Keras에서 모델 저장하고 불러오기 - Image classification with MNIST dataset.ipynb

{"nbformat":4,"nbformat_minor":0,"metadata":{"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":"3.6.8"},"colab":{"name":"Keras에서 모델 저장하고 불러오기 - Image classification with MNIST dataset.ipynb의 사본","version":"0.3.2","provenance":[{"file_id":"1Sa4daRdYmvT-F753-bGYae4OEp0V56OV","timestamp":1567145029529}]}},"cells":[{"cell_type":"markdown","metadata":{"id":"zCzvE0Dx5dVj","colab_type":"text"},"source":["# Keras에서 모델 저장하고 불러오기 - Image classification with MNIST dataset\n","MNIST image classification 코드 예제를 통해 tf.keras.callbacks.ModelCheckpoint를 이용하여 모델을 저장하고 불러오는 방법을 설명합니다."]},{"cell_type":"markdown","metadata":{"id":"qXgoX6cl5dV8","colab_type":"text"},"source":["### 라이브러리 import (h5py, tensorflow, keras 설치 필요)"]},{"cell_type":"code","metadata":{"id":"VVq2mhcB5dV9","colab_type":"code","outputId":"22125632-b268-49b2-8d80-e30400fb5c3f","colab":{}},"source":["from __future__ import absolute_import, division, print_function, unicode_literals, unicode_literals\n","\n","import os\n","import h5py\n","import tensorflow as tf\n","from tensorflow import keras\n","\n","print(tf.__version__)"],"execution_count":0,"outputs":[{"output_type":"stream","text":["/home/ubuntu/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:516: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n","  _np_qint8 = np.dtype([(\"qint8\", np.int8, 1)])\n","/home/ubuntu/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:517: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n","  _np_quint8 = np.dtype([(\"quint8\", np.uint8, 1)])\n","/home/ubuntu/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:518: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n","  _np_qint16 = np.dtype([(\"qint16\", np.int16, 1)])\n","/home/ubuntu/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:519: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n","  _np_quint16 = np.dtype([(\"quint16\", np.uint16, 1)])\n","/home/ubuntu/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:520: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n","  _np_qint32 = np.dtype([(\"qint32\", np.int32, 1)])\n","/home/ubuntu/.local/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:525: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n","  np_resource = np.dtype([(\"resource\", np.ubyte, 1)])\n"],"name":"stderr"},{"output_type":"stream","text":["1.14.0\n"],"name":"stdout"},{"output_type":"stream","text":["/home/ubuntu/.local/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:541: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n","  _np_qint8 = np.dtype([(\"qint8\", np.int8, 1)])\n","/home/ubuntu/.local/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:542: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n","  _np_quint8 = np.dtype([(\"quint8\", np.uint8, 1)])\n","/home/ubuntu/.local/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:543: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n","  _np_qint16 = np.dtype([(\"qint16\", np.int16, 1)])\n","/home/ubuntu/.local/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:544: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n","  _np_quint16 = np.dtype([(\"quint16\", np.uint16, 1)])\n","/home/ubuntu/.local/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:545: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n","  _np_qint32 = np.dtype([(\"qint32\", np.int32, 1)])\n","/home/ubuntu/.local/lib/python3.6/site-packages/tensorboard/compat/tensorflow_stub/dtypes.py:550: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n","  np_resource = np.dtype([(\"resource\", np.ubyte, 1)])\n"],"name":"stderr"}]},{"cell_type":"markdown","metadata":{"id":"4t0EylBo5dWD","colab_type":"text"},"source":["### MNIST 데이터셋 다운로드\n","데이터셋 이미지 중 첫 1000개만 사용"]},{"cell_type":"code","metadata":{"id":"2S3qLOeF5dWE","colab_type":"code","colab":{}},"source":["(train_images, train_labels), (test_images, test_labels) = tf.keras.datasets.mnist.load_data()\n","\n","train_labels = train_labels[:1000]\n","test_labels = test_labels[:1000]\n","\n","train_images = train_images[:1000].reshape(-1, 28 * 28) / 255.0\n","test_images = test_images[:1000].reshape(-1, 28 * 28) / 255.0"],"execution_count":0,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"cTzCxGo55dWJ","colab_type":"text"},"source":["### 모델 만들기"]},{"cell_type":"code","metadata":{"id":"SBcLe2645dWK","colab_type":"code","outputId":"b82c88db-6f11-4153-b1d8-4fbcd429736d","colab":{}},"source":["def create_model():\n","  model = tf.keras.models.Sequential([\n","    keras.layers.Dense(512, activation=tf.keras.activations.relu, input_shape=(784,)),\n","    keras.layers.Dropout(0.2),\n","    keras.layers.Dense(10, activation=tf.keras.activations.softmax)\n","  ])\n","\n","  model.compile(optimizer=tf.keras.optimizers.Adam(),\n","                loss=tf.keras.losses.sparse_categorical_crossentropy,\n","                metrics=['accuracy'])\n","\n","  return model\n","\n","\n","model = create_model()\n","model.summary()"],"execution_count":0,"outputs":[{"output_type":"stream","text":["WARNING:tensorflow:From /home/ubuntu/.local/lib/python3.6/site-packages/tensorflow/python/ops/init_ops.py:1251: calling VarianceScaling.__init__ (from tensorflow.python.ops.init_ops) with dtype is deprecated and will be removed in a future version.\n","Instructions for updating:\n","Call initializer instance with the dtype argument instead of passing it to the constructor\n","Model: \"sequential\"\n","_________________________________________________________________\n","Layer (type)                 Output Shape              Param #   \n","=================================================================\n","dense (Dense)                (None, 512)               401920    \n","_________________________________________________________________\n","dropout (Dropout)            (None, 512)               0         \n","_________________________________________________________________\n","dense_1 (Dense)              (None, 10)                5130      \n","=================================================================\n","Total params: 407,050\n","Trainable params: 407,050\n","Non-trainable params: 0\n","_________________________________________________________________\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"8aFWuBMc5dWP","colab_type":"text"},"source":["### 체크포인트 콜백 사용하여 체크포인트 저장, 학습 시작\n","체크포인트 파일을 만들고 매 epoch 종료시 파일 업데이트\n"]},{"cell_type":"code","metadata":{"id":"rEK3Cgo75dWQ","colab_type":"code","outputId":"2a0cc1d1-26c1-4a8b-bdf7-78069045f675","colab":{}},"source":["checkpoint_path = \"training_1/cp.ckpt\"\n","checkpoint_dir = os.path.dirname(checkpoint_path)\n","\n","# 체크포인트 콜백 만들기\n","cp_callback = tf.keras.callbacks.ModelCheckpoint(checkpoint_path,\n","                                                 save_weights_only=True,\n","                                                 verbose=1)\n","model = create_model()\n","\n","model. fit(train_images, train_labels, epochs = 10,\n","           validation_data = (test_images, test_labels),\n","           callbacks = [cp_callback])"],"execution_count":0,"outputs":[{"output_type":"stream","text":["Train on 1000 samples, validate on 1000 samples\n","Epoch 1/10\n"," 544/1000 [===============>..............] - ETA: 0s - loss: 1.4692 - acc: 0.5809\n","Epoch 00001: saving model to training_1/cp.ckpt\n","1000/1000 [==============================] - 0s 348us/sample - loss: 1.1189 - acc: 0.6800 - val_loss: 0.7396 - val_acc: 0.7700\n","Epoch 2/10\n"," 576/1000 [================>.............] - ETA: 0s - loss: 0.4708 - acc: 0.8785\n","Epoch 00002: saving model to training_1/cp.ckpt\n","1000/1000 [==============================] - 0s 153us/sample - loss: 0.4443 - acc: 0.8810 - val_loss: 0.5552 - val_acc: 0.8120\n","Epoch 3/10\n"," 544/1000 [===============>..............] - ETA: 0s - loss: 0.2701 - acc: 0.9357\n","Epoch 00003: saving model to training_1/cp.ckpt\n","1000/1000 [==============================] - 0s 154us/sample - loss: 0.2868 - acc: 0.9270 - val_loss: 0.4688 - val_acc: 0.8500\n","Epoch 4/10\n"," 544/1000 [===============>..............] - ETA: 0s - loss: 0.1821 - acc: 0.9651\n","Epoch 00004: saving model to training_1/cp.ckpt\n","1000/1000 [==============================] - 0s 153us/sample - loss: 0.2093 - acc: 0.9530 - val_loss: 0.4311 - val_acc: 0.8620\n","Epoch 5/10\n"," 544/1000 [===============>..............] - ETA: 0s - loss: 0.1750 - acc: 0.9614\n","Epoch 00005: saving model to training_1/cp.ckpt\n","1000/1000 [==============================] - 0s 155us/sample - loss: 0.1776 - acc: 0.9600 - val_loss: 0.4423 - val_acc: 0.8620\n","Epoch 6/10\n"," 576/1000 [================>.............] - ETA: 0s - loss: 0.1232 - acc: 0.9774\n","Epoch 00006: saving model to training_1/cp.ckpt\n","1000/1000 [==============================] - 0s 153us/sample - loss: 0.1151 - acc: 0.9830 - val_loss: 0.4205 - val_acc: 0.8620\n","Epoch 7/10\n"," 576/1000 [================>.............] - ETA: 0s - loss: 0.0845 - acc: 0.9913\n","Epoch 00007: saving model to training_1/cp.ckpt\n","1000/1000 [==============================] - 0s 153us/sample - loss: 0.0873 - acc: 0.9880 - val_loss: 0.3996 - val_acc: 0.8730\n","Epoch 8/10\n"," 544/1000 [===============>..............] - ETA: 0s - loss: 0.0683 - acc: 0.9926\n","Epoch 00008: saving model to training_1/cp.ckpt\n","1000/1000 [==============================] - 0s 153us/sample - loss: 0.0688 - acc: 0.9930 - val_loss: 0.4001 - val_acc: 0.8610\n","Epoch 9/10\n"," 576/1000 [================>.............] - ETA: 0s - loss: 0.0489 - acc: 0.9965\n","Epoch 00009: saving model to training_1/cp.ckpt\n","1000/1000 [==============================] - 0s 153us/sample - loss: 0.0520 - acc: 0.9970 - val_loss: 0.4044 - val_acc: 0.8690\n","Epoch 10/10\n"," 544/1000 [===============>..............] - ETA: 0s - loss: 0.0452 - acc: 0.9963\n","Epoch 00010: saving model to training_1/cp.ckpt\n","1000/1000 [==============================] - 0s 154us/sample - loss: 0.0413 - acc: 0.9980 - val_loss: 0.4161 - val_acc: 0.8680\n"],"name":"stdout"},{"output_type":"execute_result","data":{"text/plain":["<tensorflow.python.keras.callbacks.History at 0x7f06a28599e8>"]},"metadata":{"tags":[]},"execution_count":9}]},{"cell_type":"markdown","metadata":{"id":"ZGvKb2xy5dWU","colab_type":"text"},"source":["### 훈련되지 않은 새로운 모델 생성하여 테스트셋으로 성능 평가"]},{"cell_type":"code","metadata":{"id":"988zezZd5dWV","colab_type":"code","outputId":"b0417c20-45ab-4ea8-9ac4-d48d11f5e0a1","colab":{}},"source":["model = create_model()\n","\n","loss, acc = model.evaluate(test_images, test_labels)\n","print(\"학습되지 않은 모델 정확도: {:5.2f}%\".format(100*acc))"],"execution_count":0,"outputs":[{"output_type":"stream","text":["1000/1000 [==============================] - 0s 67us/sample - loss: 2.3299 - acc: 0.1270\n","훈련되지 않은 모델의 정확도: 12.70%\n"],"name":"stdout"}]},{"cell_type":"markdown","metadata":{"id":"zVy6LuWP5dWa","colab_type":"text"},"source":["### 저장된 체크포인트를 로드하여 다시 평가"]},{"cell_type":"code","metadata":{"id":"xkmKxylF5dWc","colab_type":"code","outputId":"06b20854-d19a-45f1-814f-4706326a6cf3","colab":{}},"source":["model.load_weights(checkpoint_path)\n","loss,acc = model.evaluate(test_images, test_labels)\n","print(\"복원된 모델 정확도: {:5.2f}%\".format(100*acc))"],"execution_count":0,"outputs":[{"output_type":"stream","text":["1000/1000 [==============================] - 0s 37us/sample - loss: 0.4161 - acc: 0.8680\n","복원된 모델의 정확도: 86.80%\n"],"name":"stdout"}]}]}