logit = logits[0][prediction_class]\n",
        "\n",
        "        #  Get gradient of the input with respect to prediction class\n",
        "        gradient_non_normalized = tf.norm(\n",
        "            tape.gradient(target_logit, token_ids_tensor_one_hot),axis=2)\n",
        "\n",
        "        # Normalize the gradients\n",
        "        gradient_tensor = (\n",
        "                gradient_non_normalized /\n",
        "                tf.reduce_max(gradient_non_normalized)\n",
        "            )[0].numpy().tolist()\n",
        "\n",
        "        # Get the translated tokens\n",
        "        token_words = tokenizer.convert_ids_to_tokens(token_ids)\n",
        "\n",
        "        return gradient_tensor, token_words, prediction_class\n",
        "\n",
        "# Test our function!\n",
        "gradient_tensor, token_words, prediction_class = get_gradients(text, model, tokenizer)\n",
        "\n",
        "display(HTML(f\"<b>Text</b>: <i>{text}</i>\"))\n",
        "display(HTML(f\"<b>Prediction class</b>: <i>{model.config.id2label[int(prediction_class)]}</i>\"))\n",
        "display(HTML(f\"<b>Gradient scores</b>: <br>{dict(zip(token_words, gradient_tensor))}\"))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XMhWeGOGoFFX"
      },
      "source": [
        "Let us generate a more visual explanation, i.e., a graph that visualizes the gradient scores for each word in the input text. We can also use the gradient scores as a quantity and color range, giving us another visualization tool.\n",
        "\n",
        "> **Note:** In a text classification context, the `[CLS]` and the `[SEP]` tokens can be understood as the `start_of_the_sequence` and `end_of_the_sequence` tokens. In sequence-to-sequence modeling, `[SEP]` is the `sequence_separator` token.\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 80,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "rjG8ehLYoFFX",
        "outputId": "d07b02f8-7aa4-45cd-a789-012cdc5c6340"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "<b>Text</b>: <i>This explanation is really Bad!</i>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "text/html": [
              "<b>Integrated Gradient Scores:</b> <span style=\"color:#fc6c33\"><b>This</b></span> <span style=\"color:#f23924\"><b>explanation</b></span> <span style=\"color:#ffffcc\"><b>is</b></span> <span style=\"color:#fecc68\"><b>really</b></span> <span style=\"color:#800026\"><b>Bad</b></span> <span style=\"color:#febf5a\"><b>!</b></span>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAABkoAAAEfCAYAAAAOUZrCAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAABPB0lEQVR4nO3deXjMV///8ddksklI7Ik9KEqLtNxSW20h1lqqFa0bUUtpa4miIRKC2lUQS1tqL1/a6qJFmtJSW2sram3tJBIkIQhm5veHX+Y2TZQoHTLPx3XNJfP5nM/J+7iMWV5zzjFYLBaLAAAAAAAAAAAAHJCTvQsAAAAAAAAAAACwF4ISAAAAAAAAAADgsAhKAAAAAAAAAACAwyIoAQAAAAAAAAAADougBAAAAAAAAAAAOCyCEgAAAAAAAAAA4LAISgAAAAAAAAAAgMMiKAEAAAAAAAAAAA6LoAQAAAAAAAAAADgsghIAAADAQfn5+alr167W+xs2bJDBYNCGDRse2u8wGAwaMWLEQ+sPAAAAAB42ghIAAADADubPny+DwWC9ubu7q3z58nr77beVkJBg7/Ky5dtvv31iwpD69evLYDCoVatWmc4dP35cBoNBkyZNsh7LCI/udlu2bJlNH2azWQsXLlTjxo1VsGBBubi4qHDhwmrSpIk+/PBDpaenZ1lXcnKy3N3dZTAYdODAAevxESNG/O3vz7jVr19fktS1a1flzp1bknT+/Hk5OzurU6dOd/37uHz5snLlyqV27dpJyvzv8q+3rVu33t9fNAAAAPAEcbZ3AQAAAIAji4qKUunSpXX9+nVt2rRJs2bN0rfffqt9+/bJw8PjX63lxRdf1LVr1+Tq6pqt67799lvFxMRkGZZcu3ZNzs6P39uOb775Rjt27FC1atXuq33fvn31n//8J9PxmjVrWn++du2a2rZtq7Vr16pWrVp699135ePjo4sXL+rHH39Unz59tG3bNs2dOzdTPytWrJDBYJCvr6+WLFmi0aNHS5LatWunp556ytruypUr6t27t9q2bWsNNyTJx8cnU5+FCxdW48aN9eWXX+rq1atZ/nv6/PPPdf369UxhSsa/y7+6sxYAAAAgp3j83rEAAAAADqRZs2aqXr26JKl79+4qUKCApkyZoi+//FIdO3bM8pq0tDR5eno+9FqcnJzk7u7+UPt82P09DCVLltTly5c1cuRIffXVV/d1Td26ddW+ffu/bTNgwACtXbtWU6dOVb9+/WzODRw4UEeOHFFsbGyW1y5evFjNmzdXqVKltHTpUmtQUqVKFVWpUsXaLikpSb1791aVKlX+dqZIhtdff11r1qzRV199peDg4Eznly5dKm9vb7Vo0cLm+J3/LgEAAICcjqW3AAAAgMdIw4YNJUnHjh2T9L+llP744w81b95cefLk0euvvy7p9jJPU6dO1TPPPCN3d3f5+PioV69eunTpkk2fFotFo0ePVvHixeXh4aEGDRpo//79mX733fYo2bZtm5o3b658+fLJ09NTVapUUXR0tLW+mJgYSbJZoilDVnuU7Nq1S82aNZOXl5dy586tRo0aZVrSKWMJqJ9//lmhoaEqVKiQPD091bZtWyUmJtq0TUlJ0cGDB5WSknI/f8XKkyePBgwYoK+//lo7d+68r2vu5dSpU/r444/VtGnTTCFJhnLlyqlPnz6Zjp88eVIbN25UcHCwgoODdezYMW3evPmh1NW2bVt5enpq6dKlmc6dP39ecXFxat++vdzc3B7K7wMAAACeRAQlAAAAwGPkjz/+kCQVKFDAeuzWrVsKCgpS4cKFNWnSJL388suSpF69emnQoEGqXbu2oqOjFRISoiVLligoKEg3b960Xh8REaHhw4eratWqmjhxosqUKaMmTZooLS3tnvXExsbqxRdf1O+//65+/fpp8uTJatCggb755htrDY0bN5YkLVq0yHq7m/3796tu3bras2ePBg8erOHDh+vYsWOqX7++tm3blqn9O++8oz179igyMlK9e/fW119/rbffftumzRdffKGKFSvqiy++uOd4MvTr10/58uW7771VLl++rKSkpEw3i8UiSfruu+9kMpnua5bHX3366afy9PRUy5YtVaNGDZUtW1ZLlizJdj9Z8fT0VOvWrbV27VpdvHjR5tzy5ctlMpmswdudUlJSMo31woULD6UmAAAA4HHD0lsAAACAHWV8IH39+nX9/PPPioqKUq5cudSyZUtrm/T0dL3yyisaO3as9dimTZv08ccfa8mSJXrttdesxxs0aKCmTZtqxYoVeu2115SYmKgJEyaoRYsW+vrrr62zPYYNG6b333//b2szmUzq1auXihQpot27dytv3rzWcxkBQc2aNVW+fHnFxsbeV0gQHh6umzdvatOmTSpTpowkqXPnzqpQoYIGDx6sH3/80aZ9gQIFtG7dOmvdZrNZ06ZNU0pKiry9ve/5++7Gy8tL/fv3V2RkpHbu3Knnn3/+b9t369Yty+Pnzp2Tr6+vDh48KEl69tlnbc7fuHFDqamp1vsGg8EmBJOkJUuWqHXr1sqVK5ckqUOHDvrwww8VHR39UPZ3ef3117V06VKtXLlSPXv2tB5funSpihUrpnr16mW6JjAwMNMxNzc3Xb9+/R/XAwAAADxumFECAAAA2FFgYKAKFSqkEiVKKDg4WLlz59YXX3yhYsWK2bTr3bu3zf0VK1bI29tbjRs3tvnWf7Vq1ZQ7d26tX79ekvT999/rxo0beuedd2yWxOrfv/89a9u1a5eOHTum/v3724Qkkmz6ul8mk0nr1q1TmzZtrCGJJBUpUkSvvfaaNm3aZBMqSFLPnj1tflfdunVlMpl04sQJ67GuXbvKYrGoa9eu2aonY1bJyJEj79k2IiJCsbGxmW758+eXJGvduXPntrnu22+/VaFChay3UqVK2Zz/7bfftHfvXpv9aDp27KikpCStXbs2W+O5myZNmqhQoUI2y28dO3ZMW7duVceOHeXklPltYUxMTKaxfvfddw+lHgAAAOBxw4wSAAAAwI5iYmJUvnx5OTs7y8fHRxUqVMj0wbWzs7OKFy9uc+zIkSNKSUlR4cKFs+z3/PnzkmQNFMqVK2dzvlChQsqXL9/f1paxDNhfZ0k8qMTERF29elUVKlTIdK5ixYoym806deqUnnnmGevxkiVL2rTLqPmv+7A8CG9vb+uskl27dv3t30flypWznGWRIU+ePJKkK1eu2ByvXbu2dQP3iRMn6ueff7Y5v3jxYnl6eqpMmTI6evSoJMnd3V1+fn5asmRJpk3WH4Szs7M6dOigmTNn6syZMypWrJg1NMlq2S1JqlGjBpu5AwAAwGEQlAAAAAB2dD8fSLu5uWUKT8xmswoXLnzXvSwKFSr00Gq0J6PRmOXxjKW//ql+/frpgw8+0MiRIzV16tQH7ufpp5+WJO3bt09Vq1a1Hi9UqJA1YFm8eLHNNRaLRZ9++qnS0tJUqVKlTH2eP39eV65cyTRL5UF06tRJM2bM0Keffqp3331Xn376qSpVqiR/f/9/3DcAAADwpCMoAQAAAJ5AZcuW1ffff6/atWtb97bISsZST0eOHLFZ7ioxMfGeszLKli0r6faH/383m+J+l+EqVKiQPDw8dOjQoUznDh48KCcnJ5UoUeK++npYMmaVjBgxQl26dHngfpo1ayaj0aglS5bcdZbGX/344486ffq0oqKiVLFiRZtzly5dUs+ePbVq1aoH2iD+rwICAlS2bFktXbpUjRs31v79+zVmzJh/3C8AAACQE7BHCQAAAPAEevXVV2UymTRq1KhM527duqXk5GRJt/dAcXFx0fTp021mYdzP7Innn39epUuX1tSpU639ZbizL09PT0nK1OavjEajmjRpoi+//FLHjx+3Hk9ISNDSpUtVp04deXl53bOuv0pJSdHBgweVkpKS7WslWfdgiYqKeqDrpdtLhHXr1k3fffedZsyYkWWbv86CyVh2a9CgQWrfvr3NrUePHipXrtxdZww9iNdff127du1SZGSkDAaDXnvttYfWNwAAAPAkY0YJAAAA8ASqV6+eevXqpbFjx2r37t1q0qSJXFxcdOTIEa1YsULR0dFq3769ChUqpHfffVdjx45Vy5Yt1bx5c+3atUvfffedChYs+Le/w8nJSbNmzVKrVq3k7++vkJAQFSlSRAcPHtT+/futm41Xq1ZNktS3b18FBQXJaDQqODg4yz5Hjx6t2NhY1alTR3369JGzs7PmzJmj9PR0TZgw4YH+Lr744guFhITok08+yfaG7tLtWSX9+vX7203dN27cqOvXr2c6XqVKFVWpUkXS7fDp2LFjeuedd7Rs2TK1atVKhQsXVlJSkn7++Wd9/fXX1v1Z0tPT9dlnn6lx48Zyd3fP8ne+9NJLio6O1vnz5++6F012dOrUSVFRUfryyy9Vu3Zt+fn53bXtd999p4MHD2Y6XqtWLZuZSQAAAEBOQFACAAAAPKFmz56tatWqac6cORo6dKicnZ3l5+enTp06qXbt2tZ2o0ePlru7u2bPnq3169crICBA69atu6+NwoOCgrR+/XqNHDlSkydPltlsVtmyZdWjRw9rm3bt2lnDgcWLF8tisdw1KHnmmWe0ceNGhYWFaezYsTKbzQoICNDixYsVEBDwz/9SHlD//v01derUu85KmTZtWpbHIyMjrUGJh4eH1qxZo0WLFmnRokWaMGGCUlNTlTdvXlWtWlUzZ860Lu+1evVqJScnq1WrVnetqVWrVpo8ebKWLVumvn37/sMRSuXKldN//vMf/fLLL/dcHiwiIiLL45988glBCQAAAHIcg+Vh7YIIAAAAAAAAAADwhGGPEgAAAAAAAAAA4LAISgAAAAAAAAAAgMMiKAEAAAAAAAAAAA6LoAQAAAAAAAAAADgsghIAAAAAAAAAAOCwCEoAAAAAAAAAAIDDcrZ3AQ+D2WzW2bNnlSdPHhkMBnuXAwAAAAAAAAAA7Mhisejy5csqWrSonJz+fs5IjghKzp49qxIlSti7DAAAAAAAAAAA8Bg5deqUihcv/rdtckRQkidPHkm3B+zl5WXnagAAAAAAAAAAgD2lpqaqRIkS1vzg7+SIoCRjuS0vLy+CEgAAAAAAAAAAIEn3tV1Htjdz/+mnn9SqVSsVLVpUBoNBq1atuuc1GzZs0PPPPy83Nzc99dRTmj9/fqY2MTEx8vPzk7u7uwICArR9+/bslgYAAAAAAAAAAJAt2Q5K0tLSVLVqVcXExNxX+2PHjqlFixZq0KCBdu/erf79+6t79+5au3attc3y5csVGhqqyMhI7dy5U1WrVlVQUJDOnz+f3fIAAAAAAAAAAADum8FisVge+GKDQV988YXatGlz1zZDhgzR6tWrtW/fPuux4OBgJScna82aNZKkgIAA/ec//9GMGTMkSWazWSVKlNA777yj99577551pKamytvbWykpKSy9BQAAAAAAAACAg8tObpDtGSXZtWXLFgUGBtocCwoK0pYtWyRJN27c0I4dO2zaODk5KTAw0Nrmr9LT05WammpzAwAAAAAAAAAAyK5Hvpl7fHy8fHx8bI75+PgoNTVV165d06VLl2QymbJsc/DgwSz7HDt2rEaOHPnIagYAAAAAADnffeztCuR4D77WDADkHI98RsmjEBYWppSUFOvt1KlT9i4JAAAAAAAAAAA8gR75jBJfX18lJCTYHEtISJCXl5dy5colo9Eoo9GYZRtfX98s+3Rzc5Obm9sjqxkAAAAAAAAAADiGRz6jpGbNmoqLi7M5Fhsbq5o1a0qSXF1dVa1aNZs2ZrNZcXFx1jYAAAAAAAAAAACPQraDkitXrmj37t3avXu3JOnYsWPavXu3Tp48Ken2slidO3e2tn/zzTf1559/avDgwTp48KBmzpyp//u//9OAAQOsbUJDQ/XRRx9pwYIFOnDggHr37q20tDSFhIT8w+EBAAAAAAAAAADcXbaX3vr111/VoEED6/3Q0FBJUpcuXTR//nydO3fOGppIUunSpbV69WoNGDBA0dHRKl68uD7++GMFBQVZ23To0EGJiYmKiIhQfHy8/P39tWbNmkwbvAMAAAAAAAAAADxMBovFYrF3Ef9UamqqvL29lZKSIi8vL3uXAwAAAAAAngAGg70rAOzvyf9kEACylp3c4JHvUQIAAAAAAAAAAPC4IigBAAAAAAAAAAAOi6AEAAAAAAAAAAA4LIISAAAAAAAAAADgsAhKAAAAAAAAAACAwyIoAQAAAAAAAAAADougBAAAAAAAAAAAOCyCEgAAAAAAAAAA4LAISgAAAAAAAAAAgMMiKAEAAAAAAAAAAA6LoAQAAAAAAAAAADgsghIAAAAAAAAAAOCwCEoAAAAAAAAAAIDDIigBAAAAAAAAAAAOi6AEAAAAAAAAAAA4LIISAAAAAAAAAADgsAhKAAAAAAAAAACAwyIoAQAAAAAAAAAADougBAAAAAAAAAAAOCyCEgAAAAAAAAAA4LAISgAAAAAAAAAAgMMiKAEAAAAAAAAAAA6LoAQAAAAAAAAAADgsghIAAAAAAAAAAOCwCEoAAAAAAAAAAIDDcrZ3AQCA+1Ozpr0rAOxvyxZ7VwAAAAAAAHKaB5pREhMTIz8/P7m7uysgIEDbt2+/a9v69evLYDBkurVo0cLapmvXrpnON23a9EFKAwAAAAAAAAAAuG/ZnlGyfPlyhYaGavbs2QoICNDUqVMVFBSkQ4cOqXDhwpnaf/7557px44b1/oULF1S1alW98sorNu2aNm2qTz75xHrfzc0tu6UBAAAAAAAAAABkS7ZnlEyZMkU9evRQSEiIKlWqpNmzZ8vDw0Pz5s3Lsn3+/Pnl6+trvcXGxsrDwyNTUOLm5mbTLl++fA82IgAAAAAAAAAAgPuUraDkxo0b2rFjhwIDA//XgZOTAgMDteU+Fw2fO3eugoOD5enpaXN8w4YNKly4sCpUqKDevXvrwoULd+0jPT1dqampNjcAAAAAAAAAAIDsylZQkpSUJJPJJB8fH5vjPj4+io+Pv+f127dv1759+9S9e3eb402bNtXChQsVFxen8ePH68cff1SzZs1kMpmy7Gfs2LHy9va23kqUKJGdYQAAAAAAAAAAAEh6gD1K/om5c+eqcuXKqlGjhs3x4OBg68+VK1dWlSpVVLZsWW3YsEGNGjXK1E9YWJhCQ0Ot91NTUwlLAAAAAAAAAABAtmVrRknBggVlNBqVkJBgczwhIUG+vr5/e21aWpqWLVumN954456/p0yZMipYsKCOHj2a5Xk3Nzd5eXnZ3AAAAAAAAAAAALIrW0GJq6urqlWrpri4OOsxs9msuLg41axZ82+vXbFihdLT09WpU6d7/p7Tp0/rwoULKlKkSHbKAwAAAAAAAAAAyJZsBSWSFBoaqo8++kgLFizQgQMH1Lt3b6WlpSkkJESS1LlzZ4WFhWW6bu7cuWrTpo0KFChgc/zKlSsaNGiQtm7dquPHjysuLk6tW7fWU089paCgoAccFgAAAAAAAAAAwL1le4+SDh06KDExUREREYqPj5e/v7/WrFlj3eD95MmTcnKyzV8OHTqkTZs2ad26dZn6MxqN+u2337RgwQIlJyeraNGiatKkiUaNGiU3N7cHHBbu1KKFvSsA7G/1antXAAAAAAAAAOBxZLBYLBZ7F/FPpaamytvbWykpKexXkgWCEiBnBCX3WOEQcAhbtti7AgAAkJMYDPauALC/J/+TQQDIWnZyg2wvvQUAAAAAAAAAAJBTEJQAAAAAAAAAAACHRVACAAAAAAAAAAAcFkEJAAAAAAAAAABwWAQlAAAAAAAAAADAYRGUAAAAAAAAAAAAh0VQAgAAAAAAAAAAHBZBCQAAAAAAAAAAcFgEJQAAAAAAAAAAwGERlAAAAAAAAAAAAIdFUAIAAAAAAAAAABwWQQkAAAAAAAAAAHBYBCUAAAAAAAAAAMBhEZQAAAAAAAAAAACHRVACAAAAAAAAAAAcFkEJAAAAAAAAAABwWAQlAAAAAAAAAADAYRGUAAAAAAAAAAAAh0VQAgAAAAAAAAAAHBZBCQAAAAAAAAAAcFgEJQAAAAAAAAAAwGE527sAAAAAAAAAAHhQoaH2rgCwvylT7F3Bk40ZJQAAAAAAAAAAwGERlAAAAAAAAAAAAIdFUAIAAAAAAAAAABzWA+1REhMTo4kTJyo+Pl5Vq1bV9OnTVaNGjSzbzp8/XyEhITbH3NzcdP36det9i8WiyMhIffTRR0pOTlbt2rU1a9YslStX7kHKAwAAAIBHKizM3hUA9jd2rL0rAAAAeDiyPaNk+fLlCg0NVWRkpHbu3KmqVasqKChI58+fv+s1Xl5eOnfunPV24sQJm/MTJkzQtGnTNHv2bG3btk2enp4KCgqyCVMAAAAAAAAAAAAetmwHJVOmTFGPHj0UEhKiSpUqafbs2fLw8NC8efPueo3BYJCvr6/15uPjYz1nsVg0depUhYeHq3Xr1qpSpYoWLlyos2fPatWqVQ80KAAAAAAAAAAAgPuRraDkxo0b2rFjhwIDA//XgZOTAgMDtWXLlrted+XKFZUqVUolSpRQ69attX//fuu5Y8eOKT4+3qZPb29vBQQE3LXP9PR0paam2twAAAAAAAAAAACyK1tBSVJSkkwmk82MEEny8fFRfHx8ltdUqFBB8+bN05dffqnFixfLbDarVq1aOn36tCRZr8tOn2PHjpW3t7f1VqJEiewMAwAAAAAAAAAAQNIDLL2VXTVr1lTnzp3l7++vevXq6fPPP1ehQoU0Z86cB+4zLCxMKSkp1tupU6ceYsUAAAAAAAAAAMBRZCsoKViwoIxGoxISEmyOJyQkyNfX9776cHFx0XPPPaejR49KkvW67PTp5uYmLy8vmxsAAAAAAAAAAEB2ZSsocXV1VbVq1RQXF2c9ZjabFRcXp5o1a95XHyaTSXv37lWRIkUkSaVLl5avr69Nn6mpqdq2bdt99wkAAAAAAAAAAPAgnLN7QWhoqLp06aLq1aurRo0amjp1qtLS0hQSEiJJ6ty5s4oVK6axY8dKkqKiovTCCy/oqaeeUnJysiZOnKgTJ06oe/fukiSDwaD+/ftr9OjRKleunEqXLq3hw4eraNGiatOmzcMbKQAAAAAAAAAAwF9kOyjp0KGDEhMTFRERofj4ePn7+2vNmjXWzdhPnjwpJ6f/TVS5dOmSevToofj4eOXLl0/VqlXT5s2bValSJWubwYMHKy0tTT179lRycrLq1KmjNWvWyN3d/SEMEQAAAAAAAAAAIGsGi8VisXcR/1Rqaqq8vb2VkpLCfiVZaNHC3hUA9rd6tb0r+OdYjRCQtmyxdwUAcFtYmL0rAOzv/y8k8UQzGOxdAWB/T/4ng1JoqL0rAOxvyhR7V/D4yU5ukK09SgAAAAAAAAAAAHISghIAAAAAAAAAAOCwCEoAAAAAAAAAAIDDIigBAAAAAAAAAAAOi6AEAAAAAAAAAAA4LIISAAAAAAAAAADgsAhKAAAAAAAAAACAwyIoAQAAAAAAAAAADougBAAAAAAAAAAAOCyCEgAAAAAAAAAA4LAISgAAAAAAAAAAgMMiKAEAAAAAAAAAAA6LoAQAAAAAAAAAADgsghIAAAAAAAAAAOCwCEoAAAAAAAAAAIDDIigBAAAAAAAAAAAOi6AEAAAAAAAAAAA4LIISAAAAAAAAAADgsJztXQAAAIAjWbHC3hUA9vfKK/auAAAAAAD+hxklAAAAAAAAAADAYRGUAAAAAAAAAAAAh0VQAgAAAAAAAAAAHBZBCQAAAAAAAAAAcFgEJQAAAAAAAAAAwGERlAAAAAAAAAAAAIdFUAIAAAAAAAAAABzWAwUlMTEx8vPzk7u7uwICArR9+/a7tv3oo49Ut25d5cuXT/ny5VNgYGCm9l27dpXBYLC5NW3a9EFKAwAAAAAAAAAAuG/ZDkqWL1+u0NBQRUZGaufOnapataqCgoJ0/vz5LNtv2LBBHTt21Pr167VlyxaVKFFCTZo00ZkzZ2zaNW3aVOfOnbPePv300wcbEQAAAAAAAAAAwH3KdlAyZcoU9ejRQyEhIapUqZJmz54tDw8PzZs3L8v2S5YsUZ8+feTv76+nn35aH3/8scxms+Li4mzaubm5ydfX13rLly/fg40IAAAAAAAAAADgPmUrKLlx44Z27NihwMDA/3Xg5KTAwEBt2bLlvvq4evWqbt68qfz589sc37BhgwoXLqwKFSqod+/eunDhwl37SE9PV2pqqs0NAAAAAAAAAAAgu7IVlCQlJclkMsnHx8fmuI+Pj+Lj4++rjyFDhqho0aI2YUvTpk21cOFCxcXFafz48frxxx/VrFkzmUymLPsYO3asvL29rbcSJUpkZxgAAAAAAAAAAACSJOd/85eNGzdOy5Yt04YNG+Tu7m49HhwcbP25cuXKqlKlisqWLasNGzaoUaNGmfoJCwtTaGio9X5qaiphCQAAAAAAAAAAyLZszSgpWLCgjEajEhISbI4nJCTI19f3b6+dNGmSxo0bp3Xr1qlKlSp/27ZMmTIqWLCgjh49muV5Nzc3eXl52dwAAAAAAAAAAACyK1tBiaurq6pVq2azEXvGxuw1a9a863UTJkzQqFGjtGbNGlWvXv2ev+f06dO6cOGCihQpkp3yAAAAAAAAAAAAsiVbQYkkhYaG6qOPPtKCBQt04MAB9e7dW2lpaQoJCZEkde7cWWFhYdb248eP1/DhwzVv3jz5+fkpPj5e8fHxunLliiTpypUrGjRokLZu3arjx48rLi5OrVu31lNPPaWgoKCHNEwAAAAAAAAAAIDMsr1HSYcOHZSYmKiIiAjFx8fL399fa9assW7wfvLkSTk5/S9/mTVrlm7cuKH27dvb9BMZGakRI0bIaDTqt99+04IFC5ScnKyiRYuqSZMmGjVqlNzc3P7h8AAAAAAAAAAAAO7ugTZzf/vtt/X2229neW7Dhg02948fP/63feXKlUtr1659kDIAAAAAAAAAAAD+kWwvvQUAAAAAAAAAAJBTEJQAAAAAAAAAAACHRVACAAAAAAAAAAAcFkEJAAAAAAAAAABwWAQlAAAAAAAAAADAYRGUAAAAAAAAAAAAh0VQAgAAAAAAAAAAHBZBCQAAAAAAAAAAcFgEJQAAAAAAAAAAwGERlAAAAAAAAAAAAIdFUAIAAAAAAAAAABwWQQkAAAAAAAAAAHBYBCUAAAAAAAAAAMBhEZQAAAAAAAAAAACHRVACAAAAAAAAAAAcFkEJAAAAAAAAAABwWAQlAAAAAAAAAADAYRGUAAAAAAAAAAAAh0VQAgAAAAAAAAAAHBZBCQAAAAAAAAAAcFgEJQAAAAAAAAAAwGERlAAAAAAAAAAAAIdFUAIAAAAAAAAAABwWQQkAAAAAAAAAAHBYBCUAAAAAAAAAAMBhEZQAAAAAAAAAAACH9UBBSUxMjPz8/OTu7q6AgABt3779b9uvWLFCTz/9tNzd3VW5cmV9++23NuctFosiIiJUpEgR5cqVS4GBgTpy5MiDlAYAAAAAAAAAAHDfsh2ULF++XKGhoYqMjNTOnTtVtWpVBQUF6fz581m237x5szp27Kg33nhDu3btUps2bdSmTRvt27fP2mbChAmaNm2aZs+erW3btsnT01NBQUG6fv36g48MAAAAAAAAAADgHrIdlEyZMkU9evRQSEiIKlWqpNmzZ8vDw0Pz5s3Lsn10dLSaNm2qQYMGqWLFiho1apSef/55zZgxQ9Lt2SRTp05VeHi4WrdurSpVqmjhwoU6e/asVq1a9Y8GBwAAAAAAAAAA8Hecs9P4xo0b2rFjh8LCwqzHnJycFBgYqC1btmR5zZYtWxQaGmpzLCgoyBqCHDt2TPHx8QoMDLSe9/b2VkBAgLZs2aLg4OBMfaanpys9Pd16PyUlRZKUmpqaneE4jJs37V0BYH854b+HW7fsXQFgfznhsXz1qr0rAOwvJzyW73g7AjisnPBYBpAzHss8LwM547H8sGXkBRaL5Z5tsxWUJCUlyWQyycfHx+a4j4+PDh48mOU18fHxWbaPj4+3ns84drc2fzV27FiNHDky0/ESJUrc30AAOBxvb3tXAOBh4LEM5Axdu9q7AgAPwwcf2LsCAA8Dr7GBnGHmTHtX8Pi6fPmyvO/xn122gpLHRVhYmM0sFbPZrIsXL6pAgQIyGAx2rAzILDU1VSVKlNCpU6fk5eVl73IAPCAey0DOwGMZyBl4LAM5A49lIGfgsYzHlcVi0eXLl1W0aNF7ts1WUFKwYEEZjUYlJCTYHE9ISJCvr2+W1/j6+v5t+4w/ExISVKRIEZs2/v7+Wfbp5uYmNzc3m2N58+bNzlCAf52XlxdPFkAOwGMZyBl4LAM5A49lIGfgsQzkDDyW8Ti610ySDNnazN3V1VXVqlVTXFyc9ZjZbFZcXJxq1qyZ5TU1a9a0aS9JsbGx1valS5eWr6+vTZvU1FRt27btrn0CAAAAAAAAAAA8DNleeis0NFRdunRR9erVVaNGDU2dOlVpaWkKCQmRJHXu3FnFihXT2LFjJUn9+vVTvXr1NHnyZLVo0ULLli3Tr7/+qg8//FCSZDAY1L9/f40ePVrlypVT6dKlNXz4cBUtWlRt2rR5eCMFAAAAAAAAAAD4i2wHJR06dFBiYqIiIiIUHx8vf39/rVmzxroZ+8mTJ+Xk9L+JKrVq1dLSpUsVHh6uoUOHqly5clq1apWeffZZa5vBgwcrLS1NPXv2VHJysurUqaM1a9bI3d39IQwRsC83NzdFRkZmWi4OwJOFxzKQM/BYBnIGHstAzsBjGcgZeCwjJzBYLBaLvYsAAAAAAAAAAACwh2ztUQIAAAAAAAAAAJCTEJQAAAAAAAAAAACHRVACAAAAAAAAAAAcFkEJAAAAAAAAAABwWAQlAAAAAAAAAADAYRGUAAAAAAAAAAAAh0VQAtiJ2WyWJFksFps/AQAAADw8Ga+7ATz+Jk2apJ07d9q7DACAAyIoAezAZDLJyclJZ86c0eLFi3XkyBEZDAZ7lwUAQI711w9K+YIC4BgsFoucnG6/7T1x4oT1GIDHzw8//KDBgwdr3Lhx2rt3r73LAfA3Dhw4YO8SgIeOoAT4l5lMJhmNRh04cECBgYGaPHmyVqxYYe+yADxGTCZTlsf5RizwYDK+oHD27Fl99dVXksQXFAAHkfFYDw4O1rBhw2yOAXi8NGzYUPPmzdPKlSsVFRVFWAI8pvbs2aNnnnlGERER9i4FeKic7V0A4AjMZrP1m2xGo1GHDx9Ww4YNVaVKFb3zzjtq2bKlTXuLxcIbOMBBZYSpkhQdHa1r166pYMGCCgkJkdFo5P8HIJvu/IJCcHCw/vzzT02ePFk9e/a0d2kA/kXFixfXhx9+qB07dqhatWr2LgfAX2S8Z+7atatMJpN69Oghi8Wi8PBw+fv727s8AHcoXbq03nrrLY0bN06urq4KDw+3d0nAQ0FQAjxCP/zwgxo2bGgNSSQpPT1dI0aMUOHChTVq1CjVqFFD0u0XhlevXtW1a9dUqFAhe5UMwM4yQpIWLVro+++/l5OTkywWi2bNmqXVq1fL19fXzhUCTw6LxWL9gkL9+vX13HPPKSoqSq1bt87UjgASyNmaN2+uuXPn6ptvvlG1atVsvpgAwP6cnJx08+ZNubi46I033pCrq6u6du0qT09P9e/fX88995y9SwTw/3l5een999+Xm5ubdVYJYQlyApbeAh6R8ePHKzAwUPPnz7c5fuPGDe3fv19VqlSxhiQbN25URESEnn/+eQUEBOjDDz+UxPrJgCO5deuW9ec5c+bo999/10cffaTdu3crOjpaycnJqlu3rg4dOmTHKoHHW3BwsPr27Wu9bzAYdPXqVfXv319ly5bV+++/bxOSpKenW9sBeLJlvG7+6/KVGfcbNmyoNm3aaM6cOUpOTrbO0gTweDCZTHJxcdHvv/+u4OBgbd68Wc7Ozlq6dKkmTJig3377zd4lArhDnjx5FBERodDQUEVERGj06NH2Lgn4xwhKgEekRo0aateunZ599lmb487OzipTpozWr1+v2NhYRUZGKjg4WPPmzdOzzz4rPz8/9enTRz/88AMf3AAOxNn59iTP9evXy2g0qm7dunrllVdUoUIFde3aVXPmzJGrq6tatmypw4cP27la4PFz7do1OTk5acGCBTpz5oz1+PXr13Xo0CEFBATo+eeflyRt3rxZo0aNUqVKldSoUSMtWrRI169ft1fpAP6Bn3/+WSdOnJDBYJDZbLbOEjly5Iik2zM1M76M0K1bN5lMJo0bN04mk4nX2sBjxGg06s8//1T9+vV17tw5lStXTjNnzlTnzp21fPlyjRo1irAEeMx4eXnZhCXvv/++vUsC/hGCEuARadCggRYsWKDq1avr0KFDWrhwoSQpV65cCg4OVt68eRUUFKRJkybp1Vdf1fLly/X5559r+vTpyp07t9avX2/nEQD4t02dOlWNGjXS4MGDVbJkSeXKlUsmk0lubm5q0KCBoqOj5eLioubNm1s/AAJwW65cuTRnzhzt3btXxYoVU1JSkiTJ3d1dBoNBBw4c0K+//qrRo0erU6dOmjFjhipWrKj4+HiFhYXpxIkTdh4BgOzaunWr6tWrpxEjRuj06dPW5W47duyomjVrqnv37jp9+rQ1KKlWrZr8/f21evVqXb16VRIzuIHHgcVikcViUUxMjCRpzJgxCg0N1RtvvKG5c+dqxowZ+uyzzzR69Gjt2bPHztUCuFNGWPLuu+8qPDxcw4YNs3dJwAMjKAEeIQ8PD12/fl2RkZHq2rWrPvroI0lShw4d9Nlnn2nNmjXau3evxo0bp7p160qSTp8+rdy5c6tChQr2LB2AHdSvX1/dunXT9evXtWvXLt26dUtGo9H6DdkGDRpo+vTp8vDwUKNGjXTgwAF7lww8VvLkyaOSJUsqISFBFSpU0JAhQ+Th4aExY8bol19+Uc2aNTV69Gg1adJEy5Yt0zfffKOvvvpKZ8+e1Q8//GDv8gHcp4xw44UXXlDPnj317bffatSoUTp16pQk6a233lKzZs20du1aVa5cWf3799emTZvk4eGh8ePH6/jx49YPZJlVAtifwWCQwWDQhQsX5ObmZt2PxGw2S5L69OmjiRMnauXKlZo0aZJ27dplz3IB/IWXl5fCw8PVq1cvFS5c2N7lAA+MzdyBR8hgMMjd3V3vvPOOrl27prCwMN26dUu9e/dWhQoVMoUh27Zt07Rp0+Tp6al69erZqWoA/4Y7N4/O+Nnf319vv/22bt68qUWLFmnChAkaOnSonJycrGFJ/fr1NXnyZHXp0kXbtm1TxYoV7TwS4PHj4eGhV155RR988IHy58+vQYMGqWrVqlq/fr3q1q2r0qVLy9PTU5K0e/duFShQQKVLl7Zz1QDu16VLl5Q/f35J0syZM+Xi4qJFixZJkoYOHao6deqoevXqSkhI0JQpU7RmzRrNnz9fnTt31gsvvKDmzZvru+++06uvvqoyZcrYcygA7pA7d24lJyfrzJkzKl++vM3eQx06dNCcOXO0atUqXbp0SePHj9czzzxj54oBZPDy8lJ0dLRcXV117Ngxbd68Wa+//rq9ywKyhaAEeMiy+vCzdu3aGjp0qKKiojR8+HAZjUb17NnTes21a9c0YcIEff311zp16pRiY2NVokQJew0BwCNmMpmsa6hfvXpVly5dko+Pj5ydneXv76/Q0FCZTCaFh4fLyclJ7733nk1Y0rBhQ23bto3/J4C7yJMnjyZOnChPT0+FhYXp5s2bCg8P19NPP23T7pdfftH8+fPl6+urKlWq2KlaANmxatUqtWvXTr/++qt136Ho6GhJsoYlYWFh8vPzU6lSpRQdHa29e/fqhx9+0IQJExQXF6djx47J3d1dBw8eVJkyZWQ2m63LdgF49O58z3zn/e7du2vhwoWKiorS4sWLZTQadfPmTbm4uKh48eIqVaqUihQpop9++kne3t52HAGArLi6uspkMmnhwoUaOXKkEhMT1b9/f3uXBdw3ghLgH0pPT5ebm5sk6datW3J2dlZSUpKSk5NlNptVvnx5SVJAQICGDx+uUaNGaejQoZJkDUuWL1+ur7/+WiVKlNCSJUtYdgvIwe4MSfr27auffvpJ+/bt03/+8x81atRIo0ePVtWqVTV48GBZLBbr/xd/DUsyQhI+3AGylidPHkVGRkqSIiIiJEnh4eGSbn8gExERoR9//FEHDhzQDz/8oKJFi9qtVgD3z83NTRUrVlRgYKB++OEH+fv7S8ocloSHh1ufKytXrqzKlSurZcuW+umnn7R48WKtX79ew4cPV40aNVSwYEG7jAVwRBmvha9evaorV67IxcVFefLkkbOzs0qUKKGePXtqypQpcnNz09y5c+Xi4iJJ2rRpk5KSkrRq1SoVKlRIHh4edh4JgKwYjUZ1795dJ0+eVNOmTe1dDpAtBgu71wEP7L///a9q1aqlzp07W5fw2L9/v15++WWdOXNGZrNZb775pvr37299o7Z161aNGjVK27Zt0/vvv28NS37//XcVL15cXl5edhsPgH/PSy+9pG3btqlevXoqV66cvv76a+3bt09BQUH69ttvZTAYtGfPHk2cOFFLly7VqFGj2BgPeACpqamKiorSlClTFBUVpfDwcC1cuFDjxo1TiRIlNHXqVJawA54w33//vYYMGaLDhw9r48aN1rBEkvr166dFixbplVdesYYlGfscZHyxwGKxKDw8XB9//LE+/PBDtW7dmi8eAP+CjJDk0KFDCg0N1f79++Xp6akXXnhB48ePV8GCBXX06FFNmDBBH3/8sWrVqqVWrVrp5s2bWrNmjU6cOKHNmzczqxp4Atz5BUHgSUFQAjyghIQEtWjRQkePHtW0adMUHBysK1euKCAgQEWLFlXt2rV19epVzZo1S0FBQRoxYoR1eYCMsGTnzp0aNGiQQkND7TwaAI9CWlqaNUS906JFi9S7d29Nnz5dr7zyinLnzq2zZ88qJiZGEyZMUFBQkL755htJ0r59+zRq1CitWLFC27dvV/Xq1f/tYQBPvDvDksmTJ6tbt246deqUihUrpnz58tm7PAD36c7lemJjY/Xee+/dd1iSce2dgUjlypVVqlQp63MugEcn47F38OBBvfjii/Lx8VHdunWVmpqqjRs3ysPDQz/++KMKFy6sEydO6LvvvtO0adN05MgR5cqVS6VKldLSpUtVuXJlew8FAJBDEZQAD+jSpUtKSkpSv3799PPPP2vWrFkqWLCghg0bpqlTp6p27dqSpG+++UYdO3ZUrVq1NHbsWGtYsm3bNr377rs6e/asduzYIW9vb5t1WgE82RYsWKC4uDiNGDEi00ax4eHhmj59uo4dO6b8+fNbl+27cOGCxo8fr0mTJmncuHEaPHiwJGnv3r06c+YMU5eBfyA1NVWjR4/WpEmTNGnSJL6kADyhHiQsGT58uIoXL55pX4Q33nhD27Zt06ZNm5Q3b95/eSSA40lMTFTbtm3l6empUaNGqUaNGpKktm3b6ssvv5Sfn582b94sX19fmUwm3bhxQ7t27ZKXl5d8fHxUqFAhO48AAJCTMbcYeACzZs1ScHCwPD09NX36dAUEBKhfv35asmSJChcubA1JTCaTWrZsqc8++0ybN29WWFiYdu7cKen2niUffPCBNmzYoLx58xKSADnMmTNntHjxYh0+fDjTOScnJ6WlpSk5OVnS7XVcLRaLChQooLffflvFixfX5s2bre0rV65sDUkylg8BkD1eXl4aNmyYhg0bRugIPMEMBoNu3bolSWrcuLHGjRun8uXLq27dutq9e7e1XXR0tP773//qiy++UGRkpE6cOGHzejs+Pl67d++Wp6endQ8EAA+PxWKR2Wy2vna1WCzasmWLjh07pl69ellDkqFDh2r16tXq2bOn0tLSVKdOHSUlJcloNCpXrlyqVauWnn32WUISAMAjR1ACZNOCBQv01ltvyc/PTx4eHipbtqxmzZqlgIAALVq0SAcOHNDZs2cl/e/DzyZNmmjlypXavHmzhg8fru3bt0uSqlevzvqqQA41dOhQbd261fqBbFJSkvXcU089JbPZrKVLl+ry5csyGAzWN5MlS5ZUjRo1tGfPHqWlpWXql/XTgQfn7e2tkSNHqlKlSvYuBUA2mEwmm/vOzs7Wnxs3bqyxY8feNSx55ZVX9Mknn+jo0aPW4xaLxdpu9uzZWS6TCeDBbdiwQe+9955atWql0aNH69ChQzIYDKpTp45atGihtm3bSpImTZqkCRMmaNasWdalMf/880/Vq1dP8fHxdh4FAMDRsPQWkA0LFixQSEiI+vfvr0GDBqlIkSLWc0eOHFF4eLhWrFihmJgYmw3eM8TGxiooKEht27bV0qVL5ebm9m8PAYAdtG3bVh4eHhozZoz8/PwkSa1atdLGjRs1bdo0tW7dWt7e3pKkU6dOqX379ipWrJg+++wzZpsBABzanZvBzpgxQ0ePHtWVK1f05ptv6rnnnrOeW7dunYYOHaqDBw9q48aNeu6556x9/PTTT3rxxRdt+r127ZquXr2qAgUK/HuDARzAokWLNHDgQHl4eMhsNuv06dN69dVXNX36dJtZIb/++qtatmypN954QwMHDlT+/Pn1xx9/KCgoSKdOnVL58uW1e/duOTk58XoYAPCv4GupwH36v//7P4WEhOiNN97QxIkTrSFJRtZYrlw5jRkzRk2aNNF7772nlStXZvo2eOPGjfX9999rzJgxhCRADvLX7xxcv37d5n6VKlW0bNkyTZo0SX/++aek2x/2VK5cWW+//bbGjBmjffv26ZdfflFMTIx2796t1q1b86YQAODQzGazNQhp1aqVIiIitGLFCsXGxqpmzZqaMmWKTp8+LUlq0qSJ3n//fT399NNq2LChdblbSdaQ5M4lgHLlykVIAjxk8+fPV5cuXfT6669r1apVOnnypAYOHKjPP//cuuRshrNnzyolJUUvvfSS8ufPL4vFoq+//lq5c+fWokWL9OWXX8poNPJ6GADwryEoAe7D/PnzFRwcLOn2t9UuXLggSbp586bNC7ennnpKMTExeuGFF9S3b98sw5KGDRvq6aef/veKB/DIGQwGpaenW5fVc3d3l3R70/arV69q5MiRmjx5smbOnKlJkybp1KlTKlWqlP7v//5PjRs31qRJk1SlShXVrl1bs2bN0qhRo9SlSxdJmUMYAAAcRcZyk927d9fu3bv1wQcf6JdfftHhw4fVqlUrjRkzRp988ol12dsmTZpo3LhxKlGihKpXr67ExESb59GM/vjgFXj4FixYoDfeeEMDBgzQ4MGD5e/vL+l2yFmyZEldvHhRe/fu1fHjxyVJefLkUXp6unbs2CFJ2r59u+Li4lStWjW1adNGZcqUsdNIAACOiqW3gHvIWG6rX79+KlWqlMaOHStXV1dt2bJFxYsX161bt2zWSZakP/74Q3369NG2bds0ffp0tW3bVrlz57bTCAD8G+Li4jRs2DDVqVNHkyZNUrNmzbRlyxatXbtWAQEBkqQPPvhAAwcOVK9evRQWFqaSJUtKkr777judPn1auXLlUpkyZVSrVi1Jt7/5yp4kAABHtn79er399tsKCQlRjx495O3trfXr16t169YqWLCgTp48qfDwcPXo0UPFihWTJK1evVqpqanq2LGjnasHHMP+/ftVuXJllS9fXqtXr1bZsmWt56KiojRixAgVKlRIiYmJyp07tyZOnKgGDRooLCxMq1atUpkyZXTt2jVdv35dGzZs0LPPPmvH0QAAHJXzvZsAjmvWrFl66623FBoaqmHDhilfvnzy8vJSWFiYatasqa1bt6pYsWKZwpKyZctq5syZ6tu3r7p06SJnZ2cFBwfz7TUgBytbtqx8fX01c+ZMff7557p69aqWL19u80ZvwIABkqSBAwfKYDDo3XffVZkyZdSsWbNM/RGSAAAgXb16VeXLl1f79u3l7e2tnTt3qmXLlmrbtq3GjRun9957TxMnTpS7u7tef/11lShRQi1atLBez/Mp8OgVK1ZMAwYMUExMjD788ENFRkbKw8NDc+fOVVRUlLp06aK2bdvq/Pnzmj17tt566y19+OGHmjFjhmrWrKmffvpJxYsXV79+/VShQgV7DwcA4KCYUQL8jfXr1+vTTz/VyJEjrXuSmEwmzZ8/X0OHDpWrq+tdwxJJOnz4sMLCwvT+++/zgg9wAKdOnVKNGjWUkJCgl156SatWrZJ0e5k+FxcXa7uMmSVvvfWWBg4caN3gHQAA3GaxWGQwGHTr1i2dPn1afn5+Onv2rOrXr6+nn35a06dPV6lSpbRs2TK99tprMhqN6t+/v0aOHCkPDw97lw84nNTUVI0aNUqTJ0/WsGHDVKxYMfXp00fDhg3TgAEDlD9/fknShg0b1Lt3bx09elS///67ypUrJ4vFIpPJlOn9NAAA/yaehYC/0aBBA9WuXVuurq6SbockRqNR3bp1kyQNHTpUL7zwwl3DkvLly2vZsmU2H5ACyJnMZrMOHjwoHx8f+fn5KTY2VoMGDdLEiRPl4uJi8//DgAED5OTkpAEDBujy5cuaMWMGy/MBABxaxuvsDBkzsZ2dna1fKNi/f79SUlLUrVs3lSpVSpLk6uqqhg0bqnTp0ipWrBghCWAnXl5eGj58uCwWi8aNGyeTyaTBgwdr8ODBypMnjzX8rF+/vrp166YhQ4bo8OHDKleunAwGAyEJAMDueCYC7iEjJJEko9Fonb5/t7Dkr2/yCEmAnCvjDZ90e4PYxo0bKzY2VklJSRo+fLhmzpwps9msyZMny9nZ2eb/h379+iklJUXe3t6EJAAAh3bn8+P48eO1f/9+nT17Vm3atFHDhg1VqVIlSdKRI0eUmJioAgUKSJISExO1Zs0aeXt7a86cOSyxBdhZRliSK1cujR8/XlevXs3y/fCNGzfk7e2tggUL2qFKAACyxtJbwAPICEssFovmzZtnXYZr27ZtKlq0aKawBEDOc+fj3Gw269q1a/L09LSeP3LkiIYMGaK1a9fqzTff1OTJkyVJFy9e1A8//KBmzZrZtL8zdAEAwFHc+fzXokULbd++XcWLF1euXLn0+++/y8/PT5MnT1ajRo0UHx+v6tWrq3Tp0mrUqJFOnjypTz/9VNOmTVOPHj3sPBIAGVJSUjR69GhNnjxZ7777roYPH648efJIkvbu3Wt9vK5evdoafAIAYG/MKAEegJOTU6aZJZGRkSpXrpyOHj1q3c8EQM50Z0gyYsQIbd26VdevX1ft2rU1cuRIOTs7q1y5cho/frwkafbs2TKbzerXr58+/fRTDRs2TEuXLlVwcLC1T0ISAIAjynj+GzJkiHbs2KGYmBg1atRIBQoUUL9+/TR9+nRt3rxZtWrVkq+vr5YsWaKePXtq3LhxKly4sMaOHWv90JUvHQCPB29vbw0fPlySNGnSJBkMBo0YMULHjx/X0KFD9fvvv2vLli2EJACAxwozSoB/4M6ZJTNnzlR0dLRWr16tcuXK2bs0AI9IxuNeklq2bKlNmzbpmWeeUa5cubR582a9+OKL+uCDD1ShQgU5OTnp6NGjGj58uJYvX678+fPr8uXLioiI0LBhw+w8EgAA7OevM7Dr1q2rcuXKacaMGfLw8NDGjRvVvHlzvfrqq4qIiLDuSSJJ165d06lTp+Tm5mY9fufzM4DHw50bvIeEhOjcuXP66aef9PPPP6tq1ar2Lg8AABsEJcA/dGdYkpqaKm9vb3uXBOBf0LdvX3311VcaN26c2rZtKzc3N40fP15hYWFq3Lixpk6dag1Lzpw5o9jYWB0+fFj+/v569dVXJfGhDgDAMd35/DdixAh17NhRbdq0UefOnRUWFqa4uDi1atVK7dq108SJE62ztdesWaOKFSvahCYSM0mAx1lqaqrGjh2r8ePHy2g06tdffyUkAQA8llh6C/iHMkISg8FASAI4iP379+vXX39V586d1bRpU7m5uWnjxo0aPXq0GjRooJ07d+rNN9/UzJkzVbFiRRUrVkxdu3a16YOQBADgiCwWi/X5r1OnToqNjZWfn58MBoP+/PNP7d69W61atVLbtm1tQpJvvvlGAwcO1CeffJIpKCEkAR5fXl5eGjJkiPLkyaN27drp6aeftndJAABkiU9ogIeAN2eAY3FxcVHhwoXVqVMn5c2bV3v37lWzZs300ksv6ZNPPtGoUaO0ceNGhYeH68CBA1n2QUgCAHA0ZrPZ+ro5NTVVhw4d0ogRI9S1a1cNGjRI8+bN0/PPP6///ve/mjRpkjUkOX36tNasWaPcuXMrf/789hwCgAeQN29evffee4QkAIDHGp/SAADwN8xms/XnjNUqy5cvrwULFqh8+fK6cOGCunfvrrp162r06NEqWbKkWrVqJR8fH3355Zfq0KGDLl26ZK/yAQB4bGR8SaB169bq1auXjEajXnrpJUlSQECA2rZtKycnJxUpUkSFCxeWJB05ckRz5szRJ598ol69evFBK/CE4ktCAIDHHUtvAQBwF7du3ZKz8+2nyrS0NJlMJnl5eUmS9c/k5GSdOHFCwcHBKl26tCTp+PHjKly4sCZOnCiTyaR8+fLZZwAAADxmkpOTde3aNa1bt06enp46e/asihYtqkqVKmngwIG6deuWoqKitHLlSuXPn1+JiYk6deqUIiMj1bNnT0nsSQIAAICHj0gfAIAsmM1ma0jSs2dPvfDCC2rYsKHGjRsn6X9L7l26dEnnz5+Xi4uLJOncuXPW5UEaNmyoLl26SPrfbBQAABxZ3rx5tWDBAuuMyy+++ML6nFqzZk198MEHWrJkiUqWLCknJyc1bdpUS5Ys0eDBgyXZLt8FAAAAPCwGC5/cAABwVx06dNDatWtVs2ZNpaena8OGDXr99dc1fvx4FS1aVJLUuXNnrVy5Us2aNVNycrI2btyoCRMmqH///vYtHgCAx1R8fLz69eunFStWaNKkSQoNDbU5f/PmTTk5OcloNFqPmc1mlu8BAADAI8HSWwAA3MFkMlk/lDly5IiOHj2q6OhoderUSWlpaVqxYoX69++vtLQ0TZs2TcWLF9eIESPk6+urpUuXqnTp0oqOjlbv3r0lsTwIAABZ8fX11bRp02SxWPTuu+/KycnJ+gUDs9lsnal55/MoIQkAAAAeFWaUAACQhe7du6t48eKKi4vTN998I29vb0m3v+H62WefqUePHgoMDNTMmTNVpEgRSdKFCxfk7u4uT09PSXzzFQCAe0lISNA777yjlStXaurUqerbt6+9SwIAAIADYkYJAAB/ceTIEW3ZskUHDhyQn5+frly5Yg1KXFxc1L59e0lSjx491LdvX02ePFklS5ZUgQIFrH1YLBZCEgAA7sHHx0fTp0+X0WhU//79dfXqVb333nv2LgsAAAAOhk9wAAD4i3LlyunDDz9Uq1atdOLECX3//fc2m7E7Ozurffv2+vjjj/Xll1+qe/fuSktLs+mD5bYAALg/Pj4++uCDD9S0aVO5u7vbuxwAAAA4IJbeAgA4tLstj2WxWLR582YNHTpUv/32mxYtWqSWLVvatLl165aWLFmitLQ09enT598qGQCAHOnq1avy8PCwdxkAAABwQAQlAACHdevWLTk7316Fct++fTp//ryeeeYZeXl5KVeuXLJYLNq6dasGDx6s/fv3a+HChZnCkjs3mWXjdgAA/jmeTwEAAPBvIygBADgkk8kko9EoSfrvf/+rtWvXKikpSV5eXnrzzTfVpUsXVaxYMVNYsnjxYjVv3tzO1QMAAAAAAOBhYY8SAIBDyghJ2rVrp9jYWPXu3Vtr1qxRnz59NH36dEVFRWnPnj0yGAx64YUXNGHCBFWpUkWvvPKKvvjiCztXDwAAAAAAgIeFoAQAkOPt2bNHFy5cyHR82rRp2rt3r6ZOnarBgwerSZMmCg4O1rVr1/Tdd99p9OjR2rdvnzUsGTNmjMqWLauEhAQ7jAIAAAAAAACPAkEJACBHi4uL03PPPaeZM2fq4sWL1uPp6ek6duyYihcvroYNG8rT01M///yzatasqS5duujNN9/UZ599phEjRui3336TwWBQrVq1tG7dOr355pt2HBEAAAAAAAAeJvYoAQDkeC+99JK+//57DR8+XL169VL+/PllMpm0ceNGFSlSRBUqVNCRI0cUEBCgRo0aaebMmSpUqJCaNWumX375RfXq1dPQoUNVrVo1a59sNAsAAAAAAJAzMKMEAJBj3bp1S5L01VdfqXnz5oqIiNCcOXN04cIFGY1G1a1bVxUqVNDNmzc1Y8YMlSpVSiNGjFDevHklSfny5ZPRaNSqVasUHx9v0zchCQAAAAAAQM7gbO8CAAB4VJydnWUymWQ0GrVy5Uq1b99eERERkmSdWSJJLi4uOnbsmIxGo5555hlJ0smTJ5WamqrFixerWLFiqlSpkt3GAQAAAAAAgEeHGSUAgBzNaDTKZDJJklauXKnWrVtbZ5Zk7FmSnp4uT09PnTt3TitWrNCOHTs0d+5cbdq0SZKsIYnZbLbPIAAAAAAAAPDIsEcJACBHMpvNcnL63/cBMmaWSNLLL7+sr776SlFRUerRo4cKFiyoEydOqH79+kpKSpLBYJDFYlF4eLiGDBliryEAAAAAAADgX0BQAgDIcW7duiVn59urS544cUIuLi7KlSuX8uXLZ23Trl07ff3114qKilK3bt3k4+Ojc+fOacmSJTIYDHr66afVokULSZlDFwAAAAAAAOQcBCUAgBzlzlCjW7du+vHHH3XlyhXlzZtXYWFhevHFF1WmTBlJtmHJG2+8ocKFC/9tfwAAAAAAAMh5CEoAADnGnaFGx44dtW7dOrVs2VJeXl765Zdf9Ouvv+r111/XgAED5O/vL0lq3769Vq9erWHDhunNN99UwYIF7TgCAAAAAAAA/NsISgAAOcJfZ340btxYbdq0UY8ePeTq6qpLly4pJiZGERERCgkJUVRUlIoVKyZJatOmjb766itt375d1atXt9cQAAAAAAAAYAfO9i4AAICHISMkadu2rRISEpSSkqIXX3xRrq6ukqR8+fIpPDxcZrNZI0eOVPPmzfXyyy9LklatWqVvv/2WkAQAAAAAAMABseg6AOCJZjabrT+fO3dOZrNZf/zxh86ePauLFy9Kur25uyRZLBZ1795dZcuW1ZQpU3TlyhXdvHlTktS8efNM/QEAAAAAACDnIygBADyx7lxuKzU1VUWKFNHkyZPVokULpaSkaNasWZIkZ2dn3bx5UwaDQUWLFlWFChWUmpoqFxcXubi42PTJxu0AAAAAAACOhaW3AABPrIxQo1WrVqpSpYrGjBmjp556SsOHD5fZbNbChQuVJ08effTRR9ZA5Ny5c7p48aK8vb119epVubq6ymAw2HMYAAAAAAAAsCOCEgDAE81kMmnLli0qU6aM9Vjp0qUVGRkpSZo7d64SExP12muvyc3NTXFxcdq6datiYmKUL18+e5UNAAAAAACAx4TBYrFY7F0EAAD3y2KxWGeA3LhxQ66urmrSpIksFotiY2NlMpnk5OQkg8GgY8eOaeTIkVq8eLGcnZ1VrVo1FS1aVLVq1dKAAQMy9QcAAAAAAADHw4wSAMATw2QyyWg0Wu+7urpKkmrVqqWpU6cqMTFRhQoVknQ7ACldurQiIiIkSd98843Kli2rhQsXWs9bLBb2JAEAAAAAAHBwBCUAgCdGRkgSEhIiNzc3VapUSQ0bNpTFYpGnp6euXLliE5QYDAaVKVNGQ4cOlSQtXLhQuXPn1syZM2UwGGQ2m+02FgAAAAAAADweWHoLAPBE+f777/Xuu+8qOTlZly5d0rVr1+Tl5aWLFy+qTZs2CgwMVNWqVVWoUCGVL19e6enpcnNz0x9//KFRo0Zp4cKFeueddxQdHW3voQAAAAAAAOAxQFACAHis/XW5rQyXL19WUlKSfvnlF6WkpKh3794ym80qVKiQEhMT5e3tLT8/P+XOnVurVq1SgQIFdOzYMY0ePVqffPKJBg4cqIkTJ9phRAAAAAAAAHicsPQWAOCxdevWLTk7336q+vTTT3X48GE5OzurTp06qlevnvLkyaPSpUvr0qVLWrx4scqUKaOIiAj9/vvvOnLkiH744QfVq1dPBQoUkCSVLl1aYWFhcnNzU9euXe04MgAAAAAAADwumFECAHgsmc1m60brL730kjZt2iSDwSAXFxclJiZq8ODB6tWrl/z8/CRJwcHB2rt3r/bv33/P/u4MYAAAAAAAAODYnOxdAAAAWckINXr16qUdO3ZoypQpOnz4sPbs2aO2bdtq/Pjx2rt3r27evClJeuaZZ3T+/HldvHhR0u0luzK+C2CxWKz9SSIkAQAAAAAAgBVBCQDgsZWYmKiffvpJnTt3Vtu2bVWgQAH9/vvvWrt2rTp37qyqVavKxcVFkvTiiy/qwoUL2rlzpyTJaDTKYDBIkvVPAAAAAAAA4K8ISgAAj60zZ87o0KFDat26tby9vRUbG6sWLVqoTZs2Gjt2rEqWLClJOn78uJ566ilJEitKAgAAAAAAIDsISgAAj61SpUrJx8dHu3bt0oYNG9S6dWu1a9dOEydOVJEiRSRJixcvlr+/v1JTU/X999+rcePGdq4aAAAAAAAATxIWaQcAPLaMRqPKly+vCRMmKCkpSe3atVN0dLTy588vSfrzzz8VGxuratWqydvbWxUrVpRku3E7AAAAAAAA8Hf4FAkA8Njy8vJSdHS0Ll26JJPJpGbNmqlAgQIyGAw6ceKE5s+fr6+++kqdO3dW0aJFrdcRkgAAAAAAAOB+GSws5g4AeMytW7dOL7/8sry9vdWwYUMVL15cW7du1ZYtWzRixAgNGTJE0u39Sdi4HQAAAAAAANlBUAIAeCLs3btXERER2rNnjy5duqQ6deqoQ4cO6tSpkySW2wIAAAAAAMCDISgBADwx0tPT5eLiotTUVOXOnVvOzre32iIkAQAAAAAAwIMiKAEAPDHuXFor42eW2wIAAAAAAMA/QVACAAAAAAAAAAAcFuuUAAAAAAAAAAAAh0VQAgAAAAAAAAAAHBZBCQAAAAAAAAAAcFgEJQAAAAAAAAAAwGERlAAAAAAAAAAAAIdFUAIAAAAAAAAAABwWQQkAAAAAAAAAAHBYBCUAAAAAAAAAAMBhEZQAAAAAAAAAAACHRVACAAAAAAAAAAAcFkEJAAAAAAAAAABwWP8PpO1eOCjApjUAAAAASUVORK5CYII=",
            "text/plain": [
              "<Figure size 2000x200 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "text/html": [
              "<b>Text</b>: <i>Is to complicated and boring.</i>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "text/html": [
              "<b>Integrated Gradient Scores:</b> <span style=\"color:#fecf6b\"><b>Is</b></span> <span style=\"color:#fffdc8\"><b>to</b></span> <span style=\"color:#fc5f2f\"><b>complicated</b></span> <span style=\"color:#ffffcc\"><b>and</b></span> <span style=\"color:#800026\"><b>boring</b></span> <span style=\"color:#fff9be\"><b>.</b></span>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 2000x200 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "text/html": [
              "<b>Text</b>: <i>Garbage, not worth my time and money!</i>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "text/html": [
              "<b>Integrated Gradient Scores:</b> <span style=\"color:#fff1ab\"><b>G</b></span> <span style=\"color:#f23924\"><b>##ar</b></span> <span style=\"color:#800026\"><b>##bage</b></span> <span style=\"color:#ffeda1\"><b>,</b></span> <span style=\"color:#ffe998\"><b>not</b></span> <span style=\"color:#febe59\"><b>worth</b></span> <span style=\"color:#fff6b6\"><b>my</b></span> <span style=\"color:#fff9be\"><b>time</b></span> <span style=\"color:#ffffcc\"><b>and</b></span> <span style=\"color:#fed673\"><b>money</b></span> <span style=\"color:#ffeb9c\"><b>!</b></span>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 2000x200 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "text/html": [
              "<b>Text</b>: <i>Is nice to see philosophers doing machine learning.</i>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "text/html": [
              "<b>Integrated Gradient Scores:</b> <span style=\"color:#fd913e\"><b>Is</b></span> <span style=\"color:#800026\"><b>nice</b></span> <span style=\"color:#ffe997\"><b>to</b></span> <span style=\"color:#ffe590\"><b>see</b></span> <span style=\"color:#fea546\"><b>philosophers</b></span> <span style=\"color:#fee187\"><b>doing</b></span> <span style=\"color:#fd9c42\"><b>machine</b></span> <span style=\"color:#fed36f\"><b>learning</b></span> <span style=\"color:#ffffcc\"><b>.</b></span>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 2000x200 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "text/html": [
              "<b>Text</b>: <i>This is a great and wonderful example of nlp.</i>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "text/html": [
              "<b>Integrated Gradient Scores:</b> <span style=\"color:#fc572c\"><b>This</b></span> <span style=\"color:#ffefa5\"><b>is</b></span> <span style=\"color:#ffffcc\"><b>a</b></span> <span style=\"color:#febd57\"><b>great</b></span> <span style=\"color:#ffe793\"><b>and</b></span> <span style=\"color:#800026\"><b>wonderful</b></span> <span style=\"color:#f03523\"><b>example</b></span> <span style=\"color:#fff6b6\"><b>of</b></span> <span style=\"color:#ffe590\"><b>n</b></span> <span style=\"color:#ffe794\"><b>##l</b></span> <span style=\"color:#fff1ab\"><b>##p</b></span> <span style=\"color:#fff7b9\"><b>.</b></span>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 2000x200 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "text/html": [
              "<b>Text</b>: <i>This tutorial is great and I have learned a lot.</i>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "text/html": [
              "<b>Integrated Gradient Scores:</b> <span style=\"color:#fede80\"><b>This</b></span> <span style=\"color:#800026\"><b>tutor</b></span> <span style=\"color:#fedb7b\"><b>##ial</b></span> <span style=\"color:#ffe48d\"><b>is</b></span> <span style=\"color:#fd8f3d\"><b>great</b></span> <span style=\"color:#fff4b0\"><b>and</b></span> <span style=\"color:#fff0a8\"><b>I</b></span> <span style=\"color:#fffac1\"><b>have</b></span> <span style=\"color:#fea145\"><b>learned</b></span> <span style=\"color:#ffffcc\"><b>a</b></span> <span style=\"color:#ffec9f\"><b>lot</b></span> <span style=\"color:#ffeda1\"><b>.</b></span>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 2000x200 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "import matplotlib\n",
        "import matplotlib.cm as cm\n",
        "import matplotlib.pyplot as plt\n",
        "import matplotlib.colors as mcolors\n",
        "\n",
        "from IPython.display import HTML\n",
        "\n",
        "def plot_gradient_scores(text, words, gradients, prediction_class):\n",
        "    \"\"\"\n",
        "    Plot the gradient scores for each word in the input text.\n",
        "\n",
        "    Args:\n",
        "        text (str): The input text for which gradients need to be computed.\n",
        "        words (list): A list of tokens corresponding to the input text.\n",
        "        gradients (list): A list of gradient values, each corresponding to a\n",
        "            token in the sequence.\n",
        "        prediction_label (str): The predicted label for the input text.\n",
        "\n",
        "    Returns:\n",
        "        None : the function plots a figure instead\n",
        "    \"\"\"\n",
        "\n",
        "    # Remove special tokens [CLS] and [SEP]\n",
        "    words = words[1:-1]\n",
        "    gradients = gradients[1:-1]\n",
        "\n",
        "    # Get min and max gradients\n",
        "    minima = min(gradients)\n",
        "    maxima = max(gradients)\n",
        "\n",
        "    # Normalize gradients to between 0 and 1\n",
        "    norm = matplotlib.colors.Normalize(vmin=minima, vmax=maxima, clip=True)\n",
        "    mapper = cm.ScalarMappable(norm=norm, cmap=cm.YlOrRd)\n",
        "\n",
        "    # Get colors corresponding to gradients\n",
        "    colors = [mcolors.to_hex(mapper.to_rgba(v)) for v in gradients]\n",
        "\n",
        "    # Create html string with the words colored corresponding to their gradients\n",
        "    text_with_gradients = ' '.join([f'''<span style=\"color:{colors[i]}\"><b>{words[i]}</b></span>''' for i in range(len(words))])\n",
        "    display(HTML(f\"<b>Text</b>: <i>{text}</i>\"))\n",
        "    display(HTML(f\"<b>Integrated Gradient Scores:</b> {text_with_gradients}\"))\n",
        "\n",
        "    # Plot the gradients on a bar chart\n",
        "    plt.figure(figsize=(20,2))\n",
        "    xvals = [ x + str(i) for i,x in enumerate(words)]\n",
        "    colors =  [ (0,0,1, c) for c in (gradients) ]\n",
        "    plt.tick_params(axis='both', which='minor', labelsize=29)\n",
        "    plt.title(f\"Prediction: {model.config.id2label[int(prediction_class)]}\")\n",
        "    plt.bar(xvals, gradients, color=colors, linewidth=1 )\n",
        "    plt.xticks(ticks=[i for i in range(len(words))], labels=words, fontsize=12,rotation=45)\n",
        "    plt.show()\n",
        "\n",
        "texts = [\n",
        "    'This explanation is really Bad!',\n",
        "    'Is to complicated and boring.',\n",
        "    'Garbage, not worth my time and money!',\n",
        "    'Is nice to see philosophers doing machine learning.',\n",
        "    'This is a great and wonderful example of nlp.',\n",
        "    'This tutorial is great and I have learned a lot.'\n",
        "]\n",
        "\n",
        "\n",
        "for text in texts:\n",
        "    gradient_tensor, token_words, prediction_class = get_gradients(text, model, tokenizer)\n",
        "    plot_gradient_scores(text, token_words, gradient_tensor, prediction_class)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9nWs08Y3oFFY"
      },
      "source": [
        "Awesome! We learned how to interpret the outputs of our text classifier via the integrated gradients technique. Now, let us apply this to a simple interpretability app.\n",
        "\n",
        "We have a demo app of the following URL that employs this explainability technique.\n",
        "\n",
        "- Visit our [Integrated-gradients-DistilBERT](https://huggingface.co/spaces/nicholasKluge/Integrated-gradients-DistilBERT?logs=build) Space! 🤗\n",
        "\n",
        "> **Note:** In it, we are using  [Gradio](https://gradio.app/) to construct the interface. You can learn more on how to build these apps in [this tutorial](https://github.com/Nkluge-correa/TeenyTinyCastle/blob/master/ML-Intro-Course/19_quick_AI_app.ipynb), and [Gradio's documentaion](https://www.gradio.app/docs/interface).\n",
        "\n",
        "By applying `Integrated Gradients`, we can gain valuable insights into the inner workings of machine learning models and improve trust, interpretability, and understanding in AI systems, promoting accountability and ethical considerations in these applications.\n",
        "\n",
        "---\n",
        "\n",
        "Return to the [castle](https://github.com/Nkluge-correa/TeenyTinyCastle)."
      ]
    }
  ],
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.9.13"
    },
    "orig_nbformat": 4
  },
  "nbformat": 4,
  "nbformat_minor": 0
}