diff --git "a/13_Siamese_Network.ipynb" "b/13_Siamese_Network.ipynb" new file mode 100644--- /dev/null +++ "b/13_Siamese_Network.ipynb" @@ -0,0 +1,828 @@ +{ + "cells": [ + { + "cell_type": "raw", + "metadata": {}, + "source": [ + "---\n", + "title: 14 Siamese Network\n", + "description: An implementation of Siamese Network for finding Image Similarity\n", + "---" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "\"Colab\"" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "TW4hI-5Fw_FE" + }, + "source": [ + "\"Kaggle\"" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19", + "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5", + "id": "OAAUsjMxw6ma" + }, + "source": [ + "# Siamese Network\n", + "\n", + "A Siamese neural network (sometimes called a twin neural network) is an artificial neural network that uses the same weights while working in tandem on two different input vectors to compute comparable output vectors.\n", + "\n", + "Often one of the output vectors is precomputed, thus forming a baseline against which the other output vector is compared. This is similar to comparing fingerprints but can be described more technically as a distance function for locality-sensitive hashing.\n", + "\n", + "It is possible to make a kind of structure that is functional similar to a siamese network, but implements a slightly different function. This is typically used for comparing similar instances in different type sets.\n", + "\n", + "Uses of similarity measures where a twin network might be used are such things as recognizing handwritten checks, automatic detection of faces in camera images, and matching queries with indexed documents. The perhaps most well-known application of twin networks are face recognition, where known images of people are precomputed and compared to an image from a turnstile or similar. It is not obvious at first, but there are two slightly different problems. One is recognizing a person among a large number of other persons, that is the facial recognition problem.\n", + "\n", + "Source: [Wikipedia](https://en.wikipedia.org/wiki/Siamese_neural_network)" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "_cell_guid": "79c7e3d0-c299-4dcb-8224-4455121ee9b0", + "_uuid": "d629ff2d2480ee46fbb7e2d37f6b5fab8052498a", + "execution": { + "iopub.execute_input": "2022-04-01T14:46:09.010522Z", + "iopub.status.busy": "2022-04-01T14:46:09.010194Z", + "iopub.status.idle": "2022-04-01T14:46:14.040252Z", + "shell.execute_reply": "2022-04-01T14:46:14.039438Z", + "shell.execute_reply.started": "2022-04-01T14:46:09.010490Z" + }, + "id": "ErVZSzL8w6me" + }, + "outputs": [], + "source": [ + "import numpy as np\n", + "import pandas as pd\n", + "import matplotlib.pyplot as plt\n", + "import seaborn as sns\n", + "import random\n", + "\n", + "import tensorflow as tf\n", + "import tensorflow.keras.backend as K\n", + "from tensorflow.keras.datasets import mnist\n", + "from tensorflow.keras.models import Model, Sequential\n", + "from tensorflow.keras import layers, utils, callbacks" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "prw5qfNew6mf" + }, + "source": [ + "Lets start off by creating our dataset. Our input data will be consisting of pairs of images, and output will be either 1 or 0 indicating if the pair are similar or not" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "execution": { + "iopub.execute_input": "2022-04-01T14:46:14.042352Z", + "iopub.status.busy": "2022-04-01T14:46:14.042027Z", + "iopub.status.idle": "2022-04-01T14:46:14.052334Z", + "shell.execute_reply": "2022-04-01T14:46:14.051406Z", + "shell.execute_reply.started": "2022-04-01T14:46:14.042324Z" + }, + "id": "y_h0cSuxw6mg" + }, + "outputs": [], + "source": [ + "def make_pairs(images, labels, seed=19):\n", + " np.random.seed(seed)\n", + " pairImages = []\n", + " pairLabels = []\n", + " \n", + " numClasses = len(np.unique(labels))\n", + " idx = [np.where(labels == i)[0] for i in range(numClasses)]\n", + "\n", + " for idxA in range(len(images)):\n", + " currentImage = images[idxA]\n", + " label = labels[idxA]\n", + " \n", + " idxB = np.random.choice(idx[label])\n", + " posImage = images[idxB]\n", + " \n", + " pairImages.append([currentImage, posImage])\n", + " pairLabels.append([1])\n", + "\n", + " negIdx = np.where(labels != label)[0]\n", + " negImage = images[np.random.choice(negIdx)]\n", + " \n", + " pairImages.append([currentImage, negImage])\n", + " pairLabels.append([0])\n", + " \n", + " return (np.array(pairImages), np.array(pairLabels)) " + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "4vrjuoSDw6mg" + }, + "source": [ + "We will be working with `MNIST` dataset in our notebook which comes along with the tensorflow library" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/" + }, + "execution": { + "iopub.execute_input": "2022-04-01T14:46:14.054287Z", + "iopub.status.busy": "2022-04-01T14:46:14.053699Z", + "iopub.status.idle": "2022-04-01T14:46:23.809877Z", + "shell.execute_reply": "2022-04-01T14:46:23.809059Z", + "shell.execute_reply.started": "2022-04-01T14:46:14.054245Z" + }, + "id": "PAJIVYrgw6mh", + "outputId": "fe95d301-0dfd-4ca3-c98d-25ef010cb2c9" + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "Train Data Shape: (120000, 2, 28, 28, 1)\n", + "Test Data Shape: (20000, 2, 28, 28, 1)\n", + "\n", + "\n", + "CPU times: user 16.7 s, sys: 1.45 s, total: 18.1 s\n", + "Wall time: 23.9 s\n" + ] + } + ], + "source": [ + "%%time\n", + "\n", + "(trainX, trainY), (testX, testY) = mnist.load_data()\n", + "\n", + "trainX = 1 - (trainX / 255.0)\n", + "testX = 1 - (testX / 255.0)\n", + "\n", + "trainX = np.expand_dims(trainX, axis=-1)\n", + "testX = np.expand_dims(testX, axis=-1)\n", + "\n", + "(pairTrain, labelTrain) = make_pairs(trainX, trainY)\n", + "(pairTest, labelTest) = make_pairs(testX, testY)\n", + "\n", + "print(f'\\nTrain Data Shape: {pairTrain.shape}')\n", + "print(f'Test Data Shape: {pairTest.shape}\\n\\n')" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "cBy-yF9Iw6mi" + }, + "source": [ + "Lets visualize the mnist images" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/", + "height": 400 + }, + "execution": { + "iopub.execute_input": "2022-04-01T14:46:23.812279Z", + "iopub.status.busy": "2022-04-01T14:46:23.811691Z", + "iopub.status.idle": "2022-04-01T14:46:24.381571Z", + "shell.execute_reply": "2022-04-01T14:46:24.380711Z", + "shell.execute_reply.started": "2022-04-01T14:46:23.812231Z" + }, + "id": "PJthfWA6w6mi", + "outputId": "31765a33-a803-403e-85ce-30ef61c20b1f" + }, + "outputs": [ + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": { + "needs_background": "light" + }, + "output_type": "display_data" + } + ], + "source": [ + "fig, ax = plt.subplots(2, 6, figsize=(20, 6))\n", + "random.seed(19)\n", + "idx = random.choices(range(len(trainX)), k=12)\n", + "for i in range(12):\n", + " ax[i//6][i%6].imshow(np.squeeze(trainX[idx[i]]), cmap='gray')\n", + " ax[i//6][i%6].set_title(f'Label: {trainY[idx[i]]}', fontsize=18)\n", + " ax[i//6][i%6].set_axis_off()\n", + "fig.suptitle('MNIST Images', fontsize=24);" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "RHP_-Wj7w6mj" + }, + "source": [ + "Here is a sample of our prepared dataset" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/", + "height": 400 + }, + "execution": { + "iopub.execute_input": "2022-04-01T14:46:24.384833Z", + "iopub.status.busy": "2022-04-01T14:46:24.384487Z", + "iopub.status.idle": "2022-04-01T14:46:24.791767Z", + "shell.execute_reply": "2022-04-01T14:46:24.790799Z", + "shell.execute_reply.started": "2022-04-01T14:46:24.384802Z" + }, + "id": "pOmShTLXw6mj", + "outputId": "e44261dc-6baa-4283-bc9f-25df3b8548fb" + }, + "outputs": [ + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": { + "needs_background": "light" + }, + "output_type": "display_data" + } + ], + "source": [ + "fig, ax = plt.subplots(2, 6, figsize=(20, 6))\n", + "random.seed(19)\n", + "idx = random.choices(range(len(pairTrain)), k=6)\n", + "for i in range(0, 12, 2):\n", + " ax[i//6][i%6].imshow(np.squeeze(pairTrain[idx[i//2]][0]), cmap='gray')\n", + " ax[i//6][i%6+1].imshow(np.squeeze(pairTrain[idx[i//2]][1]), cmap='gray')\n", + " ax[i//6][i%6].set_title(f'Label: {labelTrain[idx[i//2]]}', fontsize=18)\n", + " ax[i//6][i%6].set_axis_off()\n", + " ax[i//6][i%6+1].set_axis_off()\n", + "fig.suptitle('Input Pair Images', fontsize=24);" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "8N5_Ee4Aw6mk" + }, + "source": [ + "Here we define some configurations for our model" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "execution": { + "iopub.execute_input": "2022-04-01T14:46:24.793672Z", + "iopub.status.busy": "2022-04-01T14:46:24.793306Z", + "iopub.status.idle": "2022-04-01T14:46:24.800983Z", + "shell.execute_reply": "2022-04-01T14:46:24.798602Z", + "shell.execute_reply.started": "2022-04-01T14:46:24.793642Z" + }, + "id": "HftBev7Ew6mk" + }, + "outputs": [], + "source": [ + "class config():\n", + " IMG_SHAPE = (28, 28, 1)\n", + " EMBEDDING_DIM = 48\n", + " \n", + " BATCH_SIZE = 64\n", + " EPOCHS = 500" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "v_AMXzKow6mk" + }, + "source": [ + "Here we define a function to calculate euclidean distance between two vectors. This will be used by our model to calculate the euclidean distance between the vectors of the image pairs (image vectors will be created by the feature extractor of our model)" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "execution": { + "iopub.execute_input": "2022-04-01T14:46:24.803691Z", + "iopub.status.busy": "2022-04-01T14:46:24.803047Z", + "iopub.status.idle": "2022-04-01T14:46:24.812099Z", + "shell.execute_reply": "2022-04-01T14:46:24.811006Z", + "shell.execute_reply.started": "2022-04-01T14:46:24.803620Z" + }, + "id": "wb9NwizXw6ml" + }, + "outputs": [], + "source": [ + "def euclidean_distance(vectors):\n", + " (featsA, featsB) = vectors\n", + " sumSquared = K.sum(K.square(featsA - featsB), axis=1, keepdims=True)\n", + " return K.sqrt(K.maximum(sumSquared, K.epsilon()))" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "dW9hpE4_w6ml" + }, + "source": [ + "With Siamese Network, the two most commonly used loss functions are:\n", + "* contrastive loss\n", + "* triplet loss\n", + "\n", + "We will be using contrastive loss in this notebook ie:\n", + "\n", + "```Contrastive loss = mean( (1-true_value) * square(prediction) + true_value * square( max(margin - prediction, 0)))```" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "execution": { + "iopub.execute_input": "2022-04-01T14:46:24.814676Z", + "iopub.status.busy": "2022-04-01T14:46:24.814000Z", + "iopub.status.idle": "2022-04-01T14:46:24.824946Z", + "shell.execute_reply": "2022-04-01T14:46:24.823903Z", + "shell.execute_reply.started": "2022-04-01T14:46:24.814629Z" + }, + "id": "ICXLZFp1w6ml" + }, + "outputs": [], + "source": [ + "def loss(margin=1):\n", + " def contrastive_loss(y_true, y_pred):\n", + " y_true = tf.cast(y_true, y_pred.dtype)\n", + " square_pred = tf.math.square(y_pred)\n", + " margin_square = tf.math.square(tf.math.maximum(margin - (y_pred), 0))\n", + " return tf.math.reduce_mean(\n", + " (1 - y_true) * square_pred + (y_true) * margin_square\n", + " )\n", + "\n", + " return contrastive_loss" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "rtFBjwEyw6mm" + }, + "source": [ + "Finally we define our model architecture\n", + "\n", + "* The model contains two input layers\n", + "* A feature extractor through which both the images will be passed to generate feature vectors, the feature extractor typically consists of Convolutional and Pooling Layers\n", + "* The feature vectors are passed through a custom layer to get euclidean distance between the vectors\n", + "* The final layer consists of a single sigmoid unit" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "execution": { + "iopub.execute_input": "2022-04-01T14:46:24.827411Z", + "iopub.status.busy": "2022-04-01T14:46:24.826654Z", + "iopub.status.idle": "2022-04-01T14:46:27.133936Z", + "shell.execute_reply": "2022-04-01T14:46:27.133040Z", + "shell.execute_reply.started": "2022-04-01T14:46:24.827365Z" + }, + "id": "UFUpR9O5w6mm" + }, + "outputs": [], + "source": [ + "class SiameseNetwork(Model):\n", + " def __init__(self, inputShape, embeddingDim):\n", + " super(SiameseNetwork, self).__init__()\n", + " \n", + " imgA = layers.Input(shape=inputShape)\n", + " imgB = layers.Input(shape=inputShape)\n", + " featureExtractor = self.build_feature_extractor(inputShape, embeddingDim)\n", + " featsA = featureExtractor(imgA)\n", + " featsB = featureExtractor(imgB)\n", + " distance = layers.Lambda(euclidean_distance, name='euclidean_distance')([featsA, featsB])\n", + " outputs = layers.Dense(1, activation=\"sigmoid\")(distance)\n", + " self.model = Model(inputs=[imgA, imgB], outputs=outputs) \n", + " \n", + " def build_feature_extractor(self, inputShape, embeddingDim=48):\n", + "\n", + " model = Sequential([\n", + " layers.Input(inputShape),\n", + " layers.Conv2D(64, (2, 2), padding=\"same\", activation=\"relu\"),\n", + " layers.MaxPooling2D(pool_size=2),\n", + " layers.Dropout(0.3),\n", + " layers.Conv2D(64, (2, 2), padding=\"same\", activation=\"relu\"),\n", + " layers.MaxPooling2D(pool_size=2),\n", + " layers.Dropout(0.3),\n", + " layers.Conv2D(128, (1, 1), padding=\"same\", activation=\"relu\"),\n", + " layers.Flatten(),\n", + " layers.Dense(embeddingDim, activation='tanh')\n", + " ])\n", + "\n", + " return model \n", + " \n", + " def call(self, x):\n", + " return self.model(x)\n", + "\n", + "model = SiameseNetwork(inputShape=config.IMG_SHAPE, embeddingDim=config.EMBEDDING_DIM)\n", + "model.compile(loss=loss(margin=1), optimizer=\"adam\", metrics=[\"accuracy\"])" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "B-a_8ZjKw6mm" + }, + "source": [ + "This is how the model looks like" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/", + "height": 1000 + }, + "execution": { + "iopub.execute_input": "2022-04-01T14:46:27.135778Z", + "iopub.status.busy": "2022-04-01T14:46:27.135364Z", + "iopub.status.idle": "2022-04-01T14:46:27.675382Z", + "shell.execute_reply": "2022-04-01T14:46:27.674262Z", + "shell.execute_reply.started": "2022-04-01T14:46:27.135713Z" + }, + "id": "pl8nGMCtw6mn", + "outputId": "c0c4d9b4-78a7-4c86-f6ad-de3362252c80" + }, + "outputs": [ + { + "data": { + "image/png": "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\n", + "text/plain": [ + "" + ] + }, + "execution_count": 10, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "utils.plot_model(model.model, show_shapes=True, expand_nested=True)" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/" + }, + "execution": { + "iopub.execute_input": "2022-04-01T14:46:27.677082Z", + "iopub.status.busy": "2022-04-01T14:46:27.676707Z", + "iopub.status.idle": "2022-04-01T14:56:20.212695Z", + "shell.execute_reply": "2022-04-01T14:56:20.211833Z", + "shell.execute_reply.started": "2022-04-01T14:46:27.677043Z" + }, + "id": "6dC6hZFHw6mn", + "outputId": "534125d2-e152-4032-e8a3-6f8056e7cb74" + }, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Epoch 1/500\n", + "1875/1875 [==============================] - 17s 8ms/step - loss: 0.1404 - accuracy: 0.7989 - val_loss: 0.0645 - val_accuracy: 0.9389 - lr: 0.0010\n", + "Epoch 2/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0610 - accuracy: 0.9333 - val_loss: 0.0470 - val_accuracy: 0.9498 - lr: 0.0010\n", + "Epoch 3/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0500 - accuracy: 0.9429 - val_loss: 0.0420 - val_accuracy: 0.9539 - lr: 0.0010\n", + "Epoch 4/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0361 - accuracy: 0.9592 - val_loss: 0.0268 - val_accuracy: 0.9708 - lr: 0.0010\n", + "Epoch 5/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0268 - accuracy: 0.9699 - val_loss: 0.0209 - val_accuracy: 0.9769 - lr: 0.0010\n", + "Epoch 6/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0240 - accuracy: 0.9723 - val_loss: 0.0205 - val_accuracy: 0.9772 - lr: 0.0010\n", + "Epoch 7/500\n", + "1875/1875 [==============================] - 14s 8ms/step - loss: 0.0221 - accuracy: 0.9744 - val_loss: 0.0172 - val_accuracy: 0.9816 - lr: 0.0010\n", + "Epoch 8/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0204 - accuracy: 0.9767 - val_loss: 0.0173 - val_accuracy: 0.9803 - lr: 0.0010\n", + "Epoch 9/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0195 - accuracy: 0.9772 - val_loss: 0.0159 - val_accuracy: 0.9823 - lr: 0.0010\n", + "Epoch 10/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0184 - accuracy: 0.9785 - val_loss: 0.0145 - val_accuracy: 0.9837 - lr: 0.0010\n", + "Epoch 11/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0162 - accuracy: 0.9814 - val_loss: 0.0142 - val_accuracy: 0.9834 - lr: 0.0010\n", + "Epoch 12/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0150 - accuracy: 0.9827 - val_loss: 0.0136 - val_accuracy: 0.9849 - lr: 0.0010\n", + "Epoch 13/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0141 - accuracy: 0.9841 - val_loss: 0.0134 - val_accuracy: 0.9843 - lr: 0.0010\n", + "Epoch 14/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0140 - accuracy: 0.9839 - val_loss: 0.0121 - val_accuracy: 0.9859 - lr: 0.0010\n", + "Epoch 15/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0129 - accuracy: 0.9857 - val_loss: 0.0135 - val_accuracy: 0.9845 - lr: 0.0010\n", + "Epoch 16/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0123 - accuracy: 0.9861 - val_loss: 0.0113 - val_accuracy: 0.9868 - lr: 0.0010\n", + "Epoch 17/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0113 - accuracy: 0.9873 - val_loss: 0.0106 - val_accuracy: 0.9872 - lr: 0.0010\n", + "Epoch 18/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0100 - accuracy: 0.9886 - val_loss: 0.0102 - val_accuracy: 0.9877 - lr: 0.0010\n", + "Epoch 19/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0091 - accuracy: 0.9895 - val_loss: 0.0093 - val_accuracy: 0.9888 - lr: 0.0010\n", + "Epoch 20/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0086 - accuracy: 0.9901 - val_loss: 0.0083 - val_accuracy: 0.9905 - lr: 0.0010\n", + "Epoch 21/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0083 - accuracy: 0.9906 - val_loss: 0.0089 - val_accuracy: 0.9895 - lr: 0.0010\n", + "Epoch 22/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0082 - accuracy: 0.9907 - val_loss: 0.0078 - val_accuracy: 0.9912 - lr: 0.0010\n", + "Epoch 23/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0081 - accuracy: 0.9905 - val_loss: 0.0082 - val_accuracy: 0.9905 - lr: 0.0010\n", + "Epoch 24/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0074 - accuracy: 0.9914 - val_loss: 0.0080 - val_accuracy: 0.9906 - lr: 0.0010\n", + "Epoch 25/500\n", + "1870/1875 [============================>.] - ETA: 0s - loss: 0.0074 - accuracy: 0.9914\n", + "Epoch 25: ReduceLROnPlateau reducing learning rate to 0.0005000000237487257.\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0074 - accuracy: 0.9915 - val_loss: 0.0091 - val_accuracy: 0.9887 - lr: 0.0010\n", + "Epoch 26/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0057 - accuracy: 0.9935 - val_loss: 0.0073 - val_accuracy: 0.9908 - lr: 5.0000e-04\n", + "Epoch 27/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0054 - accuracy: 0.9938 - val_loss: 0.0066 - val_accuracy: 0.9923 - lr: 5.0000e-04\n", + "Epoch 28/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0053 - accuracy: 0.9940 - val_loss: 0.0066 - val_accuracy: 0.9918 - lr: 5.0000e-04\n", + "Epoch 29/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0051 - accuracy: 0.9945 - val_loss: 0.0065 - val_accuracy: 0.9928 - lr: 5.0000e-04\n", + "Epoch 30/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0052 - accuracy: 0.9943 - val_loss: 0.0066 - val_accuracy: 0.9922 - lr: 5.0000e-04\n", + "Epoch 31/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0048 - accuracy: 0.9948 - val_loss: 0.0075 - val_accuracy: 0.9909 - lr: 5.0000e-04\n", + "Epoch 32/500\n", + "1871/1875 [============================>.] - ETA: 0s - loss: 0.0047 - accuracy: 0.9947\n", + "Epoch 32: ReduceLROnPlateau reducing learning rate to 0.0002500000118743628.\n", + "1875/1875 [==============================] - 14s 8ms/step - loss: 0.0047 - accuracy: 0.9947 - val_loss: 0.0071 - val_accuracy: 0.9915 - lr: 5.0000e-04\n", + "Epoch 33/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0040 - accuracy: 0.9956 - val_loss: 0.0069 - val_accuracy: 0.9915 - lr: 2.5000e-04\n", + "Epoch 34/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0039 - accuracy: 0.9959 - val_loss: 0.0067 - val_accuracy: 0.9918 - lr: 2.5000e-04\n", + "Epoch 35/500\n", + "1871/1875 [============================>.] - ETA: 0s - loss: 0.0038 - accuracy: 0.9959\n", + "Epoch 35: ReduceLROnPlateau reducing learning rate to 0.0001250000059371814.\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0038 - accuracy: 0.9959 - val_loss: 0.0065 - val_accuracy: 0.9919 - lr: 2.5000e-04\n", + "Epoch 36/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0034 - accuracy: 0.9964 - val_loss: 0.0064 - val_accuracy: 0.9922 - lr: 1.2500e-04\n", + "Epoch 37/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0033 - accuracy: 0.9965 - val_loss: 0.0061 - val_accuracy: 0.9925 - lr: 1.2500e-04\n", + "Epoch 38/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0033 - accuracy: 0.9965 - val_loss: 0.0062 - val_accuracy: 0.9922 - lr: 1.2500e-04\n", + "Epoch 39/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0032 - accuracy: 0.9967 - val_loss: 0.0060 - val_accuracy: 0.9930 - lr: 1.2500e-04\n", + "Epoch 40/500\n", + "1870/1875 [============================>.] - ETA: 0s - loss: 0.0032 - accuracy: 0.9966\n", + "Epoch 40: ReduceLROnPlateau reducing learning rate to 6.25000029685907e-05.\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0032 - accuracy: 0.9966 - val_loss: 0.0062 - val_accuracy: 0.9926 - lr: 1.2500e-04\n", + "Epoch 41/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0029 - accuracy: 0.9970 - val_loss: 0.0060 - val_accuracy: 0.9930 - lr: 6.2500e-05\n", + "Epoch 42/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0029 - accuracy: 0.9969 - val_loss: 0.0059 - val_accuracy: 0.9933 - lr: 6.2500e-05\n", + "Epoch 43/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0030 - accuracy: 0.9968 - val_loss: 0.0059 - val_accuracy: 0.9929 - lr: 6.2500e-05\n", + "Epoch 44/500\n", + "1875/1875 [==============================] - 14s 8ms/step - loss: 0.0029 - accuracy: 0.9971 - val_loss: 0.0058 - val_accuracy: 0.9930 - lr: 6.2500e-05\n", + "Epoch 45/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0028 - accuracy: 0.9973 - val_loss: 0.0058 - val_accuracy: 0.9930 - lr: 6.2500e-05\n", + "Epoch 46/500\n", + "1871/1875 [============================>.] - ETA: 0s - loss: 0.0028 - accuracy: 0.9970\n", + "Epoch 46: ReduceLROnPlateau reducing learning rate to 3.125000148429535e-05.\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0029 - accuracy: 0.9970 - val_loss: 0.0058 - val_accuracy: 0.9931 - lr: 6.2500e-05\n", + "Epoch 47/500\n", + "1875/1875 [==============================] - 14s 8ms/step - loss: 0.0027 - accuracy: 0.9974 - val_loss: 0.0058 - val_accuracy: 0.9931 - lr: 3.1250e-05\n", + "Epoch 48/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0028 - accuracy: 0.9972 - val_loss: 0.0057 - val_accuracy: 0.9933 - lr: 3.1250e-05\n", + "Epoch 49/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0027 - accuracy: 0.9973 - val_loss: 0.0057 - val_accuracy: 0.9934 - lr: 3.1250e-05\n", + "Epoch 50/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0027 - accuracy: 0.9973 - val_loss: 0.0057 - val_accuracy: 0.9933 - lr: 3.1250e-05\n", + "Epoch 51/500\n", + "1868/1875 [============================>.] - ETA: 0s - loss: 0.0027 - accuracy: 0.9973\n", + "Epoch 51: ReduceLROnPlateau reducing learning rate to 1.5625000742147677e-05.\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0027 - accuracy: 0.9973 - val_loss: 0.0056 - val_accuracy: 0.9933 - lr: 3.1250e-05\n", + "Epoch 52/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0026 - accuracy: 0.9973 - val_loss: 0.0056 - val_accuracy: 0.9934 - lr: 1.5625e-05\n", + "Epoch 53/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0027 - accuracy: 0.9973 - val_loss: 0.0056 - val_accuracy: 0.9933 - lr: 1.5625e-05\n", + "Epoch 54/500\n", + "1874/1875 [============================>.] - ETA: 0s - loss: 0.0026 - accuracy: 0.9974\n", + "Epoch 54: ReduceLROnPlateau reducing learning rate to 7.812500371073838e-06.\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0026 - accuracy: 0.9973 - val_loss: 0.0056 - val_accuracy: 0.9933 - lr: 1.5625e-05\n", + "Epoch 55/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0026 - accuracy: 0.9973 - val_loss: 0.0056 - val_accuracy: 0.9931 - lr: 7.8125e-06\n", + "Epoch 56/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0026 - accuracy: 0.9975 - val_loss: 0.0056 - val_accuracy: 0.9933 - lr: 7.8125e-06\n", + "Epoch 57/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0027 - accuracy: 0.9973 - val_loss: 0.0056 - val_accuracy: 0.9932 - lr: 7.8125e-06\n", + "Epoch 58/500\n", + "1871/1875 [============================>.] - ETA: 0s - loss: 0.0027 - accuracy: 0.9973\n", + "Epoch 58: ReduceLROnPlateau reducing learning rate to 3.906250185536919e-06.\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0027 - accuracy: 0.9973 - val_loss: 0.0056 - val_accuracy: 0.9933 - lr: 7.8125e-06\n", + "Epoch 59/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0026 - accuracy: 0.9973 - val_loss: 0.0056 - val_accuracy: 0.9932 - lr: 3.9063e-06\n", + "Epoch 60/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0026 - accuracy: 0.9973 - val_loss: 0.0056 - val_accuracy: 0.9933 - lr: 3.9063e-06\n", + "Epoch 61/500\n", + "1869/1875 [============================>.] - ETA: 0s - loss: 0.0026 - accuracy: 0.9974\n", + "Epoch 61: ReduceLROnPlateau reducing learning rate to 1.9531250927684596e-06.\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0026 - accuracy: 0.9974 - val_loss: 0.0056 - val_accuracy: 0.9933 - lr: 3.9063e-06\n", + "Epoch 62/500\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0028 - accuracy: 0.9971 - val_loss: 0.0056 - val_accuracy: 0.9931 - lr: 1.9531e-06\n", + "Epoch 63/500\n", + "1875/1875 [==============================] - 14s 8ms/step - loss: 0.0026 - accuracy: 0.9975 - val_loss: 0.0056 - val_accuracy: 0.9932 - lr: 1.9531e-06\n", + "Epoch 64/500\n", + "1874/1875 [============================>.] - ETA: 0s - loss: 0.0027 - accuracy: 0.9971\n", + "Epoch 64: ReduceLROnPlateau reducing learning rate to 1e-06.\n", + "1875/1875 [==============================] - 14s 8ms/step - loss: 0.0027 - accuracy: 0.9972 - val_loss: 0.0056 - val_accuracy: 0.9932 - lr: 1.9531e-06\n", + "Epoch 65/500\n", + "1869/1875 [============================>.] - ETA: 0s - loss: 0.0026 - accuracy: 0.9974Restoring model weights from the end of the best epoch: 55.\n", + "1875/1875 [==============================] - 14s 7ms/step - loss: 0.0026 - accuracy: 0.9974 - val_loss: 0.0056 - val_accuracy: 0.9932 - lr: 1.0000e-06\n", + "Epoch 65: early stopping\n" + ] + } + ], + "source": [ + "es = callbacks.EarlyStopping(monitor='val_loss', patience=10, verbose=1, restore_best_weights=True, min_delta=1e-4)\n", + "\n", + "rlp = callbacks.ReduceLROnPlateau(monitor='val_loss', factor=0.5, patience=3, min_lr=1e-6, mode='min', verbose=1)\n", + "\n", + "history = model.fit(\n", + " [pairTrain[:, 0], pairTrain[:, 1]], labelTrain[:],\n", + " validation_data=([pairTest[:, 0], pairTest[:, 1]], labelTest[:]),\n", + " batch_size=config.BATCH_SIZE, \n", + " epochs=config.EPOCHS,\n", + " callbacks=[es, rlp]\n", + ")" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/", + "height": 400 + }, + "execution": { + "iopub.execute_input": "2022-04-01T14:56:20.214650Z", + "iopub.status.busy": "2022-04-01T14:56:20.214291Z", + "iopub.status.idle": "2022-04-01T14:56:21.973724Z", + "shell.execute_reply": "2022-04-01T14:56:21.972847Z", + "shell.execute_reply.started": "2022-04-01T14:56:20.214612Z" + }, + "id": "yaoZMPSMw6mo", + "outputId": "3cfa4e39-bfe0-4f8e-d96c-0cadfdcc471b" + }, + "outputs": [ + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": { + "needs_background": "light" + }, + "output_type": "display_data" + } + ], + "source": [ + "fig, ax = plt.subplots(2, 6, figsize=(20, 6))\n", + "random.seed(19)\n", + "idx = random.choices(range(len(pairTest)), k=6)\n", + "preds = model.predict([pairTest[:, 0], pairTest[:, 1]])\n", + "for i in range(0, 12, 2):\n", + " ax[i//6][i%6].imshow(np.squeeze(pairTest[idx[i//2]][0]), cmap='gray')\n", + " ax[i//6][i%6+1].imshow(np.squeeze(pairTest[idx[i//2]][1]), cmap='gray')\n", + " ax[i//6][i%6].set_title(f'Label: {labelTest[idx[i//2]]}', fontsize=18)\n", + " ax[i//6][i%6+1].set_title(f'Predicted: {np.round(preds[idx[i//2]], 2)}', fontsize=18)\n", + " ax[i//6][i%6].set_axis_off()\n", + " ax[i//6][i%6+1].set_axis_off()\n", + "fig.suptitle('Test Pair Images', fontsize=24);" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/", + "height": 540 + }, + "execution": { + "iopub.execute_input": "2022-04-01T14:56:21.975349Z", + "iopub.status.busy": "2022-04-01T14:56:21.974978Z", + "iopub.status.idle": "2022-04-01T14:56:22.521939Z", + "shell.execute_reply": "2022-04-01T14:56:22.521024Z", + "shell.execute_reply.started": "2022-04-01T14:56:21.975318Z" + }, + "id": "-YHSoBnXw6mo", + "outputId": "47484ce9-dda1-4f64-e810-6832aff53b80" + }, + "outputs": [ + { + "data": { + "image/png": "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\n", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "sns.set_style('darkgrid')\n", + "\n", + "fig, ax = plt.subplots(2, 1, figsize=(20, 8))\n", + "df = pd.DataFrame(history.history)\n", + "df[['accuracy', 'val_accuracy']].plot(ax=ax[0])\n", + "df[['loss', 'val_loss']].plot(ax=ax[1])\n", + "ax[0].set_title('Model Accuracy', fontsize=12)\n", + "ax[1].set_title('Model Loss', fontsize=12)\n", + "fig.suptitle('Siamese Network: Learning Curve', fontsize=18);" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "id": "WIYIkFCew6mo" + }, + "source": [ + "# References\n", + "[Fisher Discriminant Triplet and Contrastive Losses for Training Siamese Networks](https://arxiv.org/pdf/2004.04674v1.pdf)" + ] + } + ], + "metadata": { + "accelerator": "GPU", + "colab": { + "collapsed_sections": [], + "name": "Siamese-Network.ipynb", + "provenance": [] + }, + "kernelspec": { + "display_name": "Python 3 (ipykernel)", + "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": "3.9.7" + } + }, + "nbformat": 4, + "nbformat_minor": 1 +}