Adding baseline_model

baseline-model.ipynb

+{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.7.12","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"import tensorflow as tf\nfrom tensorflow import keras\nfrom keras import layers\nimport sklearn\nfrom sklearn.metrics import classification_report, confusion_matrix\nfrom keras.preprocessing.image import ImageDataGenerator\nimport matplotlib.pyplot as plt\nimport numpy as np\nimport itertools","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","execution":{"iopub.status.busy":"2022-11-25T05:09:47.442123Z","iopub.execute_input":"2022-11-25T05:09:47.442547Z","iopub.status.idle":"2022-11-25T05:09:53.017428Z","shell.execute_reply.started":"2022-11-25T05:09:47.442463Z","shell.execute_reply":"2022-11-25T05:09:53.016430Z"},"trusted":true},"execution_count":1,"outputs":[]},{"cell_type":"code","source":"train_ds_path = '/kaggle/input/cropimagedataset/idata/Image Dataset/ImageDataset/train/'\nvalid_ds_path = '/kaggle/input/cropimagedataset/idata/Image Dataset/ImageDataset/valid/'\n\ntrain_ds = keras.utils.image_dataset_from_directory(\n    directory=train_ds_path,\n    labels='inferred',\n    label_mode='categorical',\n    batch_size=32,\n    image_size=(224, 224))\n\nvalidation_ds = keras.utils.image_dataset_from_directory(\n    directory=valid_ds_path,\n    labels='inferred',\n    label_mode='categorical',\n#     batch_size=None,\n    batch_size=32,\n    image_size=(224, 224))","metadata":{"execution":{"iopub.status.busy":"2022-11-25T05:09:53.019860Z","iopub.execute_input":"2022-11-25T05:09:53.020643Z","iopub.status.idle":"2022-11-25T05:10:02.745234Z","shell.execute_reply.started":"2022-11-25T05:09:53.020599Z","shell.execute_reply":"2022-11-25T05:10:02.744302Z"},"trusted":true},"execution_count":2,"outputs":[{"name":"stdout","text":"Found 12009 files belonging to 6 classes.\n","output_type":"stream"},{"name":"stderr","text":"2022-11-25 05:09:59.189771: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-11-25 05:09:59.383665: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-11-25 05:09:59.384471: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-11-25 05:09:59.387380: I tensorflow/core/platform/cpu_feature_guard.cc:142] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 AVX512F FMA\nTo enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n2022-11-25 05:09:59.387649: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-11-25 05:09:59.388355: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-11-25 05:09:59.389079: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-11-25 05:10:02.062034: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-11-25 05:10:02.062880: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-11-25 05:10:02.063604: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:937] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n2022-11-25 05:10:02.064285: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1510] Created device /job:localhost/replica:0/task:0/device:GPU:0 with 15401 MB memory:  -> device: 0, name: Tesla P100-PCIE-16GB, pci bus id: 0000:00:04.0, compute capability: 6.0\n","output_type":"stream"},{"name":"stdout","text":"Found 1331 files belonging to 6 classes.\n","output_type":"stream"}]},{"cell_type":"code","source":"image_shape = (224, 224, 3)\nclasses = 6","metadata":{"execution":{"iopub.status.busy":"2022-11-25T05:10:02.749256Z","iopub.execute_input":"2022-11-25T05:10:02.751710Z","iopub.status.idle":"2022-11-25T05:10:02.757317Z","shell.execute_reply.started":"2022-11-25T05:10:02.751672Z","shell.execute_reply":"2022-11-25T05:10:02.755446Z"},"trusted":true},"execution_count":3,"outputs":[]},{"cell_type":"code","source":"baseline_model = keras.Sequential([\n    keras.Input(shape=image_shape),\n    layers.Conv2D(32, 3, activation='relu'),\n    layers.MaxPooling2D(2),\n    \n    layers.Conv2D(32, 3, strides=2, activation='relu'),\n    layers.MaxPooling2D(2),\n    \n    layers.Conv2D(64, 3, activation='relu'),\n    layers.MaxPooling2D(2),\n    \n    layers.Flatten(),\n    layers.Dense(64, activation='relu'),\n    layers.Dense(6, activation='softmax')\n])\nbaseline_model.summary()","metadata":{"execution":{"iopub.status.busy":"2022-11-25T05:10:02.760472Z","iopub.execute_input":"2022-11-25T05:10:02.761005Z","iopub.status.idle":"2022-11-25T05:10:02.893421Z","shell.execute_reply.started":"2022-11-25T05:10:02.760967Z","shell.execute_reply":"2022-11-25T05:10:02.892511Z"},"trusted":true},"execution_count":4,"outputs":[{"name":"stdout","text":"Model: \"sequential\"\n_________________________________________________________________\nLayer (type)                 Output Shape              Param #   \n=================================================================\nconv2d (Conv2D)              (None, 222, 222, 32)      896       \n_________________________________________________________________\nmax_pooling2d (MaxPooling2D) (None, 111, 111, 32)      0         \n_________________________________________________________________\nconv2d_1 (Conv2D)            (None, 55, 55, 32)        9248      \n_________________________________________________________________\nmax_pooling2d_1 (MaxPooling2 (None, 27, 27, 32)        0         \n_________________________________________________________________\nconv2d_2 (Conv2D)            (None, 25, 25, 64)        18496     \n_________________________________________________________________\nmax_pooling2d_2 (MaxPooling2 (None, 12, 12, 64)        0         \n_________________________________________________________________\nflatten (Flatten)            (None, 9216)              0         \n_________________________________________________________________\ndense (Dense)                (None, 64)                589888    \n_________________________________________________________________\ndense_1 (Dense)              (None, 6)                 390       \n=================================================================\nTotal params: 618,918\nTrainable params: 618,918\nNon-trainable params: 0\n_________________________________________________________________\n","output_type":"stream"}]},{"cell_type":"code","source":"Adam = keras.optimizers.Adam(learning_rate=0.0001, beta_1=0.9, beta_2=0.999, epsilon=None, decay=1e-5, amsgrad=False)\nbaseline_model.compile(optimizer=Adam, loss='categorical_crossentropy', metrics=['accuracy'])","metadata":{"execution":{"iopub.status.busy":"2022-11-25T05:10:02.897355Z","iopub.execute_input":"2022-11-25T05:10:02.897989Z","iopub.status.idle":"2022-11-25T05:10:02.917303Z","shell.execute_reply.started":"2022-11-25T05:10:02.897953Z","shell.execute_reply":"2022-11-25T05:10:02.916502Z"},"trusted":true},"execution_count":5,"outputs":[]},{"cell_type":"code","source":"epochs = 20\nbaseline_model.fit(train_ds, epochs=epochs, validation_data=validation_ds)","metadata":{"execution":{"iopub.status.busy":"2022-11-25T05:10:02.921122Z","iopub.execute_input":"2022-11-25T05:10:02.921791Z","iopub.status.idle":"2022-11-25T05:16:54.965351Z","shell.execute_reply.started":"2022-11-25T05:10:02.921739Z","shell.execute_reply":"2022-11-25T05:16:54.964365Z"},"trusted":true},"execution_count":6,"outputs":[{"name":"stdout","text":"Epoch 1/20\n","output_type":"stream"},{"name":"stderr","text":"2022-11-25 05:10:03.357457: I tensorflow/compiler/mlir/mlir_graph_optimization_pass.cc:185] None of the MLIR Optimization Passes are enabled (registered 2)\n2022-11-25 05:10:05.167767: I tensorflow/stream_executor/cuda/cuda_dnn.cc:369] Loaded cuDNN version 8005\n","output_type":"stream"},{"name":"stdout","text":"376/376 [==============================] - 40s 84ms/step - loss: 1.7829 - accuracy: 0.5432 - val_loss: 0.9859 - val_accuracy: 0.6619\nEpoch 2/20\n376/376 [==============================] - 19s 51ms/step - loss: 0.6016 - accuracy: 0.7883 - val_loss: 0.6113 - val_accuracy: 0.8054\nEpoch 3/20\n376/376 [==============================] - 18s 48ms/step - loss: 0.3862 - accuracy: 0.8708 - val_loss: 0.5834 - val_accuracy: 0.7934\nEpoch 4/20\n376/376 [==============================] - 18s 47ms/step - loss: 0.2756 - accuracy: 0.9103 - val_loss: 0.6506 - val_accuracy: 0.7701\nEpoch 5/20\n376/376 [==============================] - 18s 48ms/step - loss: 0.2005 - accuracy: 0.9353 - val_loss: 0.5044 - val_accuracy: 0.8490\nEpoch 6/20\n376/376 [==============================] - 18s 47ms/step - loss: 0.1395 - accuracy: 0.9574 - val_loss: 0.5171 - val_accuracy: 0.8234\nEpoch 7/20\n376/376 [==============================] - 19s 49ms/step - loss: 0.1031 - accuracy: 0.9694 - val_loss: 0.4270 - val_accuracy: 0.8768\nEpoch 8/20\n376/376 [==============================] - 19s 51ms/step - loss: 0.0841 - accuracy: 0.9756 - val_loss: 0.5460 - val_accuracy: 0.8580\nEpoch 9/20\n376/376 [==============================] - 18s 48ms/step - loss: 0.0530 - accuracy: 0.9859 - val_loss: 0.5130 - val_accuracy: 0.8355\nEpoch 10/20\n376/376 [==============================] - 18s 48ms/step - loss: 0.0404 - accuracy: 0.9906 - val_loss: 0.5209 - val_accuracy: 0.8663\nEpoch 11/20\n376/376 [==============================] - 18s 48ms/step - loss: 0.0368 - accuracy: 0.9910 - val_loss: 0.6774 - val_accuracy: 0.8520\nEpoch 12/20\n376/376 [==============================] - 18s 47ms/step - loss: 0.0392 - accuracy: 0.9893 - val_loss: 0.4939 - val_accuracy: 0.8738\nEpoch 13/20\n376/376 [==============================] - 19s 49ms/step - loss: 0.0203 - accuracy: 0.9959 - val_loss: 0.4904 - val_accuracy: 0.8881\nEpoch 14/20\n376/376 [==============================] - 18s 46ms/step - loss: 0.0094 - accuracy: 0.9991 - val_loss: 0.5356 - val_accuracy: 0.8783\nEpoch 15/20\n376/376 [==============================] - 19s 49ms/step - loss: 0.0486 - accuracy: 0.9849 - val_loss: 0.5375 - val_accuracy: 0.8490\nEpoch 16/20\n376/376 [==============================] - 19s 49ms/step - loss: 0.0207 - accuracy: 0.9948 - val_loss: 0.5074 - val_accuracy: 0.8903\nEpoch 17/20\n376/376 [==============================] - 18s 46ms/step - loss: 0.0056 - accuracy: 0.9995 - val_loss: 0.5810 - val_accuracy: 0.8760\nEpoch 18/20\n376/376 [==============================] - 19s 49ms/step - loss: 0.0051 - accuracy: 0.9995 - val_loss: 0.5438 - val_accuracy: 0.8926\nEpoch 19/20\n376/376 [==============================] - 18s 47ms/step - loss: 0.0521 - accuracy: 0.9830 - val_loss: 0.5872 - val_accuracy: 0.8813\nEpoch 20/20\n376/376 [==============================] - 19s 49ms/step - loss: 0.0091 - accuracy: 0.9987 - val_loss: 0.5567 - val_accuracy: 0.8881\n","output_type":"stream"},{"execution_count":6,"output_type":"execute_result","data":{"text/plain":"<keras.callbacks.History at 0x7f70b5417410>"},"metadata":{}}]},{"cell_type":"code","source":"history = baseline_model.history\nplt.plot(history.history['accuracy'])\nplt.plot(history.history['val_accuracy'])\nplt.title('model accuracy')\nplt.ylabel('accuracy')\nplt.xlabel('epoch')\nplt.legend(['train', 'val'], loc='best')\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2022-11-25T05:16:54.967036Z","iopub.execute_input":"2022-11-25T05:16:54.967415Z","iopub.status.idle":"2022-11-25T05:16:55.201764Z","shell.execute_reply.started":"2022-11-25T05:16:54.967361Z","shell.execute_reply":"2022-11-25T05:16:55.200726Z"},"trusted":true},"execution_count":7,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 432x288 with 1 Axes>","image/png":"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\n"},"metadata":{"needs_background":"light"}}]},{"cell_type":"code","source":"plt.plot(history.history['loss'])\nplt.plot(history.history['val_loss'])\nplt.title('model loss')\nplt.ylabel('loss')\nplt.xlabel('epoch')\nplt.legend(['train', 'val'], loc='best')\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2022-11-25T05:16:55.204321Z","iopub.execute_input":"2022-11-25T05:16:55.204717Z","iopub.status.idle":"2022-11-25T05:16:55.413250Z","shell.execute_reply.started":"2022-11-25T05:16:55.204680Z","shell.execute_reply":"2022-11-25T05:16:55.412287Z"},"trusted":true},"execution_count":8,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 432x288 with 1 Axes>","image/png":"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\n"},"metadata":{"needs_background":"light"}}]},{"cell_type":"code","source":"test_ds = keras.utils.image_dataset_from_directory(\n#     directory='/Users/manasgabani/Downloads/IITB/CS725/project/idata/Image Dataset/test_data/test_dup/',\n    directory='/kaggle/input/testdata/Test/',\n    labels='inferred',\n    label_mode='categorical',\n    batch_size=1,\n    image_size=(224, 224))","metadata":{"execution":{"iopub.status.busy":"2022-11-25T05:16:55.414631Z","iopub.execute_input":"2022-11-25T05:16:55.415717Z","iopub.status.idle":"2022-11-25T05:16:55.650770Z","shell.execute_reply.started":"2022-11-25T05:16:55.415679Z","shell.execute_reply":"2022-11-25T05:16:55.649749Z"},"trusted":true},"execution_count":9,"outputs":[{"name":"stdout","text":"Found 1330 files belonging to 6 classes.\n","output_type":"stream"}]},{"cell_type":"code","source":"predictions = np.array([])\nlabels =  np.array([])\nfor x, y in test_ds:\n    predictions = np.concatenate([predictions, np.argmax(baseline_model.predict(x), axis = -1)])\n    labels = np.concatenate([labels, np.argmax(y.numpy(), axis=-1)])","metadata":{"execution":{"iopub.status.busy":"2022-11-25T05:16:55.654129Z","iopub.execute_input":"2022-11-25T05:16:55.654435Z","iopub.status.idle":"2022-11-25T05:17:44.304992Z","shell.execute_reply.started":"2022-11-25T05:16:55.654406Z","shell.execute_reply":"2022-11-25T05:17:44.304033Z"},"trusted":true},"execution_count":10,"outputs":[]},{"cell_type":"code","source":"# conf_matrix = confusion_matrix(y_true=test_labels, y_pred=test_predictions.argmax(axis=1))\nconf_matrix = confusion_matrix(y_true=labels, y_pred=predictions)\nconf_matrix_plot_lables = test_ds.class_names\nconf_matrix","metadata":{"execution":{"iopub.status.busy":"2022-11-25T05:17:44.306826Z","iopub.execute_input":"2022-11-25T05:17:44.307233Z","iopub.status.idle":"2022-11-25T05:17:44.318061Z","shell.execute_reply.started":"2022-11-25T05:17:44.307198Z","shell.execute_reply":"2022-11-25T05:17:44.317015Z"},"trusted":true},"execution_count":11,"outputs":[{"execution_count":11,"output_type":"execute_result","data":{"text/plain":"array([[209,   0,   7,   2,   0,   3],\n       [  0,  27,   3,   0,   0,   0],\n       [  8,   3, 193,   4,   0,  14],\n       [  2,   2,  10, 166,   6,  36],\n       [  0,   0,   0,   4, 212,   0],\n       [  4,   3,  18,  27,   4, 363]])"},"metadata":{}}]},{"cell_type":"code","source":"plt.imshow(conf_matrix, interpolation='nearest', cmap=plt.cm.Blues)\nplt.title('Confusion Matrix')\nplt.colorbar()\ntick_marks = np.arange(len(conf_matrix_plot_lables))\nplt.xticks(tick_marks, conf_matrix_plot_lables, rotation=45)\nplt.yticks(tick_marks, conf_matrix_plot_lables)\n\nthresh = conf_matrix.max() / 2.\nfor i, j in itertools.product(range(conf_matrix.shape[0]), range(conf_matrix.shape[1])):\n    plt.text(j, i, conf_matrix[i, j],\n        horizontalalignment=\"center\",\n        color=\"white\" if conf_matrix[i, j] > thresh else \"black\")\n\nplt.tight_layout()\nplt.ylabel('True label')\nplt.xlabel('Predicted label')","metadata":{"execution":{"iopub.status.busy":"2022-11-25T05:17:44.319666Z","iopub.execute_input":"2022-11-25T05:17:44.320078Z","iopub.status.idle":"2022-11-25T05:17:44.943240Z","shell.execute_reply.started":"2022-11-25T05:17:44.320043Z","shell.execute_reply":"2022-11-25T05:17:44.942354Z"},"trusted":true},"execution_count":12,"outputs":[{"execution_count":12,"output_type":"execute_result","data":{"text/plain":"Text(0.5, -37.95460061251481, 'Predicted label')"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"<Figure size 432x288 with 2 Axes>","image/png":"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\n"},"metadata":{"needs_background":"light"}}]},{"cell_type":"code","source":"test_score = baseline_model.evaluate(test_ds,\n                                 batch_size=32,\n                                 verbose=0)\ntest_score","metadata":{"execution":{"iopub.status.busy":"2022-11-25T05:17:44.947741Z","iopub.execute_input":"2022-11-25T05:17:44.949882Z","iopub.status.idle":"2022-11-25T05:17:50.085095Z","shell.execute_reply.started":"2022-11-25T05:17:44.949845Z","shell.execute_reply":"2022-11-25T05:17:50.083933Z"},"trusted":true},"execution_count":13,"outputs":[{"execution_count":13,"output_type":"execute_result","data":{"text/plain":"[0.6564521789550781, 0.8796992301940918]"},"metadata":{}}]},{"cell_type":"code","source":"test_score = baseline_model.evaluate(test_ds,\n                                 batch_size=None,\n                                 verbose=0)\ntest_score","metadata":{"execution":{"iopub.status.busy":"2022-11-25T05:17:50.087630Z","iopub.execute_input":"2022-11-25T05:17:50.088351Z","iopub.status.idle":"2022-11-25T05:17:55.216026Z","shell.execute_reply.started":"2022-11-25T05:17:50.088313Z","shell.execute_reply":"2022-11-25T05:17:55.214929Z"},"trusted":true},"execution_count":14,"outputs":[{"execution_count":14,"output_type":"execute_result","data":{"text/plain":"[0.6564521193504333, 0.8796992301940918]"},"metadata":{}}]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]}]}
\ No newline at end of file