[hint]Solution6_TrainSomeLayers.ipynb

{"cells":[{"cell_type":"markdown","source":["## Load Data"],"metadata":{"id":"9Jx0UG8WZPQ0"},"id":"9Jx0UG8WZPQ0"},{"cell_type":"code","execution_count":1,"id":"e175eeac","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":4765,"status":"ok","timestamp":1669968158333,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"},"user_tz":-420},"id":"e175eeac","outputId":"d0ce7c27-f59d-4aa2-9d3d-418a0d54a6f7"},"outputs":[{"output_type":"stream","name":"stdout","text":["Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/train-labels-idx1-ubyte.gz\n","29515/29515 [==============================] - 0s 0us/step\n","Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/train-images-idx3-ubyte.gz\n","26421880/26421880 [==============================] - 0s 0us/step\n","Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/t10k-labels-idx1-ubyte.gz\n","5148/5148 [==============================] - 0s 0us/step\n","Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/t10k-images-idx3-ubyte.gz\n","4422102/4422102 [==============================] - 0s 0us/step\n","(60000, 28, 28)\n","(60000,)\n","(10000, 28, 28)\n","(10000,)\n"]}],"source":["# fashion_mnist\n","import tensorflow as tf\n","import tensorflow.keras as keras\n","import numpy as np\n","tf.keras.utils.set_random_seed(1)\n","tf.config.experimental.enable_op_determinism()\n","\n","(X_train, y_train), (X_test, y_test) = keras.datasets.fashion_mnist.load_data()\n","print(X_train.shape)\n","print(y_train.shape)\n","print(X_test.shape)\n","print(y_test.shape)"]},{"cell_type":"markdown","source":["## Reshape Data"],"metadata":{"id":"f068e1ea"},"id":"f068e1ea"},{"cell_type":"code","execution_count":2,"id":"8aabf29a","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":11,"status":"ok","timestamp":1669968158333,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"},"user_tz":-420},"id":"8aabf29a","outputId":"79822f28-486a-4584-b48c-d1ae241e4e54"},"outputs":[{"output_type":"stream","name":"stdout","text":["(60000, 784)\n","(60000,)\n","(10000, 784)\n","(10000,)\n"]}],"source":["X_train = X_train.reshape(60000, -1)\n","X_test  = X_test.reshape(10000, -1)\n","\n","print(X_train.shape)\n","print(y_train.shape)\n","print(X_test.shape)\n","print(y_test.shape)"]},{"cell_type":"markdown","source":["## Train Some Layers"],"metadata":{"id":"c3167cdb"},"id":"c3167cdb"},{"cell_type":"code","execution_count":3,"id":"c46b65da","metadata":{"id":"c46b65da","executionInfo":{"status":"ok","timestamp":1669968158333,"user_tz":-420,"elapsed":6,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"}}},"outputs":[],"source":["# solution 6: Train some layers\n","\n","batch_size = 512\n","initializer = tf.keras.initializers.RandomNormal()\n","\n","# Ví dụ sub model 1\n","first = keras.Sequential()\n","first.add(keras.layers.Dense(128, activation='sigmoid',\n","                             kernel_initializer=initializer))\n","first.add(keras.layers.Dense(128, activation='sigmoid',\n","                             kernel_initializer=initializer))\n","\n","#################### YOUR CODE HERE ####################\n","# Các bạn thực hiện sub model 2 tương tự như cách thức của Ví dụ sub model 1\n","second = keras.Sequential()\n","second.add(keras.layers.Dense(128, activation='sigmoid',\n","                              kernel_initializer=initializer))\n","second.add(keras.layers.Dense(128, activation='sigmoid',\n","                              kernel_initializer=initializer))\n","########################################################\n","\n","#################### YOUR CODE HERE ####################\n","# Các bạn thực hiện sub model 3 tương tự như cách thức của Ví dụ sub model 1\n","third = keras.Sequential()\n","third.add(keras.layers.Dense(128, activation='sigmoid',\n","                              kernel_initializer=initializer))\n","third.add(keras.layers.Dense(128, activation='sigmoid',\n","                              kernel_initializer=initializer))\n","########################################################\n","\n","#################### YOUR CODE HERE ####################\n","# Các bạn thực hiện sub model 4 tương tự như cách thức của Ví dụ sub model 1\n","fourth = keras.Sequential()\n","fourth.add(keras.layers.Dense(128, activation='sigmoid',\n","                              kernel_initializer=initializer))\n","########################################################"]},{"cell_type":"markdown","source":["### Train 1"],"metadata":{"id":"uEDi740xZ6ka"},"id":"uEDi740xZ6ka"},{"cell_type":"code","execution_count":4,"id":"248dd07e","metadata":{"id":"248dd07e","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1669968301981,"user_tz":-420,"elapsed":143654,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"}},"outputId":"cc18d5a6-1d3a-4531-9c0b-108a37de6e83"},"outputs":[{"output_type":"stream","name":"stdout","text":["Epoch 1/100\n","118/118 - 4s - loss: 2.2259 - accuracy: 0.3154 - val_loss: 2.1474 - val_accuracy: 0.4716 - 4s/epoch - 32ms/step\n","Epoch 2/100\n","118/118 - 2s - loss: 2.0814 - accuracy: 0.5499 - val_loss: 2.0170 - val_accuracy: 0.5804 - 2s/epoch - 19ms/step\n","Epoch 3/100\n","118/118 - 2s - loss: 1.9466 - accuracy: 0.6050 - val_loss: 1.8824 - val_accuracy: 0.6063 - 2s/epoch - 16ms/step\n","Epoch 4/100\n","118/118 - 1s - loss: 1.8130 - accuracy: 0.6331 - val_loss: 1.7517 - val_accuracy: 0.6635 - 1s/epoch - 11ms/step\n","Epoch 5/100\n","118/118 - 1s - loss: 1.6847 - accuracy: 0.6602 - val_loss: 1.6287 - val_accuracy: 0.6625 - 1s/epoch - 11ms/step\n","Epoch 6/100\n","118/118 - 1s - loss: 1.5674 - accuracy: 0.6716 - val_loss: 1.5195 - val_accuracy: 0.6701 - 1s/epoch - 11ms/step\n","Epoch 7/100\n","118/118 - 1s - loss: 1.4633 - accuracy: 0.6819 - val_loss: 1.4224 - val_accuracy: 0.6785 - 1s/epoch - 11ms/step\n","Epoch 8/100\n","118/118 - 1s - loss: 1.3708 - accuracy: 0.6891 - val_loss: 1.3369 - val_accuracy: 0.6863 - 1s/epoch - 11ms/step\n","Epoch 9/100\n","118/118 - 1s - loss: 1.2896 - accuracy: 0.6985 - val_loss: 1.2609 - val_accuracy: 0.6880 - 1s/epoch - 11ms/step\n","Epoch 10/100\n","118/118 - 1s - loss: 1.2169 - accuracy: 0.7039 - val_loss: 1.1923 - val_accuracy: 0.7019 - 1s/epoch - 11ms/step\n","Epoch 11/100\n","118/118 - 1s - loss: 1.1521 - accuracy: 0.7141 - val_loss: 1.1323 - val_accuracy: 0.7116 - 1s/epoch - 11ms/step\n","Epoch 12/100\n","118/118 - 1s - loss: 1.0944 - accuracy: 0.7205 - val_loss: 1.0792 - val_accuracy: 0.7177 - 1s/epoch - 11ms/step\n","Epoch 13/100\n","118/118 - 1s - loss: 1.0427 - accuracy: 0.7301 - val_loss: 1.0313 - val_accuracy: 0.7224 - 1s/epoch - 11ms/step\n","Epoch 14/100\n","118/118 - 1s - loss: 0.9970 - accuracy: 0.7363 - val_loss: 0.9887 - val_accuracy: 0.7260 - 1s/epoch - 11ms/step\n","Epoch 15/100\n","118/118 - 1s - loss: 0.9557 - accuracy: 0.7422 - val_loss: 0.9519 - val_accuracy: 0.7332 - 1s/epoch - 11ms/step\n","Epoch 16/100\n","118/118 - 1s - loss: 0.9185 - accuracy: 0.7478 - val_loss: 0.9155 - val_accuracy: 0.7445 - 1s/epoch - 11ms/step\n","Epoch 17/100\n","118/118 - 1s - loss: 0.8849 - accuracy: 0.7551 - val_loss: 0.8852 - val_accuracy: 0.7482 - 1s/epoch - 11ms/step\n","Epoch 18/100\n","118/118 - 1s - loss: 0.8539 - accuracy: 0.7604 - val_loss: 0.8570 - val_accuracy: 0.7505 - 1s/epoch - 11ms/step\n","Epoch 19/100\n","118/118 - 1s - loss: 0.8260 - accuracy: 0.7655 - val_loss: 0.8305 - val_accuracy: 0.7564 - 1s/epoch - 11ms/step\n","Epoch 20/100\n","118/118 - 1s - loss: 0.8004 - accuracy: 0.7698 - val_loss: 0.8055 - val_accuracy: 0.7635 - 1s/epoch - 11ms/step\n","Epoch 21/100\n","118/118 - 1s - loss: 0.7768 - accuracy: 0.7760 - val_loss: 0.7860 - val_accuracy: 0.7615 - 1s/epoch - 11ms/step\n","Epoch 22/100\n","118/118 - 1s - loss: 0.7552 - accuracy: 0.7800 - val_loss: 0.7637 - val_accuracy: 0.7684 - 1s/epoch - 11ms/step\n","Epoch 23/100\n","118/118 - 1s - loss: 0.7344 - accuracy: 0.7837 - val_loss: 0.7473 - val_accuracy: 0.7760 - 1s/epoch - 10ms/step\n","Epoch 24/100\n","118/118 - 1s - loss: 0.7168 - accuracy: 0.7864 - val_loss: 0.7291 - val_accuracy: 0.7763 - 1s/epoch - 11ms/step\n","Epoch 25/100\n","118/118 - 1s - loss: 0.6994 - accuracy: 0.7910 - val_loss: 0.7114 - val_accuracy: 0.7763 - 1s/epoch - 11ms/step\n","Epoch 26/100\n","118/118 - 1s - loss: 0.6832 - accuracy: 0.7954 - val_loss: 0.6970 - val_accuracy: 0.7829 - 1s/epoch - 11ms/step\n","Epoch 27/100\n","118/118 - 1s - loss: 0.6681 - accuracy: 0.7982 - val_loss: 0.6845 - val_accuracy: 0.7849 - 1s/epoch - 11ms/step\n","Epoch 28/100\n","118/118 - 1s - loss: 0.6541 - accuracy: 0.8002 - val_loss: 0.6693 - val_accuracy: 0.7886 - 1s/epoch - 10ms/step\n","Epoch 29/100\n","118/118 - 1s - loss: 0.6404 - accuracy: 0.8039 - val_loss: 0.6585 - val_accuracy: 0.7908 - 1s/epoch - 11ms/step\n","Epoch 30/100\n","118/118 - 1s - loss: 0.6287 - accuracy: 0.8066 - val_loss: 0.6485 - val_accuracy: 0.7958 - 1s/epoch - 11ms/step\n","Epoch 31/100\n","118/118 - 1s - loss: 0.6167 - accuracy: 0.8086 - val_loss: 0.6377 - val_accuracy: 0.7964 - 1s/epoch - 10ms/step\n","Epoch 32/100\n","118/118 - 1s - loss: 0.6073 - accuracy: 0.8106 - val_loss: 0.6271 - val_accuracy: 0.7981 - 1s/epoch - 10ms/step\n","Epoch 33/100\n","118/118 - 1s - loss: 0.5962 - accuracy: 0.8128 - val_loss: 0.6189 - val_accuracy: 0.7986 - 1s/epoch - 10ms/step\n","Epoch 34/100\n","118/118 - 1s - loss: 0.5864 - accuracy: 0.8155 - val_loss: 0.6093 - val_accuracy: 0.8017 - 1s/epoch - 10ms/step\n","Epoch 35/100\n","118/118 - 1s - loss: 0.5772 - accuracy: 0.8172 - val_loss: 0.6013 - val_accuracy: 0.8056 - 1s/epoch - 10ms/step\n","Epoch 36/100\n","118/118 - 1s - loss: 0.5687 - accuracy: 0.8191 - val_loss: 0.5939 - val_accuracy: 0.8059 - 1s/epoch - 11ms/step\n","Epoch 37/100\n","118/118 - 1s - loss: 0.5604 - accuracy: 0.8222 - val_loss: 0.5876 - val_accuracy: 0.8049 - 1s/epoch - 11ms/step\n","Epoch 38/100\n","118/118 - 1s - loss: 0.5535 - accuracy: 0.8233 - val_loss: 0.5812 - val_accuracy: 0.8067 - 1s/epoch - 10ms/step\n","Epoch 39/100\n","118/118 - 1s - loss: 0.5452 - accuracy: 0.8248 - val_loss: 0.5715 - val_accuracy: 0.8105 - 1s/epoch - 10ms/step\n","Epoch 40/100\n","118/118 - 1s - loss: 0.5383 - accuracy: 0.8260 - val_loss: 0.5646 - val_accuracy: 0.8136 - 1s/epoch - 10ms/step\n","Epoch 41/100\n","118/118 - 1s - loss: 0.5324 - accuracy: 0.8279 - val_loss: 0.5640 - val_accuracy: 0.8097 - 1s/epoch - 10ms/step\n","Epoch 42/100\n","118/118 - 1s - loss: 0.5268 - accuracy: 0.8293 - val_loss: 0.5551 - val_accuracy: 0.8123 - 1s/epoch - 10ms/step\n","Epoch 43/100\n","118/118 - 1s - loss: 0.5210 - accuracy: 0.8302 - val_loss: 0.5509 - val_accuracy: 0.8152 - 1s/epoch - 11ms/step\n","Epoch 44/100\n","118/118 - 1s - loss: 0.5153 - accuracy: 0.8324 - val_loss: 0.5449 - val_accuracy: 0.8176 - 1s/epoch - 11ms/step\n","Epoch 45/100\n","118/118 - 1s - loss: 0.5101 - accuracy: 0.8327 - val_loss: 0.5413 - val_accuracy: 0.8194 - 1s/epoch - 11ms/step\n","Epoch 46/100\n","118/118 - 1s - loss: 0.5052 - accuracy: 0.8339 - val_loss: 0.5380 - val_accuracy: 0.8165 - 1s/epoch - 10ms/step\n","Epoch 47/100\n","118/118 - 1s - loss: 0.5005 - accuracy: 0.8348 - val_loss: 0.5378 - val_accuracy: 0.8170 - 1s/epoch - 10ms/step\n","Epoch 48/100\n","118/118 - 1s - loss: 0.4959 - accuracy: 0.8356 - val_loss: 0.5281 - val_accuracy: 0.8189 - 1s/epoch - 10ms/step\n","Epoch 49/100\n","118/118 - 1s - loss: 0.4910 - accuracy: 0.8370 - val_loss: 0.5253 - val_accuracy: 0.8209 - 1s/epoch - 11ms/step\n","Epoch 50/100\n","118/118 - 1s - loss: 0.4869 - accuracy: 0.8376 - val_loss: 0.5236 - val_accuracy: 0.8204 - 1s/epoch - 10ms/step\n","Epoch 51/100\n","118/118 - 1s - loss: 0.4832 - accuracy: 0.8391 - val_loss: 0.5179 - val_accuracy: 0.8224 - 1s/epoch - 10ms/step\n","Epoch 52/100\n","118/118 - 1s - loss: 0.4796 - accuracy: 0.8397 - val_loss: 0.5140 - val_accuracy: 0.8221 - 1s/epoch - 11ms/step\n","Epoch 53/100\n","118/118 - 1s - loss: 0.4745 - accuracy: 0.8407 - val_loss: 0.5109 - val_accuracy: 0.8245 - 1s/epoch - 11ms/step\n","Epoch 54/100\n","118/118 - 1s - loss: 0.4717 - accuracy: 0.8415 - val_loss: 0.5113 - val_accuracy: 0.8223 - 1s/epoch - 11ms/step\n","Epoch 55/100\n","118/118 - 1s - loss: 0.4689 - accuracy: 0.8419 - val_loss: 0.5070 - val_accuracy: 0.8251 - 1s/epoch - 10ms/step\n","Epoch 56/100\n","118/118 - 1s - loss: 0.4650 - accuracy: 0.8442 - val_loss: 0.5029 - val_accuracy: 0.8266 - 1s/epoch - 10ms/step\n","Epoch 57/100\n","118/118 - 1s - loss: 0.4618 - accuracy: 0.8450 - val_loss: 0.5017 - val_accuracy: 0.8269 - 1s/epoch - 10ms/step\n","Epoch 58/100\n","118/118 - 1s - loss: 0.4594 - accuracy: 0.8449 - val_loss: 0.5002 - val_accuracy: 0.8306 - 1s/epoch - 10ms/step\n","Epoch 59/100\n","118/118 - 1s - loss: 0.4560 - accuracy: 0.8450 - val_loss: 0.4964 - val_accuracy: 0.8280 - 1s/epoch - 10ms/step\n","Epoch 60/100\n","118/118 - 1s - loss: 0.4517 - accuracy: 0.8466 - val_loss: 0.4942 - val_accuracy: 0.8278 - 1s/epoch - 10ms/step\n","Epoch 61/100\n","118/118 - 1s - loss: 0.4479 - accuracy: 0.8476 - val_loss: 0.4921 - val_accuracy: 0.8310 - 1s/epoch - 10ms/step\n","Epoch 62/100\n","118/118 - 1s - loss: 0.4461 - accuracy: 0.8490 - val_loss: 0.4895 - val_accuracy: 0.8327 - 1s/epoch - 11ms/step\n","Epoch 63/100\n","118/118 - 1s - loss: 0.4437 - accuracy: 0.8489 - val_loss: 0.4859 - val_accuracy: 0.8343 - 1s/epoch - 11ms/step\n","Epoch 64/100\n","118/118 - 1s - loss: 0.4423 - accuracy: 0.8497 - val_loss: 0.4835 - val_accuracy: 0.8338 - 1s/epoch - 11ms/step\n","Epoch 65/100\n","118/118 - 1s - loss: 0.4385 - accuracy: 0.8495 - val_loss: 0.4820 - val_accuracy: 0.8309 - 1s/epoch - 11ms/step\n","Epoch 66/100\n","118/118 - 1s - loss: 0.4363 - accuracy: 0.8514 - val_loss: 0.4841 - val_accuracy: 0.8304 - 1s/epoch - 11ms/step\n","Epoch 67/100\n","118/118 - 1s - loss: 0.4334 - accuracy: 0.8511 - val_loss: 0.4766 - val_accuracy: 0.8332 - 1s/epoch - 11ms/step\n","Epoch 68/100\n","118/118 - 1s - loss: 0.4313 - accuracy: 0.8522 - val_loss: 0.4776 - val_accuracy: 0.8341 - 1s/epoch - 11ms/step\n","Epoch 69/100\n","118/118 - 1s - loss: 0.4301 - accuracy: 0.8530 - val_loss: 0.4773 - val_accuracy: 0.8364 - 1s/epoch - 11ms/step\n","Epoch 70/100\n","118/118 - 1s - loss: 0.4275 - accuracy: 0.8532 - val_loss: 0.4764 - val_accuracy: 0.8341 - 1s/epoch - 10ms/step\n","Epoch 71/100\n","118/118 - 1s - loss: 0.4251 - accuracy: 0.8540 - val_loss: 0.4810 - val_accuracy: 0.8353 - 1s/epoch - 11ms/step\n","Epoch 72/100\n","118/118 - 1s - loss: 0.4245 - accuracy: 0.8537 - val_loss: 0.4783 - val_accuracy: 0.8319 - 1s/epoch - 10ms/step\n","Epoch 73/100\n","118/118 - 1s - loss: 0.4245 - accuracy: 0.8549 - val_loss: 0.4771 - val_accuracy: 0.8324 - 1s/epoch - 11ms/step\n","Epoch 74/100\n","118/118 - 1s - loss: 0.4216 - accuracy: 0.8553 - val_loss: 0.4737 - val_accuracy: 0.8320 - 1s/epoch - 11ms/step\n","Epoch 75/100\n","118/118 - 1s - loss: 0.4165 - accuracy: 0.8567 - val_loss: 0.4659 - val_accuracy: 0.8371 - 1s/epoch - 10ms/step\n","Epoch 76/100\n","118/118 - 1s - loss: 0.4160 - accuracy: 0.8574 - val_loss: 0.4694 - val_accuracy: 0.8341 - 1s/epoch - 11ms/step\n","Epoch 77/100\n","118/118 - 1s - loss: 0.4147 - accuracy: 0.8574 - val_loss: 0.4698 - val_accuracy: 0.8367 - 1s/epoch - 10ms/step\n","Epoch 78/100\n","118/118 - 1s - loss: 0.4132 - accuracy: 0.8584 - val_loss: 0.4644 - val_accuracy: 0.8380 - 1s/epoch - 10ms/step\n","Epoch 79/100\n","118/118 - 1s - loss: 0.4104 - accuracy: 0.8589 - val_loss: 0.4627 - val_accuracy: 0.8386 - 1s/epoch - 10ms/step\n","Epoch 80/100\n","118/118 - 1s - loss: 0.4115 - accuracy: 0.8586 - val_loss: 0.4612 - val_accuracy: 0.8389 - 1s/epoch - 10ms/step\n","Epoch 81/100\n","118/118 - 1s - loss: 0.4092 - accuracy: 0.8595 - val_loss: 0.4622 - val_accuracy: 0.8376 - 1s/epoch - 10ms/step\n","Epoch 82/100\n","118/118 - 1s - loss: 0.4049 - accuracy: 0.8605 - val_loss: 0.4546 - val_accuracy: 0.8414 - 1s/epoch - 10ms/step\n","Epoch 83/100\n","118/118 - 1s - loss: 0.4033 - accuracy: 0.8602 - val_loss: 0.4574 - val_accuracy: 0.8413 - 1s/epoch - 10ms/step\n","Epoch 84/100\n","118/118 - 1s - loss: 0.4024 - accuracy: 0.8615 - val_loss: 0.4560 - val_accuracy: 0.8417 - 1s/epoch - 10ms/step\n","Epoch 85/100\n","118/118 - 1s - loss: 0.4004 - accuracy: 0.8621 - val_loss: 0.4548 - val_accuracy: 0.8449 - 1s/epoch - 10ms/step\n","Epoch 86/100\n","118/118 - 1s - loss: 0.3999 - accuracy: 0.8614 - val_loss: 0.4544 - val_accuracy: 0.8410 - 1s/epoch - 11ms/step\n","Epoch 87/100\n","118/118 - 1s - loss: 0.4001 - accuracy: 0.8611 - val_loss: 0.4543 - val_accuracy: 0.8405 - 1s/epoch - 10ms/step\n","Epoch 88/100\n","118/118 - 1s - loss: 0.3961 - accuracy: 0.8635 - val_loss: 0.4583 - val_accuracy: 0.8354 - 1s/epoch - 10ms/step\n","Epoch 89/100\n","118/118 - 1s - loss: 0.3946 - accuracy: 0.8626 - val_loss: 0.4539 - val_accuracy: 0.8377 - 1s/epoch - 10ms/step\n","Epoch 90/100\n","118/118 - 1s - loss: 0.3917 - accuracy: 0.8642 - val_loss: 0.4479 - val_accuracy: 0.8419 - 1s/epoch - 10ms/step\n","Epoch 91/100\n","118/118 - 1s - loss: 0.3920 - accuracy: 0.8646 - val_loss: 0.4513 - val_accuracy: 0.8401 - 1s/epoch - 10ms/step\n","Epoch 92/100\n","118/118 - 1s - loss: 0.3911 - accuracy: 0.8635 - val_loss: 0.4546 - val_accuracy: 0.8373 - 1s/epoch - 10ms/step\n","Epoch 93/100\n","118/118 - 1s - loss: 0.3896 - accuracy: 0.8648 - val_loss: 0.4483 - val_accuracy: 0.8427 - 1s/epoch - 10ms/step\n","Epoch 94/100\n","118/118 - 1s - loss: 0.3878 - accuracy: 0.8659 - val_loss: 0.4464 - val_accuracy: 0.8415 - 1s/epoch - 10ms/step\n","Epoch 95/100\n","118/118 - 1s - loss: 0.3855 - accuracy: 0.8659 - val_loss: 0.4429 - val_accuracy: 0.8432 - 1s/epoch - 10ms/step\n","Epoch 96/100\n","118/118 - 1s - loss: 0.3868 - accuracy: 0.8661 - val_loss: 0.4443 - val_accuracy: 0.8407 - 1s/epoch - 10ms/step\n","Epoch 97/100\n","118/118 - 1s - loss: 0.3833 - accuracy: 0.8669 - val_loss: 0.4453 - val_accuracy: 0.8458 - 1s/epoch - 10ms/step\n","Epoch 98/100\n","118/118 - 1s - loss: 0.3807 - accuracy: 0.8674 - val_loss: 0.4440 - val_accuracy: 0.8400 - 1s/epoch - 10ms/step\n","Epoch 99/100\n","118/118 - 1s - loss: 0.3812 - accuracy: 0.8678 - val_loss: 0.4502 - val_accuracy: 0.8392 - 1s/epoch - 10ms/step\n","Epoch 100/100\n","118/118 - 1s - loss: 0.3809 - accuracy: 0.8675 - val_loss: 0.4401 - val_accuracy: 0.8427 - 1s/epoch - 10ms/step\n"]}],"source":["# Ví dụ train lần 1 với sub model 1 (tạo model mới và thêm Input layer, sub model 1, Output layer)\n","first_model = keras.Sequential()\n","first_model.add(keras.Input(shape=(784,)))\n","first_model.add(first)\n","first_model.add(keras.layers.Dense(10))\n","\n","first_model.compile(optimizer=tf.keras.optimizers.SGD(), \n","              loss=keras.losses.SparseCategoricalCrossentropy(from_logits=True), \n","              metrics=['accuracy'])\n","\n","history = first_model.fit(X_train, y_train, batch_size,\n","                          validation_data=(X_test, y_test),\n","                          epochs=100, verbose=2)"]},{"cell_type":"code","execution_count":4,"id":"8e35d7eb","metadata":{"id":"8e35d7eb","executionInfo":{"status":"ok","timestamp":1669968301982,"user_tz":-420,"elapsed":10,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"}}},"outputs":[],"source":[]},{"cell_type":"code","execution_count":5,"id":"aIXuw9GyK_ux","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":296},"executionInfo":{"elapsed":766,"status":"ok","timestamp":1669968302740,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"},"user_tz":-420},"id":"aIXuw9GyK_ux","outputId":"e43fca9b-f56a-4d0e-f900-48cfbf675bac"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.legend.Legend at 0x7f134b3419d0>"]},"metadata":{},"execution_count":5},{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 1 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}}],"source":["import matplotlib.pyplot as plt\n","\n","plt.plot(history.history['loss'], label='loss')\n","plt.plot(history.history['val_loss'], label='val_loss')\n","plt.xlabel('iteration')\n","plt.ylabel('Loss')\n","plt.legend()"]},{"cell_type":"code","execution_count":5,"id":"TVjfnDdYK_u0","metadata":{"id":"TVjfnDdYK_u0","executionInfo":{"status":"ok","timestamp":1669968302741,"user_tz":-420,"elapsed":19,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"}}},"outputs":[],"source":[]},{"cell_type":"code","execution_count":6,"id":"X4zt706EK_u1","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":296},"executionInfo":{"elapsed":18,"status":"ok","timestamp":1669968302742,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"},"user_tz":-420},"id":"X4zt706EK_u1","outputId":"15918946-54e1-4a2c-ae09-3f27b6d1d126"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.legend.Legend at 0x7f134cb571f0>"]},"metadata":{},"execution_count":6},{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 1 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}}],"source":["import matplotlib.pyplot as plt\n","\n","plt.plot(history.history['accuracy'], label='train_accuracy')\n","plt.plot(history.history['val_accuracy'], label='val_accuracy')\n","plt.xlabel('iteration')\n","plt.ylabel('Accuracy')\n","plt.legend()"]},{"cell_type":"markdown","source":["### Train 2"],"metadata":{"id":"bgoQGrFALDX0"},"id":"bgoQGrFALDX0"},{"cell_type":"code","execution_count":7,"id":"4b997d5b","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":124585,"status":"ok","timestamp":1669968427315,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"},"user_tz":-420},"id":"4b997d5b","outputId":"f078bbcf-491e-4047-ca77-59542f6b2ee7"},"outputs":[{"output_type":"stream","name":"stdout","text":["Epoch 1/100\n","118/118 - 2s - loss: 2.3250 - accuracy: 0.1176 - val_loss: 2.2998 - val_accuracy: 0.1000 - 2s/epoch - 17ms/step\n","Epoch 2/100\n","118/118 - 1s - loss: 2.2984 - accuracy: 0.1311 - val_loss: 2.2971 - val_accuracy: 0.1000 - 1s/epoch - 11ms/step\n","Epoch 3/100\n","118/118 - 1s - loss: 2.2958 - accuracy: 0.1858 - val_loss: 2.2946 - val_accuracy: 0.1000 - 1s/epoch - 11ms/step\n","Epoch 4/100\n","118/118 - 1s - loss: 2.2932 - accuracy: 0.2047 - val_loss: 2.2918 - val_accuracy: 0.1937 - 1s/epoch - 11ms/step\n","Epoch 5/100\n","118/118 - 1s - loss: 2.2904 - accuracy: 0.2561 - val_loss: 2.2890 - val_accuracy: 0.2027 - 1s/epoch - 11ms/step\n","Epoch 6/100\n","118/118 - 1s - loss: 2.2876 - accuracy: 0.3282 - val_loss: 2.2861 - val_accuracy: 0.2071 - 1s/epoch - 10ms/step\n","Epoch 7/100\n","118/118 - 1s - loss: 2.2846 - accuracy: 0.3104 - val_loss: 2.2829 - val_accuracy: 0.4367 - 1s/epoch - 10ms/step\n","Epoch 8/100\n","118/118 - 1s - loss: 2.2814 - accuracy: 0.3823 - val_loss: 2.2796 - val_accuracy: 0.5496 - 1s/epoch - 10ms/step\n","Epoch 9/100\n","118/118 - 1s - loss: 2.2780 - accuracy: 0.3879 - val_loss: 2.2762 - val_accuracy: 0.4124 - 1s/epoch - 10ms/step\n","Epoch 10/100\n","118/118 - 1s - loss: 2.2743 - accuracy: 0.3715 - val_loss: 2.2726 - val_accuracy: 0.2623 - 1s/epoch - 10ms/step\n","Epoch 11/100\n","118/118 - 1s - loss: 2.2704 - accuracy: 0.4344 - val_loss: 2.2685 - val_accuracy: 0.4004 - 1s/epoch - 11ms/step\n","Epoch 12/100\n","118/118 - 1s - loss: 2.2662 - accuracy: 0.4458 - val_loss: 2.2642 - val_accuracy: 0.3662 - 1s/epoch - 10ms/step\n","Epoch 13/100\n","118/118 - 1s - loss: 2.2615 - accuracy: 0.4423 - val_loss: 2.2591 - val_accuracy: 0.5290 - 1s/epoch - 10ms/step\n","Epoch 14/100\n","118/118 - 1s - loss: 2.2564 - accuracy: 0.5546 - val_loss: 2.2537 - val_accuracy: 0.6374 - 1s/epoch - 10ms/step\n","Epoch 15/100\n","118/118 - 1s - loss: 2.2507 - accuracy: 0.5404 - val_loss: 2.2478 - val_accuracy: 0.4944 - 1s/epoch - 10ms/step\n","Epoch 16/100\n","118/118 - 1s - loss: 2.2443 - accuracy: 0.5000 - val_loss: 2.2414 - val_accuracy: 0.4622 - 1s/epoch - 10ms/step\n","Epoch 17/100\n","118/118 - 1s - loss: 2.2373 - accuracy: 0.4912 - val_loss: 2.2338 - val_accuracy: 0.5543 - 1s/epoch - 10ms/step\n","Epoch 18/100\n","118/118 - 1s - loss: 2.2294 - accuracy: 0.5209 - val_loss: 2.2254 - val_accuracy: 0.4115 - 1s/epoch - 10ms/step\n","Epoch 19/100\n","118/118 - 1s - loss: 2.2204 - accuracy: 0.5349 - val_loss: 2.2158 - val_accuracy: 0.5140 - 1s/epoch - 10ms/step\n","Epoch 20/100\n","118/118 - 1s - loss: 2.2103 - accuracy: 0.5610 - val_loss: 2.2049 - val_accuracy: 0.6002 - 1s/epoch - 10ms/step\n","Epoch 21/100\n","118/118 - 1s - loss: 2.1986 - accuracy: 0.5257 - val_loss: 2.1927 - val_accuracy: 0.6421 - 1s/epoch - 10ms/step\n","Epoch 22/100\n","118/118 - 1s - loss: 2.1854 - accuracy: 0.5633 - val_loss: 2.1784 - val_accuracy: 0.5367 - 1s/epoch - 10ms/step\n","Epoch 23/100\n","118/118 - 1s - loss: 2.1702 - accuracy: 0.5138 - val_loss: 2.1625 - val_accuracy: 0.4601 - 1s/epoch - 10ms/step\n","Epoch 24/100\n","118/118 - 1s - loss: 2.1529 - accuracy: 0.5185 - val_loss: 2.1439 - val_accuracy: 0.5549 - 1s/epoch - 10ms/step\n","Epoch 25/100\n","118/118 - 1s - loss: 2.1332 - accuracy: 0.5158 - val_loss: 2.1228 - val_accuracy: 0.5948 - 1s/epoch - 10ms/step\n","Epoch 26/100\n","118/118 - 1s - loss: 2.1107 - accuracy: 0.5163 - val_loss: 2.0990 - val_accuracy: 0.4470 - 1s/epoch - 10ms/step\n","Epoch 27/100\n","118/118 - 1s - loss: 2.0853 - accuracy: 0.5072 - val_loss: 2.0721 - val_accuracy: 0.5263 - 1s/epoch - 10ms/step\n","Epoch 28/100\n","118/118 - 1s - loss: 2.0569 - accuracy: 0.5283 - val_loss: 2.0425 - val_accuracy: 0.5138 - 1s/epoch - 11ms/step\n","Epoch 29/100\n","118/118 - 1s - loss: 2.0258 - accuracy: 0.5367 - val_loss: 2.0099 - val_accuracy: 0.4779 - 1s/epoch - 11ms/step\n","Epoch 30/100\n","118/118 - 1s - loss: 1.9920 - accuracy: 0.5136 - val_loss: 1.9751 - val_accuracy: 0.5628 - 1s/epoch - 11ms/step\n","Epoch 31/100\n","118/118 - 1s - loss: 1.9562 - accuracy: 0.5368 - val_loss: 1.9385 - val_accuracy: 0.5949 - 1s/epoch - 11ms/step\n","Epoch 32/100\n","118/118 - 1s - loss: 1.9191 - accuracy: 0.5639 - val_loss: 1.9013 - val_accuracy: 0.5900 - 1s/epoch - 10ms/step\n","Epoch 33/100\n","118/118 - 1s - loss: 1.8817 - accuracy: 0.5631 - val_loss: 1.8641 - val_accuracy: 0.5871 - 1s/epoch - 11ms/step\n","Epoch 34/100\n","118/118 - 1s - loss: 1.8447 - accuracy: 0.5814 - val_loss: 1.8280 - val_accuracy: 0.6143 - 1s/epoch - 11ms/step\n","Epoch 35/100\n","118/118 - 1s - loss: 1.8091 - accuracy: 0.6154 - val_loss: 1.7933 - val_accuracy: 0.5666 - 1s/epoch - 11ms/step\n","Epoch 36/100\n","118/118 - 1s - loss: 1.7753 - accuracy: 0.6203 - val_loss: 1.7608 - val_accuracy: 0.5979 - 1s/epoch - 11ms/step\n","Epoch 37/100\n","118/118 - 1s - loss: 1.7436 - accuracy: 0.6358 - val_loss: 1.7302 - val_accuracy: 0.6682 - 1s/epoch - 11ms/step\n","Epoch 38/100\n","118/118 - 1s - loss: 1.7140 - accuracy: 0.6481 - val_loss: 1.7019 - val_accuracy: 0.6775 - 1s/epoch - 10ms/step\n","Epoch 39/100\n","118/118 - 1s - loss: 1.6864 - accuracy: 0.6737 - val_loss: 1.6756 - val_accuracy: 0.6900 - 1s/epoch - 10ms/step\n","Epoch 40/100\n","118/118 - 1s - loss: 1.6608 - accuracy: 0.6844 - val_loss: 1.6510 - val_accuracy: 0.7075 - 1s/epoch - 10ms/step\n","Epoch 41/100\n","118/118 - 1s - loss: 1.6367 - accuracy: 0.6908 - val_loss: 1.6280 - val_accuracy: 0.7289 - 1s/epoch - 10ms/step\n","Epoch 42/100\n","118/118 - 1s - loss: 1.6139 - accuracy: 0.7088 - val_loss: 1.6059 - val_accuracy: 0.6624 - 1s/epoch - 10ms/step\n","Epoch 43/100\n","118/118 - 1s - loss: 1.5922 - accuracy: 0.7145 - val_loss: 1.5847 - val_accuracy: 0.6738 - 1s/epoch - 10ms/step\n","Epoch 44/100\n","118/118 - 1s - loss: 1.5712 - accuracy: 0.7120 - val_loss: 1.5643 - val_accuracy: 0.7169 - 1s/epoch - 10ms/step\n","Epoch 45/100\n","118/118 - 1s - loss: 1.5506 - accuracy: 0.7155 - val_loss: 1.5443 - val_accuracy: 0.7270 - 1s/epoch - 10ms/step\n","Epoch 46/100\n","118/118 - 1s - loss: 1.5304 - accuracy: 0.7383 - val_loss: 1.5244 - val_accuracy: 0.7312 - 1s/epoch - 10ms/step\n","Epoch 47/100\n","118/118 - 1s - loss: 1.5103 - accuracy: 0.7387 - val_loss: 1.5046 - val_accuracy: 0.7103 - 1s/epoch - 10ms/step\n","Epoch 48/100\n","118/118 - 1s - loss: 1.4902 - accuracy: 0.7474 - val_loss: 1.4847 - val_accuracy: 0.6945 - 1s/epoch - 10ms/step\n","Epoch 49/100\n","118/118 - 1s - loss: 1.4698 - accuracy: 0.7418 - val_loss: 1.4645 - val_accuracy: 0.7651 - 1s/epoch - 10ms/step\n","Epoch 50/100\n","118/118 - 1s - loss: 1.4490 - accuracy: 0.7589 - val_loss: 1.4439 - val_accuracy: 0.7093 - 1s/epoch - 10ms/step\n","Epoch 51/100\n","118/118 - 1s - loss: 1.4278 - accuracy: 0.7550 - val_loss: 1.4227 - val_accuracy: 0.7453 - 1s/epoch - 11ms/step\n","Epoch 52/100\n","118/118 - 1s - loss: 1.4062 - accuracy: 0.7709 - val_loss: 1.4012 - val_accuracy: 0.7299 - 1s/epoch - 11ms/step\n","Epoch 53/100\n","118/118 - 1s - loss: 1.3840 - accuracy: 0.7609 - val_loss: 1.3793 - val_accuracy: 0.7386 - 1s/epoch - 11ms/step\n","Epoch 54/100\n","118/118 - 1s - loss: 1.3614 - accuracy: 0.7590 - val_loss: 1.3569 - val_accuracy: 0.7498 - 1s/epoch - 10ms/step\n","Epoch 55/100\n","118/118 - 1s - loss: 1.3383 - accuracy: 0.7760 - val_loss: 1.3338 - val_accuracy: 0.7683 - 1s/epoch - 10ms/step\n","Epoch 56/100\n","118/118 - 1s - loss: 1.3147 - accuracy: 0.7815 - val_loss: 1.3106 - val_accuracy: 0.7702 - 1s/epoch - 10ms/step\n","Epoch 57/100\n","118/118 - 1s - loss: 1.2909 - accuracy: 0.7848 - val_loss: 1.2871 - val_accuracy: 0.7790 - 1s/epoch - 10ms/step\n","Epoch 58/100\n","118/118 - 1s - loss: 1.2670 - accuracy: 0.7905 - val_loss: 1.2636 - val_accuracy: 0.7698 - 1s/epoch - 11ms/step\n","Epoch 59/100\n","118/118 - 1s - loss: 1.2430 - accuracy: 0.7945 - val_loss: 1.2403 - val_accuracy: 0.7776 - 1s/epoch - 10ms/step\n","Epoch 60/100\n","118/118 - 1s - loss: 1.2191 - accuracy: 0.7926 - val_loss: 1.2171 - val_accuracy: 0.7856 - 1s/epoch - 11ms/step\n","Epoch 61/100\n","118/118 - 1s - loss: 1.1955 - accuracy: 0.7983 - val_loss: 1.1942 - val_accuracy: 0.7909 - 1s/epoch - 10ms/step\n","Epoch 62/100\n","118/118 - 1s - loss: 1.1722 - accuracy: 0.8032 - val_loss: 1.1718 - val_accuracy: 0.7804 - 1s/epoch - 10ms/step\n","Epoch 63/100\n","118/118 - 1s - loss: 1.1494 - accuracy: 0.8037 - val_loss: 1.1498 - val_accuracy: 0.7811 - 1s/epoch - 11ms/step\n","Epoch 64/100\n","118/118 - 1s - loss: 1.1270 - accuracy: 0.8042 - val_loss: 1.1285 - val_accuracy: 0.7887 - 1s/epoch - 11ms/step\n","Epoch 65/100\n","118/118 - 1s - loss: 1.1053 - accuracy: 0.8075 - val_loss: 1.1077 - val_accuracy: 0.7965 - 1s/epoch - 10ms/step\n","Epoch 66/100\n","118/118 - 1s - loss: 1.0841 - accuracy: 0.8116 - val_loss: 1.0876 - val_accuracy: 0.7794 - 1s/epoch - 10ms/step\n","Epoch 67/100\n","118/118 - 1s - loss: 1.0636 - accuracy: 0.8122 - val_loss: 1.0680 - val_accuracy: 0.7850 - 1s/epoch - 10ms/step\n","Epoch 68/100\n","118/118 - 1s - loss: 1.0436 - accuracy: 0.8122 - val_loss: 1.0490 - val_accuracy: 0.7922 - 1s/epoch - 10ms/step\n","Epoch 69/100\n","118/118 - 1s - loss: 1.0243 - accuracy: 0.8115 - val_loss: 1.0307 - val_accuracy: 0.8014 - 1s/epoch - 10ms/step\n","Epoch 70/100\n","118/118 - 1s - loss: 1.0055 - accuracy: 0.8171 - val_loss: 1.0128 - val_accuracy: 0.7853 - 1s/epoch - 10ms/step\n","Epoch 71/100\n","118/118 - 1s - loss: 0.9874 - accuracy: 0.8156 - val_loss: 0.9955 - val_accuracy: 0.8021 - 1s/epoch - 11ms/step\n","Epoch 72/100\n","118/118 - 1s - loss: 0.9696 - accuracy: 0.8180 - val_loss: 0.9789 - val_accuracy: 0.8018 - 1s/epoch - 11ms/step\n","Epoch 73/100\n","118/118 - 1s - loss: 0.9525 - accuracy: 0.8204 - val_loss: 0.9625 - val_accuracy: 0.8022 - 1s/epoch - 11ms/step\n","Epoch 74/100\n","118/118 - 1s - loss: 0.9357 - accuracy: 0.8217 - val_loss: 0.9469 - val_accuracy: 0.8022 - 1s/epoch - 10ms/step\n","Epoch 75/100\n","118/118 - 1s - loss: 0.9194 - accuracy: 0.8236 - val_loss: 0.9313 - val_accuracy: 0.8059 - 1s/epoch - 11ms/step\n","Epoch 76/100\n","118/118 - 1s - loss: 0.9035 - accuracy: 0.8229 - val_loss: 0.9164 - val_accuracy: 0.8093 - 1s/epoch - 11ms/step\n","Epoch 77/100\n","118/118 - 1s - loss: 0.8881 - accuracy: 0.8256 - val_loss: 0.9018 - val_accuracy: 0.8079 - 1s/epoch - 11ms/step\n","Epoch 78/100\n","118/118 - 1s - loss: 0.8730 - accuracy: 0.8254 - val_loss: 0.8877 - val_accuracy: 0.8088 - 1s/epoch - 10ms/step\n","Epoch 79/100\n","118/118 - 1s - loss: 0.8583 - accuracy: 0.8263 - val_loss: 0.8739 - val_accuracy: 0.8037 - 1s/epoch - 10ms/step\n","Epoch 80/100\n","118/118 - 1s - loss: 0.8439 - accuracy: 0.8258 - val_loss: 0.8604 - val_accuracy: 0.8120 - 1s/epoch - 10ms/step\n","Epoch 81/100\n","118/118 - 1s - loss: 0.8299 - accuracy: 0.8290 - val_loss: 0.8472 - val_accuracy: 0.8101 - 1s/epoch - 10ms/step\n","Epoch 82/100\n","118/118 - 1s - loss: 0.8162 - accuracy: 0.8291 - val_loss: 0.8343 - val_accuracy: 0.8072 - 1s/epoch - 10ms/step\n","Epoch 83/100\n","118/118 - 1s - loss: 0.8028 - accuracy: 0.8288 - val_loss: 0.8219 - val_accuracy: 0.8093 - 1s/epoch - 10ms/step\n","Epoch 84/100\n","118/118 - 1s - loss: 0.7898 - accuracy: 0.8305 - val_loss: 0.8096 - val_accuracy: 0.8127 - 1s/epoch - 10ms/step\n","Epoch 85/100\n","118/118 - 1s - loss: 0.7770 - accuracy: 0.8314 - val_loss: 0.7978 - val_accuracy: 0.8114 - 1s/epoch - 10ms/step\n","Epoch 86/100\n","118/118 - 1s - loss: 0.7646 - accuracy: 0.8315 - val_loss: 0.7862 - val_accuracy: 0.8141 - 1s/epoch - 10ms/step\n","Epoch 87/100\n","118/118 - 1s - loss: 0.7525 - accuracy: 0.8325 - val_loss: 0.7749 - val_accuracy: 0.8117 - 1s/epoch - 11ms/step\n","Epoch 88/100\n","118/118 - 1s - loss: 0.7407 - accuracy: 0.8330 - val_loss: 0.7640 - val_accuracy: 0.8116 - 1s/epoch - 11ms/step\n","Epoch 89/100\n","118/118 - 1s - loss: 0.7293 - accuracy: 0.8328 - val_loss: 0.7535 - val_accuracy: 0.8139 - 1s/epoch - 10ms/step\n","Epoch 90/100\n","118/118 - 1s - loss: 0.7181 - accuracy: 0.8346 - val_loss: 0.7432 - val_accuracy: 0.8104 - 1s/epoch - 10ms/step\n","Epoch 91/100\n","118/118 - 1s - loss: 0.7073 - accuracy: 0.8331 - val_loss: 0.7330 - val_accuracy: 0.8153 - 1s/epoch - 10ms/step\n","Epoch 92/100\n","118/118 - 1s - loss: 0.6968 - accuracy: 0.8353 - val_loss: 0.7233 - val_accuracy: 0.8153 - 1s/epoch - 10ms/step\n","Epoch 93/100\n","118/118 - 1s - loss: 0.6866 - accuracy: 0.8357 - val_loss: 0.7141 - val_accuracy: 0.8150 - 1s/epoch - 10ms/step\n","Epoch 94/100\n","118/118 - 1s - loss: 0.6768 - accuracy: 0.8356 - val_loss: 0.7050 - val_accuracy: 0.8154 - 1s/epoch - 10ms/step\n","Epoch 95/100\n","118/118 - 1s - loss: 0.6672 - accuracy: 0.8368 - val_loss: 0.6963 - val_accuracy: 0.8132 - 1s/epoch - 10ms/step\n","Epoch 96/100\n","118/118 - 1s - loss: 0.6581 - accuracy: 0.8366 - val_loss: 0.6877 - val_accuracy: 0.8173 - 1s/epoch - 10ms/step\n","Epoch 97/100\n","118/118 - 1s - loss: 0.6492 - accuracy: 0.8378 - val_loss: 0.6799 - val_accuracy: 0.8148 - 1s/epoch - 10ms/step\n","Epoch 98/100\n","118/118 - 1s - loss: 0.6406 - accuracy: 0.8384 - val_loss: 0.6720 - val_accuracy: 0.8168 - 1s/epoch - 10ms/step\n","Epoch 99/100\n","118/118 - 1s - loss: 0.6324 - accuracy: 0.8385 - val_loss: 0.6644 - val_accuracy: 0.8179 - 1s/epoch - 10ms/step\n","Epoch 100/100\n","118/118 - 1s - loss: 0.6244 - accuracy: 0.8383 - val_loss: 0.6574 - val_accuracy: 0.8163 - 1s/epoch - 10ms/step\n"]}],"source":["# Fix first and train second\n","#################### YOUR CODE HERE ####################\n","# Các bạn thực hiện tương tự như Ví dụ train lần 1 \n","# Train lần 2: Ghép sub mode 1 và 2 lại với nhau. Train với 100 epoch nhưng freeze (fix) weight của sub model 1 \n","# (tạo model mới và thêm Input layer, sub model 1, sub model 2, Output layer)\n","second_model = keras.Sequential()\n","second_model.add(keras.Input(shape=(784,)))\n","second_model.add(first)\n","second_model.add(second)\n","second_model.add(keras.layers.Dense(10))\n","\n","# freeze sub model 1\n","first.trainable = False\n","\n","########################################################\n","second_model.compile(optimizer=tf.keras.optimizers.SGD(), \n","              loss=keras.losses.SparseCategoricalCrossentropy(from_logits=True), \n","              metrics=['accuracy'])\n","\n","history = second_model.fit(X_train, y_train, batch_size,\n","                           validation_data=(X_test, y_test),\n","                           epochs=100, verbose=2)"]},{"cell_type":"code","execution_count":7,"id":"97uC3rrhLKbY","metadata":{"id":"97uC3rrhLKbY","executionInfo":{"status":"ok","timestamp":1669968427316,"user_tz":-420,"elapsed":29,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"}}},"outputs":[],"source":[]},{"cell_type":"code","execution_count":8,"id":"jgy-YB1XLKuE","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":296},"executionInfo":{"elapsed":28,"status":"ok","timestamp":1669968427316,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"},"user_tz":-420},"id":"jgy-YB1XLKuE","outputId":"bdf15c98-9725-47fb-ae5e-4894482e5a3d"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.legend.Legend at 0x7f134ccd94c0>"]},"metadata":{},"execution_count":8},{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 1 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}}],"source":["import matplotlib.pyplot as plt\n","\n","plt.plot(history.history['loss'], label='loss')\n","plt.plot(history.history['val_loss'], label='val_loss')\n","plt.xlabel('iteration')\n","plt.ylabel('Loss')\n","plt.legend()"]},{"cell_type":"code","execution_count":9,"id":"6E5M8H8-LKuI","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":296},"executionInfo":{"elapsed":712,"status":"ok","timestamp":1669968428007,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"},"user_tz":-420},"id":"6E5M8H8-LKuI","outputId":"1c7572ef-aaa4-40d4-df38-98b8ed0a4cd3"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.legend.Legend at 0x7f134cc4c8e0>"]},"metadata":{},"execution_count":9},{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 1 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}}],"source":["import matplotlib.pyplot as plt\n","\n","plt.plot(history.history['accuracy'], label='train_accuracy')\n","plt.plot(history.history['val_accuracy'], label='val_accuracy')\n","plt.xlabel('iteration')\n","plt.ylabel('Accuracy')\n","plt.legend()"]},{"cell_type":"markdown","source":["### Train 3"],"metadata":{"id":"3f4eb957"},"id":"3f4eb957"},{"cell_type":"code","execution_count":10,"id":"267ca2c6","metadata":{"id":"267ca2c6","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1669968595325,"user_tz":-420,"elapsed":167329,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"}},"outputId":"f0fcedbe-2c3f-4b28-dfbc-70b4d7d8d4a4"},"outputs":[{"output_type":"stream","name":"stdout","text":["Epoch 1/100\n","118/118 - 3s - loss: 2.2054 - accuracy: 0.1975 - val_loss: 1.9769 - val_accuracy: 0.5147 - 3s/epoch - 21ms/step\n","Epoch 2/100\n","118/118 - 2s - loss: 1.8336 - accuracy: 0.6348 - val_loss: 1.7116 - val_accuracy: 0.6655 - 2s/epoch - 14ms/step\n","Epoch 3/100\n","118/118 - 2s - loss: 1.6083 - accuracy: 0.6905 - val_loss: 1.5241 - val_accuracy: 0.6851 - 2s/epoch - 14ms/step\n","Epoch 4/100\n","118/118 - 2s - loss: 1.4449 - accuracy: 0.7122 - val_loss: 1.3846 - val_accuracy: 0.7314 - 2s/epoch - 14ms/step\n","Epoch 5/100\n","118/118 - 3s - loss: 1.3197 - accuracy: 0.7412 - val_loss: 1.2758 - val_accuracy: 0.7421 - 3s/epoch - 25ms/step\n","Epoch 6/100\n","118/118 - 2s - loss: 1.2203 - accuracy: 0.7597 - val_loss: 1.1882 - val_accuracy: 0.7520 - 2s/epoch - 17ms/step\n","Epoch 7/100\n","118/118 - 2s - loss: 1.1385 - accuracy: 0.7724 - val_loss: 1.1151 - val_accuracy: 0.7649 - 2s/epoch - 14ms/step\n","Epoch 8/100\n","118/118 - 2s - loss: 1.0699 - accuracy: 0.7835 - val_loss: 1.0547 - val_accuracy: 0.7740 - 2s/epoch - 14ms/step\n","Epoch 9/100\n","118/118 - 2s - loss: 1.0113 - accuracy: 0.7916 - val_loss: 1.0021 - val_accuracy: 0.7773 - 2s/epoch - 15ms/step\n","Epoch 10/100\n","118/118 - 2s - loss: 0.9609 - accuracy: 0.7964 - val_loss: 0.9567 - val_accuracy: 0.7865 - 2s/epoch - 14ms/step\n","Epoch 11/100\n","118/118 - 2s - loss: 0.9169 - accuracy: 0.8058 - val_loss: 0.9176 - val_accuracy: 0.7888 - 2s/epoch - 14ms/step\n","Epoch 12/100\n","118/118 - 2s - loss: 0.8783 - accuracy: 0.8089 - val_loss: 0.8817 - val_accuracy: 0.7870 - 2s/epoch - 15ms/step\n","Epoch 13/100\n","118/118 - 2s - loss: 0.8438 - accuracy: 0.8098 - val_loss: 0.8509 - val_accuracy: 0.7921 - 2s/epoch - 14ms/step\n","Epoch 14/100\n","118/118 - 2s - loss: 0.8127 - accuracy: 0.8137 - val_loss: 0.8224 - val_accuracy: 0.7969 - 2s/epoch - 14ms/step\n","Epoch 15/100\n","118/118 - 2s - loss: 0.7851 - accuracy: 0.8157 - val_loss: 0.7982 - val_accuracy: 0.7973 - 2s/epoch - 14ms/step\n","Epoch 16/100\n","118/118 - 2s - loss: 0.7599 - accuracy: 0.8185 - val_loss: 0.7769 - val_accuracy: 0.7992 - 2s/epoch - 14ms/step\n","Epoch 17/100\n","118/118 - 2s - loss: 0.7376 - accuracy: 0.8209 - val_loss: 0.7569 - val_accuracy: 0.8038 - 2s/epoch - 14ms/step\n","Epoch 18/100\n","118/118 - 2s - loss: 0.7175 - accuracy: 0.8234 - val_loss: 0.7369 - val_accuracy: 0.8021 - 2s/epoch - 14ms/step\n","Epoch 19/100\n","118/118 - 2s - loss: 0.6982 - accuracy: 0.8256 - val_loss: 0.7204 - val_accuracy: 0.8075 - 2s/epoch - 14ms/step\n","Epoch 20/100\n","118/118 - 2s - loss: 0.6816 - accuracy: 0.8270 - val_loss: 0.7052 - val_accuracy: 0.8073 - 2s/epoch - 14ms/step\n","Epoch 21/100\n","118/118 - 2s - loss: 0.6656 - accuracy: 0.8278 - val_loss: 0.6904 - val_accuracy: 0.8085 - 2s/epoch - 14ms/step\n","Epoch 22/100\n","118/118 - 2s - loss: 0.6510 - accuracy: 0.8289 - val_loss: 0.6793 - val_accuracy: 0.8140 - 2s/epoch - 14ms/step\n","Epoch 23/100\n","118/118 - 2s - loss: 0.6386 - accuracy: 0.8305 - val_loss: 0.6651 - val_accuracy: 0.8130 - 2s/epoch - 14ms/step\n","Epoch 24/100\n","118/118 - 2s - loss: 0.6249 - accuracy: 0.8335 - val_loss: 0.6546 - val_accuracy: 0.8122 - 2s/epoch - 14ms/step\n","Epoch 25/100\n","118/118 - 2s - loss: 0.6142 - accuracy: 0.8331 - val_loss: 0.6434 - val_accuracy: 0.8165 - 2s/epoch - 14ms/step\n","Epoch 26/100\n","118/118 - 2s - loss: 0.6020 - accuracy: 0.8352 - val_loss: 0.6350 - val_accuracy: 0.8140 - 2s/epoch - 14ms/step\n","Epoch 27/100\n","118/118 - 2s - loss: 0.5914 - accuracy: 0.8368 - val_loss: 0.6278 - val_accuracy: 0.8136 - 2s/epoch - 14ms/step\n","Epoch 28/100\n","118/118 - 2s - loss: 0.5825 - accuracy: 0.8379 - val_loss: 0.6204 - val_accuracy: 0.8148 - 2s/epoch - 14ms/step\n","Epoch 29/100\n","118/118 - 2s - loss: 0.5733 - accuracy: 0.8393 - val_loss: 0.6084 - val_accuracy: 0.8182 - 2s/epoch - 14ms/step\n","Epoch 30/100\n","118/118 - 2s - loss: 0.5658 - accuracy: 0.8389 - val_loss: 0.6033 - val_accuracy: 0.8173 - 2s/epoch - 14ms/step\n","Epoch 31/100\n","118/118 - 2s - loss: 0.5572 - accuracy: 0.8406 - val_loss: 0.5960 - val_accuracy: 0.8203 - 2s/epoch - 14ms/step\n","Epoch 32/100\n","118/118 - 2s - loss: 0.5481 - accuracy: 0.8431 - val_loss: 0.5897 - val_accuracy: 0.8210 - 2s/epoch - 14ms/step\n","Epoch 33/100\n","118/118 - 2s - loss: 0.5407 - accuracy: 0.8446 - val_loss: 0.5800 - val_accuracy: 0.8245 - 2s/epoch - 14ms/step\n","Epoch 34/100\n","118/118 - 2s - loss: 0.5336 - accuracy: 0.8454 - val_loss: 0.5811 - val_accuracy: 0.8183 - 2s/epoch - 14ms/step\n","Epoch 35/100\n","118/118 - 2s - loss: 0.5276 - accuracy: 0.8462 - val_loss: 0.5771 - val_accuracy: 0.8232 - 2s/epoch - 14ms/step\n","Epoch 36/100\n","118/118 - 2s - loss: 0.5218 - accuracy: 0.8463 - val_loss: 0.5705 - val_accuracy: 0.8228 - 2s/epoch - 14ms/step\n","Epoch 37/100\n","118/118 - 2s - loss: 0.5160 - accuracy: 0.8487 - val_loss: 0.5618 - val_accuracy: 0.8247 - 2s/epoch - 14ms/step\n","Epoch 38/100\n","118/118 - 2s - loss: 0.5089 - accuracy: 0.8488 - val_loss: 0.5539 - val_accuracy: 0.8282 - 2s/epoch - 14ms/step\n","Epoch 39/100\n","118/118 - 2s - loss: 0.5043 - accuracy: 0.8497 - val_loss: 0.5576 - val_accuracy: 0.8267 - 2s/epoch - 14ms/step\n","Epoch 40/100\n","118/118 - 2s - loss: 0.4989 - accuracy: 0.8499 - val_loss: 0.5473 - val_accuracy: 0.8245 - 2s/epoch - 14ms/step\n","Epoch 41/100\n","118/118 - 2s - loss: 0.4949 - accuracy: 0.8514 - val_loss: 0.5418 - val_accuracy: 0.8274 - 2s/epoch - 14ms/step\n","Epoch 42/100\n","118/118 - 2s - loss: 0.4872 - accuracy: 0.8524 - val_loss: 0.5506 - val_accuracy: 0.8244 - 2s/epoch - 14ms/step\n","Epoch 43/100\n","118/118 - 2s - loss: 0.4850 - accuracy: 0.8533 - val_loss: 0.5359 - val_accuracy: 0.8268 - 2s/epoch - 14ms/step\n","Epoch 44/100\n","118/118 - 2s - loss: 0.4784 - accuracy: 0.8545 - val_loss: 0.5368 - val_accuracy: 0.8283 - 2s/epoch - 14ms/step\n","Epoch 45/100\n","118/118 - 2s - loss: 0.4758 - accuracy: 0.8544 - val_loss: 0.5247 - val_accuracy: 0.8306 - 2s/epoch - 14ms/step\n","Epoch 46/100\n","118/118 - 2s - loss: 0.4719 - accuracy: 0.8546 - val_loss: 0.5280 - val_accuracy: 0.8294 - 2s/epoch - 14ms/step\n","Epoch 47/100\n","118/118 - 2s - loss: 0.4685 - accuracy: 0.8547 - val_loss: 0.5274 - val_accuracy: 0.8280 - 2s/epoch - 14ms/step\n","Epoch 48/100\n","118/118 - 2s - loss: 0.4639 - accuracy: 0.8551 - val_loss: 0.5172 - val_accuracy: 0.8330 - 2s/epoch - 14ms/step\n","Epoch 49/100\n","118/118 - 2s - loss: 0.4598 - accuracy: 0.8557 - val_loss: 0.5192 - val_accuracy: 0.8327 - 2s/epoch - 14ms/step\n","Epoch 50/100\n","118/118 - 2s - loss: 0.4563 - accuracy: 0.8568 - val_loss: 0.5127 - val_accuracy: 0.8335 - 2s/epoch - 14ms/step\n","Epoch 51/100\n","118/118 - 2s - loss: 0.4545 - accuracy: 0.8576 - val_loss: 0.5123 - val_accuracy: 0.8358 - 2s/epoch - 14ms/step\n","Epoch 52/100\n","118/118 - 2s - loss: 0.4501 - accuracy: 0.8578 - val_loss: 0.5099 - val_accuracy: 0.8326 - 2s/epoch - 14ms/step\n","Epoch 53/100\n","118/118 - 2s - loss: 0.4472 - accuracy: 0.8586 - val_loss: 0.5067 - val_accuracy: 0.8339 - 2s/epoch - 14ms/step\n","Epoch 54/100\n","118/118 - 2s - loss: 0.4429 - accuracy: 0.8590 - val_loss: 0.5095 - val_accuracy: 0.8338 - 2s/epoch - 14ms/step\n","Epoch 55/100\n","118/118 - 2s - loss: 0.4394 - accuracy: 0.8601 - val_loss: 0.4993 - val_accuracy: 0.8363 - 2s/epoch - 14ms/step\n","Epoch 56/100\n","118/118 - 2s - loss: 0.4357 - accuracy: 0.8619 - val_loss: 0.5038 - val_accuracy: 0.8330 - 2s/epoch - 14ms/step\n","Epoch 57/100\n","118/118 - 2s - loss: 0.4367 - accuracy: 0.8609 - val_loss: 0.5003 - val_accuracy: 0.8359 - 2s/epoch - 14ms/step\n","Epoch 58/100\n","118/118 - 2s - loss: 0.4365 - accuracy: 0.8600 - val_loss: 0.4961 - val_accuracy: 0.8369 - 2s/epoch - 14ms/step\n","Epoch 59/100\n","118/118 - 2s - loss: 0.4306 - accuracy: 0.8620 - val_loss: 0.4897 - val_accuracy: 0.8371 - 2s/epoch - 14ms/step\n","Epoch 60/100\n","118/118 - 2s - loss: 0.4303 - accuracy: 0.8611 - val_loss: 0.4908 - val_accuracy: 0.8385 - 2s/epoch - 14ms/step\n","Epoch 61/100\n","118/118 - 2s - loss: 0.4241 - accuracy: 0.8631 - val_loss: 0.4949 - val_accuracy: 0.8332 - 2s/epoch - 14ms/step\n","Epoch 62/100\n","118/118 - 2s - loss: 0.4231 - accuracy: 0.8636 - val_loss: 0.4949 - val_accuracy: 0.8370 - 2s/epoch - 14ms/step\n","Epoch 63/100\n","118/118 - 2s - loss: 0.4216 - accuracy: 0.8638 - val_loss: 0.4837 - val_accuracy: 0.8375 - 2s/epoch - 14ms/step\n","Epoch 64/100\n","118/118 - 2s - loss: 0.4176 - accuracy: 0.8642 - val_loss: 0.4916 - val_accuracy: 0.8361 - 2s/epoch - 14ms/step\n","Epoch 65/100\n","118/118 - 2s - loss: 0.4169 - accuracy: 0.8644 - val_loss: 0.4848 - val_accuracy: 0.8348 - 2s/epoch - 14ms/step\n","Epoch 66/100\n","118/118 - 2s - loss: 0.4152 - accuracy: 0.8664 - val_loss: 0.4821 - val_accuracy: 0.8397 - 2s/epoch - 14ms/step\n","Epoch 67/100\n","118/118 - 2s - loss: 0.4096 - accuracy: 0.8675 - val_loss: 0.4805 - val_accuracy: 0.8380 - 2s/epoch - 14ms/step\n","Epoch 68/100\n","118/118 - 2s - loss: 0.4093 - accuracy: 0.8669 - val_loss: 0.4801 - val_accuracy: 0.8428 - 2s/epoch - 14ms/step\n","Epoch 69/100\n","118/118 - 2s - loss: 0.4100 - accuracy: 0.8670 - val_loss: 0.4777 - val_accuracy: 0.8392 - 2s/epoch - 14ms/step\n","Epoch 70/100\n","118/118 - 2s - loss: 0.4071 - accuracy: 0.8672 - val_loss: 0.4706 - val_accuracy: 0.8435 - 2s/epoch - 14ms/step\n","Epoch 71/100\n","118/118 - 2s - loss: 0.4046 - accuracy: 0.8682 - val_loss: 0.4733 - val_accuracy: 0.8439 - 2s/epoch - 14ms/step\n","Epoch 72/100\n","118/118 - 2s - loss: 0.4078 - accuracy: 0.8676 - val_loss: 0.4871 - val_accuracy: 0.8342 - 2s/epoch - 14ms/step\n","Epoch 73/100\n","118/118 - 2s - loss: 0.4040 - accuracy: 0.8683 - val_loss: 0.4751 - val_accuracy: 0.8373 - 2s/epoch - 14ms/step\n","Epoch 74/100\n","118/118 - 2s - loss: 0.4020 - accuracy: 0.8687 - val_loss: 0.4765 - val_accuracy: 0.8377 - 2s/epoch - 14ms/step\n","Epoch 75/100\n","118/118 - 2s - loss: 0.4014 - accuracy: 0.8685 - val_loss: 0.4808 - val_accuracy: 0.8375 - 2s/epoch - 14ms/step\n","Epoch 76/100\n","118/118 - 2s - loss: 0.3990 - accuracy: 0.8697 - val_loss: 0.4704 - val_accuracy: 0.8422 - 2s/epoch - 14ms/step\n","Epoch 77/100\n","118/118 - 2s - loss: 0.3950 - accuracy: 0.8703 - val_loss: 0.4735 - val_accuracy: 0.8385 - 2s/epoch - 14ms/step\n","Epoch 78/100\n","118/118 - 2s - loss: 0.3973 - accuracy: 0.8694 - val_loss: 0.4671 - val_accuracy: 0.8423 - 2s/epoch - 14ms/step\n","Epoch 79/100\n","118/118 - 2s - loss: 0.3938 - accuracy: 0.8708 - val_loss: 0.4695 - val_accuracy: 0.8416 - 2s/epoch - 14ms/step\n","Epoch 80/100\n","118/118 - 2s - loss: 0.3938 - accuracy: 0.8703 - val_loss: 0.4751 - val_accuracy: 0.8365 - 2s/epoch - 14ms/step\n","Epoch 81/100\n","118/118 - 2s - loss: 0.3919 - accuracy: 0.8704 - val_loss: 0.4646 - val_accuracy: 0.8430 - 2s/epoch - 14ms/step\n","Epoch 82/100\n","118/118 - 2s - loss: 0.3909 - accuracy: 0.8713 - val_loss: 0.4652 - val_accuracy: 0.8426 - 2s/epoch - 14ms/step\n","Epoch 83/100\n","118/118 - 2s - loss: 0.3898 - accuracy: 0.8713 - val_loss: 0.4675 - val_accuracy: 0.8427 - 2s/epoch - 14ms/step\n","Epoch 84/100\n","118/118 - 2s - loss: 0.3895 - accuracy: 0.8705 - val_loss: 0.4659 - val_accuracy: 0.8400 - 2s/epoch - 14ms/step\n","Epoch 85/100\n","118/118 - 2s - loss: 0.3863 - accuracy: 0.8723 - val_loss: 0.4632 - val_accuracy: 0.8421 - 2s/epoch - 14ms/step\n","Epoch 86/100\n","118/118 - 2s - loss: 0.3833 - accuracy: 0.8735 - val_loss: 0.4644 - val_accuracy: 0.8433 - 2s/epoch - 14ms/step\n","Epoch 87/100\n","118/118 - 2s - loss: 0.3849 - accuracy: 0.8724 - val_loss: 0.4679 - val_accuracy: 0.8396 - 2s/epoch - 14ms/step\n","Epoch 88/100\n","118/118 - 2s - loss: 0.3829 - accuracy: 0.8732 - val_loss: 0.4663 - val_accuracy: 0.8392 - 2s/epoch - 14ms/step\n","Epoch 89/100\n","118/118 - 2s - loss: 0.3809 - accuracy: 0.8738 - val_loss: 0.4607 - val_accuracy: 0.8414 - 2s/epoch - 14ms/step\n","Epoch 90/100\n","118/118 - 2s - loss: 0.3794 - accuracy: 0.8737 - val_loss: 0.4556 - val_accuracy: 0.8461 - 2s/epoch - 14ms/step\n","Epoch 91/100\n","118/118 - 2s - loss: 0.3782 - accuracy: 0.8748 - val_loss: 0.4676 - val_accuracy: 0.8384 - 2s/epoch - 13ms/step\n","Epoch 92/100\n","118/118 - 2s - loss: 0.3788 - accuracy: 0.8745 - val_loss: 0.4571 - val_accuracy: 0.8456 - 2s/epoch - 14ms/step\n","Epoch 93/100\n","118/118 - 2s - loss: 0.3755 - accuracy: 0.8751 - val_loss: 0.4657 - val_accuracy: 0.8430 - 2s/epoch - 14ms/step\n","Epoch 94/100\n","118/118 - 2s - loss: 0.3756 - accuracy: 0.8749 - val_loss: 0.4733 - val_accuracy: 0.8357 - 2s/epoch - 14ms/step\n","Epoch 95/100\n","118/118 - 2s - loss: 0.3750 - accuracy: 0.8747 - val_loss: 0.4524 - val_accuracy: 0.8479 - 2s/epoch - 14ms/step\n","Epoch 96/100\n","118/118 - 2s - loss: 0.3689 - accuracy: 0.8763 - val_loss: 0.4558 - val_accuracy: 0.8458 - 2s/epoch - 14ms/step\n","Epoch 97/100\n","118/118 - 2s - loss: 0.3747 - accuracy: 0.8743 - val_loss: 0.4521 - val_accuracy: 0.8459 - 2s/epoch - 14ms/step\n","Epoch 98/100\n","118/118 - 2s - loss: 0.3718 - accuracy: 0.8752 - val_loss: 0.4590 - val_accuracy: 0.8396 - 2s/epoch - 15ms/step\n","Epoch 99/100\n","118/118 - 2s - loss: 0.3738 - accuracy: 0.8749 - val_loss: 0.4558 - val_accuracy: 0.8444 - 2s/epoch - 14ms/step\n","Epoch 100/100\n","118/118 - 2s - loss: 0.3701 - accuracy: 0.8755 - val_loss: 0.4670 - val_accuracy: 0.8371 - 2s/epoch - 14ms/step\n"]}],"source":["# train first and second\n","#################### YOUR CODE HERE ####################\n","# Các bạn thực hiện tương tự như Ví dụ train lần 1 \n","# Train lần 3: Tiếp tục train thêm 100 epoch nhưng lần này không fix weight của sub model 1\n","# (tạo model mới và thêm Input layer, sub model 1, sub model 2, Output layer)\n","second_model = keras.Sequential()\n","second_model.add(keras.Input(shape=(784,)))\n","second_model.add(first)\n","second_model.add(second)\n","second_model.add(keras.layers.Dense(10))\n","\n","# unfreeze sub model 1\n","first.trainable = True\n","\n","########################################################\n","\n","second_model.compile(optimizer=tf.keras.optimizers.SGD(), \n","              loss=keras.losses.SparseCategoricalCrossentropy(from_logits=True), \n","              metrics=['accuracy'])\n","\n","history = second_model.fit(X_train, y_train, batch_size,\n","                          validation_data=(X_test, y_test),\n","                          epochs=100, verbose=2)"]},{"cell_type":"code","execution_count":10,"id":"Eca1wNk9LOIt","metadata":{"id":"Eca1wNk9LOIt","executionInfo":{"status":"ok","timestamp":1669968595325,"user_tz":-420,"elapsed":28,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"}}},"outputs":[],"source":[]},{"cell_type":"code","execution_count":11,"id":"2twyTdgsLOXa","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":296},"executionInfo":{"elapsed":28,"status":"ok","timestamp":1669968595326,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"},"user_tz":-420},"id":"2twyTdgsLOXa","outputId":"67182d75-f84b-4c02-bdf9-a96ebfb93e76"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.legend.Legend at 0x7f134b9dc040>"]},"metadata":{},"execution_count":11},{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 1 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}}],"source":["import matplotlib.pyplot as plt\n","\n","plt.plot(history.history['loss'], label='loss')\n","plt.plot(history.history['val_loss'], label='val_loss')\n","plt.xlabel('iteration')\n","plt.ylabel('Loss')\n","plt.legend()"]},{"cell_type":"code","execution_count":11,"id":"BWrLQlBrLOXe","metadata":{"id":"BWrLQlBrLOXe","executionInfo":{"status":"ok","timestamp":1669968595326,"user_tz":-420,"elapsed":21,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"}}},"outputs":[],"source":[]},{"cell_type":"code","execution_count":12,"id":"VwBoIbt7LOXg","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":297},"executionInfo":{"elapsed":551,"status":"ok","timestamp":1669968595856,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"},"user_tz":-420},"id":"VwBoIbt7LOXg","outputId":"13ea60b8-6cd0-494a-ce71-22349ad97efc"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.legend.Legend at 0x7f134c2d8430>"]},"metadata":{},"execution_count":12},{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 1 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}}],"source":["import matplotlib.pyplot as plt\n","\n","plt.plot(history.history['accuracy'], label='train_accuracy')\n","plt.plot(history.history['val_accuracy'], label='val_accuracy')\n","plt.xlabel('iteration')\n","plt.ylabel('Accuracy')\n","plt.legend()"]},{"cell_type":"markdown","source":["### Train 4"],"metadata":{"id":"01243d95"},"id":"01243d95"},{"cell_type":"code","execution_count":13,"id":"8e6d4c87","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":142656,"status":"ok","timestamp":1669968738506,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"},"user_tz":-420},"id":"8e6d4c87","outputId":"1f828f70-9645-432b-cb8b-fa0605ca86f4"},"outputs":[{"output_type":"stream","name":"stdout","text":["Epoch 1/100\n","118/118 - 2s - loss: 2.3085 - accuracy: 0.1172 - val_loss: 2.2981 - val_accuracy: 0.1000 - 2s/epoch - 18ms/step\n","Epoch 2/100\n","118/118 - 1s - loss: 2.2964 - accuracy: 0.1682 - val_loss: 2.2948 - val_accuracy: 0.1000 - 1s/epoch - 12ms/step\n","Epoch 3/100\n","118/118 - 1s - loss: 2.2931 - accuracy: 0.2262 - val_loss: 2.2917 - val_accuracy: 0.1000 - 1s/epoch - 13ms/step\n","Epoch 4/100\n","118/118 - 1s - loss: 2.2896 - accuracy: 0.2299 - val_loss: 2.2881 - val_accuracy: 0.1722 - 1s/epoch - 12ms/step\n","Epoch 5/100\n","118/118 - 1s - loss: 2.2860 - accuracy: 0.2967 - val_loss: 2.2844 - val_accuracy: 0.2393 - 1s/epoch - 12ms/step\n","Epoch 6/100\n","118/118 - 1s - loss: 2.2822 - accuracy: 0.4152 - val_loss: 2.2805 - val_accuracy: 0.2976 - 1s/epoch - 12ms/step\n","Epoch 7/100\n","118/118 - 1s - loss: 2.2781 - accuracy: 0.4138 - val_loss: 2.2762 - val_accuracy: 0.4530 - 1s/epoch - 12ms/step\n","Epoch 8/100\n","118/118 - 1s - loss: 2.2737 - accuracy: 0.4545 - val_loss: 2.2716 - val_accuracy: 0.5338 - 1s/epoch - 12ms/step\n","Epoch 9/100\n","118/118 - 1s - loss: 2.2689 - accuracy: 0.4712 - val_loss: 2.2667 - val_accuracy: 0.5476 - 1s/epoch - 12ms/step\n","Epoch 10/100\n","118/118 - 1s - loss: 2.2636 - accuracy: 0.4834 - val_loss: 2.2614 - val_accuracy: 0.5395 - 1s/epoch - 12ms/step\n","Epoch 11/100\n","118/118 - 1s - loss: 2.2578 - accuracy: 0.5200 - val_loss: 2.2552 - val_accuracy: 0.6751 - 1s/epoch - 12ms/step\n","Epoch 12/100\n","118/118 - 1s - loss: 2.2512 - accuracy: 0.5730 - val_loss: 2.2485 - val_accuracy: 0.5288 - 1s/epoch - 11ms/step\n","Epoch 13/100\n","118/118 - 1s - loss: 2.2439 - accuracy: 0.5792 - val_loss: 2.2406 - val_accuracy: 0.6321 - 1s/epoch - 12ms/step\n","Epoch 14/100\n","118/118 - 1s - loss: 2.2356 - accuracy: 0.6384 - val_loss: 2.2319 - val_accuracy: 0.6142 - 1s/epoch - 12ms/step\n","Epoch 15/100\n","118/118 - 1s - loss: 2.2262 - accuracy: 0.6178 - val_loss: 2.2220 - val_accuracy: 0.6091 - 1s/epoch - 11ms/step\n","Epoch 16/100\n","118/118 - 1s - loss: 2.2154 - accuracy: 0.6007 - val_loss: 2.2108 - val_accuracy: 0.5502 - 1s/epoch - 12ms/step\n","Epoch 17/100\n","118/118 - 1s - loss: 2.2032 - accuracy: 0.6027 - val_loss: 2.1975 - val_accuracy: 0.5423 - 1s/epoch - 12ms/step\n","Epoch 18/100\n","118/118 - 1s - loss: 2.1889 - accuracy: 0.6200 - val_loss: 2.1825 - val_accuracy: 0.6007 - 1s/epoch - 12ms/step\n","Epoch 19/100\n","118/118 - 1s - loss: 2.1726 - accuracy: 0.6492 - val_loss: 2.1649 - val_accuracy: 0.6712 - 1s/epoch - 11ms/step\n","Epoch 20/100\n","118/118 - 1s - loss: 2.1536 - accuracy: 0.6406 - val_loss: 2.1447 - val_accuracy: 0.6777 - 1s/epoch - 12ms/step\n","Epoch 21/100\n","118/118 - 1s - loss: 2.1316 - accuracy: 0.6211 - val_loss: 2.1213 - val_accuracy: 0.6955 - 1s/epoch - 12ms/step\n","Epoch 22/100\n","118/118 - 1s - loss: 2.1061 - accuracy: 0.6602 - val_loss: 2.0939 - val_accuracy: 0.5541 - 1s/epoch - 11ms/step\n","Epoch 23/100\n","118/118 - 1s - loss: 2.0765 - accuracy: 0.6327 - val_loss: 2.0627 - val_accuracy: 0.5694 - 1s/epoch - 12ms/step\n","Epoch 24/100\n","118/118 - 1s - loss: 2.0426 - accuracy: 0.5983 - val_loss: 2.0266 - val_accuracy: 0.5906 - 1s/epoch - 11ms/step\n","Epoch 25/100\n","118/118 - 1s - loss: 2.0040 - accuracy: 0.6224 - val_loss: 1.9857 - val_accuracy: 0.6116 - 1s/epoch - 12ms/step\n","Epoch 26/100\n","118/118 - 1s - loss: 1.9603 - accuracy: 0.6123 - val_loss: 1.9402 - val_accuracy: 0.6506 - 1s/epoch - 12ms/step\n","Epoch 27/100\n","118/118 - 1s - loss: 1.9120 - accuracy: 0.6345 - val_loss: 1.8899 - val_accuracy: 0.6005 - 1s/epoch - 13ms/step\n","Epoch 28/100\n","118/118 - 2s - loss: 1.8593 - accuracy: 0.6222 - val_loss: 1.8362 - val_accuracy: 0.6343 - 2s/epoch - 13ms/step\n","Epoch 29/100\n","118/118 - 1s - loss: 1.8033 - accuracy: 0.6454 - val_loss: 1.7793 - val_accuracy: 0.6299 - 1s/epoch - 12ms/step\n","Epoch 30/100\n","118/118 - 1s - loss: 1.7448 - accuracy: 0.6546 - val_loss: 1.7207 - val_accuracy: 0.5953 - 1s/epoch - 12ms/step\n","Epoch 31/100\n","118/118 - 1s - loss: 1.6849 - accuracy: 0.6504 - val_loss: 1.6614 - val_accuracy: 0.6687 - 1s/epoch - 12ms/step\n","Epoch 32/100\n","118/118 - 1s - loss: 1.6249 - accuracy: 0.6845 - val_loss: 1.6025 - val_accuracy: 0.6943 - 1s/epoch - 11ms/step\n","Epoch 33/100\n","118/118 - 1s - loss: 1.5657 - accuracy: 0.6978 - val_loss: 1.5449 - val_accuracy: 0.7036 - 1s/epoch - 12ms/step\n","Epoch 34/100\n","118/118 - 1s - loss: 1.5077 - accuracy: 0.7167 - val_loss: 1.4892 - val_accuracy: 0.7240 - 1s/epoch - 12ms/step\n","Epoch 35/100\n","118/118 - 1s - loss: 1.4519 - accuracy: 0.7355 - val_loss: 1.4355 - val_accuracy: 0.7193 - 1s/epoch - 11ms/step\n","Epoch 36/100\n","118/118 - 1s - loss: 1.3984 - accuracy: 0.7448 - val_loss: 1.3846 - val_accuracy: 0.7370 - 1s/epoch - 11ms/step\n","Epoch 37/100\n","118/118 - 1s - loss: 1.3476 - accuracy: 0.7602 - val_loss: 1.3362 - val_accuracy: 0.7489 - 1s/epoch - 12ms/step\n","Epoch 38/100\n","118/118 - 1s - loss: 1.2993 - accuracy: 0.7717 - val_loss: 1.2907 - val_accuracy: 0.7575 - 1s/epoch - 13ms/step\n","Epoch 39/100\n","118/118 - 1s - loss: 1.2538 - accuracy: 0.7825 - val_loss: 1.2480 - val_accuracy: 0.7550 - 1s/epoch - 12ms/step\n","Epoch 40/100\n","118/118 - 1s - loss: 1.2111 - accuracy: 0.7878 - val_loss: 1.2079 - val_accuracy: 0.7625 - 1s/epoch - 11ms/step\n","Epoch 41/100\n","118/118 - 1s - loss: 1.1708 - accuracy: 0.7927 - val_loss: 1.1703 - val_accuracy: 0.7702 - 1s/epoch - 11ms/step\n","Epoch 42/100\n","118/118 - 1s - loss: 1.1330 - accuracy: 0.7989 - val_loss: 1.1350 - val_accuracy: 0.7814 - 1s/epoch - 11ms/step\n","Epoch 43/100\n","118/118 - 1s - loss: 1.0973 - accuracy: 0.8056 - val_loss: 1.1018 - val_accuracy: 0.7776 - 1s/epoch - 11ms/step\n","Epoch 44/100\n","118/118 - 1s - loss: 1.0638 - accuracy: 0.8084 - val_loss: 1.0706 - val_accuracy: 0.7792 - 1s/epoch - 11ms/step\n","Epoch 45/100\n","118/118 - 1s - loss: 1.0320 - accuracy: 0.8110 - val_loss: 1.0412 - val_accuracy: 0.7795 - 1s/epoch - 11ms/step\n","Epoch 46/100\n","118/118 - 1s - loss: 1.0020 - accuracy: 0.8123 - val_loss: 1.0134 - val_accuracy: 0.7909 - 1s/epoch - 11ms/step\n","Epoch 47/100\n","118/118 - 1s - loss: 0.9736 - accuracy: 0.8179 - val_loss: 0.9872 - val_accuracy: 0.7855 - 1s/epoch - 11ms/step\n","Epoch 48/100\n","118/118 - 1s - loss: 0.9466 - accuracy: 0.8184 - val_loss: 0.9622 - val_accuracy: 0.7849 - 1s/epoch - 11ms/step\n","Epoch 49/100\n","118/118 - 1s - loss: 0.9209 - accuracy: 0.8191 - val_loss: 0.9386 - val_accuracy: 0.7951 - 1s/epoch - 12ms/step\n","Epoch 50/100\n","118/118 - 1s - loss: 0.8964 - accuracy: 0.8217 - val_loss: 0.9161 - val_accuracy: 0.8051 - 1s/epoch - 12ms/step\n","Epoch 51/100\n","118/118 - 1s - loss: 0.8730 - accuracy: 0.8261 - val_loss: 0.8947 - val_accuracy: 0.8031 - 1s/epoch - 12ms/step\n","Epoch 52/100\n","118/118 - 1s - loss: 0.8508 - accuracy: 0.8277 - val_loss: 0.8743 - val_accuracy: 0.8024 - 1s/epoch - 11ms/step\n","Epoch 53/100\n","118/118 - 1s - loss: 0.8295 - accuracy: 0.8303 - val_loss: 0.8550 - val_accuracy: 0.8003 - 1s/epoch - 12ms/step\n","Epoch 54/100\n","118/118 - 1s - loss: 0.8092 - accuracy: 0.8310 - val_loss: 0.8365 - val_accuracy: 0.8066 - 1s/epoch - 12ms/step\n","Epoch 55/100\n","118/118 - 1s - loss: 0.7898 - accuracy: 0.8341 - val_loss: 0.8187 - val_accuracy: 0.8042 - 1s/epoch - 11ms/step\n","Epoch 56/100\n","118/118 - 1s - loss: 0.7712 - accuracy: 0.8352 - val_loss: 0.8018 - val_accuracy: 0.8098 - 1s/epoch - 12ms/step\n","Epoch 57/100\n","118/118 - 1s - loss: 0.7534 - accuracy: 0.8379 - val_loss: 0.7858 - val_accuracy: 0.8091 - 1s/epoch - 11ms/step\n","Epoch 58/100\n","118/118 - 1s - loss: 0.7365 - accuracy: 0.8386 - val_loss: 0.7705 - val_accuracy: 0.8158 - 1s/epoch - 11ms/step\n","Epoch 59/100\n","118/118 - 1s - loss: 0.7203 - accuracy: 0.8417 - val_loss: 0.7560 - val_accuracy: 0.8104 - 1s/epoch - 12ms/step\n","Epoch 60/100\n","118/118 - 1s - loss: 0.7049 - accuracy: 0.8413 - val_loss: 0.7421 - val_accuracy: 0.8159 - 1s/epoch - 12ms/step\n","Epoch 61/100\n","118/118 - 1s - loss: 0.6901 - accuracy: 0.8448 - val_loss: 0.7289 - val_accuracy: 0.8137 - 1s/epoch - 12ms/step\n","Epoch 62/100\n","118/118 - 1s - loss: 0.6761 - accuracy: 0.8451 - val_loss: 0.7164 - val_accuracy: 0.8199 - 1s/epoch - 12ms/step\n","Epoch 63/100\n","118/118 - 1s - loss: 0.6627 - accuracy: 0.8479 - val_loss: 0.7045 - val_accuracy: 0.8193 - 1s/epoch - 11ms/step\n","Epoch 64/100\n","118/118 - 1s - loss: 0.6499 - accuracy: 0.8484 - val_loss: 0.6932 - val_accuracy: 0.8203 - 1s/epoch - 11ms/step\n","Epoch 65/100\n","118/118 - 1s - loss: 0.6378 - accuracy: 0.8496 - val_loss: 0.6826 - val_accuracy: 0.8190 - 1s/epoch - 12ms/step\n","Epoch 66/100\n","118/118 - 1s - loss: 0.6263 - accuracy: 0.8504 - val_loss: 0.6725 - val_accuracy: 0.8231 - 1s/epoch - 12ms/step\n","Epoch 67/100\n","118/118 - 1s - loss: 0.6154 - accuracy: 0.8524 - val_loss: 0.6629 - val_accuracy: 0.8218 - 1s/epoch - 11ms/step\n","Epoch 68/100\n","118/118 - 1s - loss: 0.6050 - accuracy: 0.8532 - val_loss: 0.6538 - val_accuracy: 0.8218 - 1s/epoch - 12ms/step\n","Epoch 69/100\n","118/118 - 1s - loss: 0.5952 - accuracy: 0.8547 - val_loss: 0.6455 - val_accuracy: 0.8205 - 1s/epoch - 11ms/step\n","Epoch 70/100\n","118/118 - 1s - loss: 0.5859 - accuracy: 0.8534 - val_loss: 0.6373 - val_accuracy: 0.8262 - 1s/epoch - 11ms/step\n","Epoch 71/100\n","118/118 - 1s - loss: 0.5771 - accuracy: 0.8560 - val_loss: 0.6297 - val_accuracy: 0.8243 - 1s/epoch - 11ms/step\n","Epoch 72/100\n","118/118 - 1s - loss: 0.5687 - accuracy: 0.8569 - val_loss: 0.6226 - val_accuracy: 0.8238 - 1s/epoch - 12ms/step\n","Epoch 73/100\n","118/118 - 1s - loss: 0.5609 - accuracy: 0.8569 - val_loss: 0.6158 - val_accuracy: 0.8256 - 1s/epoch - 12ms/step\n","Epoch 74/100\n","118/118 - 1s - loss: 0.5533 - accuracy: 0.8582 - val_loss: 0.6095 - val_accuracy: 0.8258 - 1s/epoch - 13ms/step\n","Epoch 75/100\n","118/118 - 1s - loss: 0.5462 - accuracy: 0.8583 - val_loss: 0.6034 - val_accuracy: 0.8265 - 1s/epoch - 13ms/step\n","Epoch 76/100\n","118/118 - 1s - loss: 0.5395 - accuracy: 0.8597 - val_loss: 0.5978 - val_accuracy: 0.8257 - 1s/epoch - 12ms/step\n","Epoch 77/100\n","118/118 - 1s - loss: 0.5331 - accuracy: 0.8596 - val_loss: 0.5923 - val_accuracy: 0.8283 - 1s/epoch - 12ms/step\n","Epoch 78/100\n","118/118 - 1s - loss: 0.5271 - accuracy: 0.8604 - val_loss: 0.5874 - val_accuracy: 0.8277 - 1s/epoch - 12ms/step\n","Epoch 79/100\n","118/118 - 1s - loss: 0.5213 - accuracy: 0.8607 - val_loss: 0.5827 - val_accuracy: 0.8303 - 1s/epoch - 12ms/step\n","Epoch 80/100\n","118/118 - 1s - loss: 0.5159 - accuracy: 0.8613 - val_loss: 0.5781 - val_accuracy: 0.8287 - 1s/epoch - 12ms/step\n","Epoch 81/100\n","118/118 - 1s - loss: 0.5107 - accuracy: 0.8614 - val_loss: 0.5738 - val_accuracy: 0.8300 - 1s/epoch - 12ms/step\n","Epoch 82/100\n","118/118 - 1s - loss: 0.5058 - accuracy: 0.8619 - val_loss: 0.5697 - val_accuracy: 0.8305 - 1s/epoch - 11ms/step\n","Epoch 83/100\n","118/118 - 1s - loss: 0.5011 - accuracy: 0.8623 - val_loss: 0.5660 - val_accuracy: 0.8302 - 1s/epoch - 11ms/step\n","Epoch 84/100\n","118/118 - 1s - loss: 0.4967 - accuracy: 0.8627 - val_loss: 0.5624 - val_accuracy: 0.8309 - 1s/epoch - 11ms/step\n","Epoch 85/100\n","118/118 - 1s - loss: 0.4925 - accuracy: 0.8630 - val_loss: 0.5590 - val_accuracy: 0.8319 - 1s/epoch - 11ms/step\n","Epoch 86/100\n","118/118 - 1s - loss: 0.4884 - accuracy: 0.8637 - val_loss: 0.5558 - val_accuracy: 0.8302 - 1s/epoch - 11ms/step\n","Epoch 87/100\n","118/118 - 1s - loss: 0.4846 - accuracy: 0.8637 - val_loss: 0.5526 - val_accuracy: 0.8320 - 1s/epoch - 11ms/step\n","Epoch 88/100\n","118/118 - 1s - loss: 0.4810 - accuracy: 0.8640 - val_loss: 0.5498 - val_accuracy: 0.8331 - 1s/epoch - 12ms/step\n","Epoch 89/100\n","118/118 - 1s - loss: 0.4775 - accuracy: 0.8644 - val_loss: 0.5471 - val_accuracy: 0.8311 - 1s/epoch - 12ms/step\n","Epoch 90/100\n","118/118 - 1s - loss: 0.4742 - accuracy: 0.8642 - val_loss: 0.5445 - val_accuracy: 0.8334 - 1s/epoch - 12ms/step\n","Epoch 91/100\n","118/118 - 1s - loss: 0.4710 - accuracy: 0.8647 - val_loss: 0.5419 - val_accuracy: 0.8326 - 1s/epoch - 12ms/step\n","Epoch 92/100\n","118/118 - 1s - loss: 0.4680 - accuracy: 0.8647 - val_loss: 0.5395 - val_accuracy: 0.8338 - 1s/epoch - 12ms/step\n","Epoch 93/100\n","118/118 - 1s - loss: 0.4651 - accuracy: 0.8652 - val_loss: 0.5373 - val_accuracy: 0.8329 - 1s/epoch - 12ms/step\n","Epoch 94/100\n","118/118 - 1s - loss: 0.4624 - accuracy: 0.8651 - val_loss: 0.5352 - val_accuracy: 0.8328 - 1s/epoch - 12ms/step\n","Epoch 95/100\n","118/118 - 1s - loss: 0.4597 - accuracy: 0.8650 - val_loss: 0.5331 - val_accuracy: 0.8340 - 1s/epoch - 11ms/step\n","Epoch 96/100\n","118/118 - 1s - loss: 0.4572 - accuracy: 0.8655 - val_loss: 0.5311 - val_accuracy: 0.8339 - 1s/epoch - 11ms/step\n","Epoch 97/100\n","118/118 - 1s - loss: 0.4548 - accuracy: 0.8656 - val_loss: 0.5294 - val_accuracy: 0.8324 - 1s/epoch - 12ms/step\n","Epoch 98/100\n","118/118 - 1s - loss: 0.4525 - accuracy: 0.8655 - val_loss: 0.5277 - val_accuracy: 0.8339 - 1s/epoch - 12ms/step\n","Epoch 99/100\n","118/118 - 1s - loss: 0.4503 - accuracy: 0.8658 - val_loss: 0.5259 - val_accuracy: 0.8342 - 1s/epoch - 11ms/step\n","Epoch 100/100\n","118/118 - 1s - loss: 0.4482 - accuracy: 0.8659 - val_loss: 0.5245 - val_accuracy: 0.8342 - 1s/epoch - 12ms/step\n"]}],"source":["# fix first and second, train third\n","#################### YOUR CODE HERE ####################\n","# Các bạn thực hiện tương tự như Ví dụ train lần 1 \n","# Train lần 4: Ghép sub mode 1, 2 và 3 lại với nhau. Train với 100 epoch nhưng fix weight của sub model 1 và 2\n","# (tạo model mới và thêm Input layer, sub model 1, sub model 2, sub model 3, Output layer)\n","third_model = keras.Sequential()\n","third_model.add(keras.Input(shape=(784,)))\n","third_model.add(first)\n","third_model.add(second)\n","third_model.add(third)\n","third_model.add(keras.layers.Dense(10))\n","\n","# freeze sub model 1 và sub model 2\n","first.trainable = False\n","second.trainable = False\n","########################################################\n","\n","\n","third_model.compile(optimizer=tf.keras.optimizers.SGD(), \n","              loss=keras.losses.SparseCategoricalCrossentropy(from_logits=True), \n","              metrics=['accuracy'])\n","\n","history = third_model.fit(X_train, y_train, batch_size,\n","                          validation_data=(X_test, y_test),\n","                          epochs=100, verbose=2)"]},{"cell_type":"code","execution_count":13,"id":"sfXVU58oLSV7","metadata":{"id":"sfXVU58oLSV7","executionInfo":{"status":"ok","timestamp":1669968738509,"user_tz":-420,"elapsed":18,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"}}},"outputs":[],"source":[]},{"cell_type":"code","execution_count":14,"id":"LQauDOpjLSoq","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":296},"executionInfo":{"elapsed":562,"status":"ok","timestamp":1669968739058,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"},"user_tz":-420},"id":"LQauDOpjLSoq","outputId":"44e837b4-0dec-42b8-ba7f-84dcd66258f7"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.legend.Legend at 0x7f134b8446d0>"]},"metadata":{},"execution_count":14},{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 1 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}}],"source":["import matplotlib.pyplot as plt\n","\n","plt.plot(history.history['loss'], label='loss')\n","plt.plot(history.history['val_loss'], label='val_loss')\n","plt.xlabel('iteration')\n","plt.ylabel('Loss')\n","plt.legend()"]},{"cell_type":"code","execution_count":14,"id":"R8oxp78TLSor","metadata":{"id":"R8oxp78TLSor","executionInfo":{"status":"ok","timestamp":1669968739058,"user_tz":-420,"elapsed":10,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"}}},"outputs":[],"source":[]},{"cell_type":"code","execution_count":15,"id":"hJ5dHDh-LSos","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":299},"executionInfo":{"elapsed":9,"status":"ok","timestamp":1669968739058,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"},"user_tz":-420},"id":"hJ5dHDh-LSos","outputId":"3d8d19c7-c124-49d6-eba9-f873ab5e4412"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.legend.Legend at 0x7f134b717c70>"]},"metadata":{},"execution_count":15},{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 1 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}}],"source":["import matplotlib.pyplot as plt\n","\n","plt.plot(history.history['accuracy'], label='train_accuracy')\n","plt.plot(history.history['val_accuracy'], label='val_accuracy')\n","plt.xlabel('iteration')\n","plt.ylabel('Accuracy')\n","plt.legend()"]},{"cell_type":"markdown","source":["### Train 5"],"metadata":{"id":"8623181e"},"id":"8623181e"},{"cell_type":"code","execution_count":16,"id":"c84dafda","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":262646,"status":"ok","timestamp":1669969001699,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"},"user_tz":-420},"id":"c84dafda","outputId":"e1480188-9609-484b-8405-0025bb9e70f8"},"outputs":[{"output_type":"stream","name":"stdout","text":["Epoch 1/100\n","118/118 - 3s - loss: 1.9834 - accuracy: 0.4269 - val_loss: 1.7431 - val_accuracy: 0.6561 - 3s/epoch - 26ms/step\n","Epoch 2/100\n","118/118 - 2s - loss: 1.5907 - accuracy: 0.7012 - val_loss: 1.4713 - val_accuracy: 0.7056 - 2s/epoch - 17ms/step\n","Epoch 3/100\n","118/118 - 2s - loss: 1.3644 - accuracy: 0.7392 - val_loss: 1.2910 - val_accuracy: 0.7371 - 2s/epoch - 17ms/step\n","Epoch 4/100\n","118/118 - 2s - loss: 1.2042 - accuracy: 0.7800 - val_loss: 1.1575 - val_accuracy: 0.7709 - 2s/epoch - 17ms/step\n","Epoch 5/100\n","118/118 - 2s - loss: 1.0814 - accuracy: 0.8170 - val_loss: 1.0533 - val_accuracy: 0.8075 - 2s/epoch - 18ms/step\n","Epoch 6/100\n","118/118 - 4s - loss: 0.9843 - accuracy: 0.8437 - val_loss: 0.9695 - val_accuracy: 0.8199 - 4s/epoch - 30ms/step\n","Epoch 7/100\n","118/118 - 2s - loss: 0.9050 - accuracy: 0.8499 - val_loss: 0.9034 - val_accuracy: 0.8242 - 2s/epoch - 20ms/step\n","Epoch 8/100\n","118/118 - 2s - loss: 0.8411 - accuracy: 0.8548 - val_loss: 0.8493 - val_accuracy: 0.8281 - 2s/epoch - 20ms/step\n","Epoch 9/100\n","118/118 - 3s - loss: 0.7878 - accuracy: 0.8579 - val_loss: 0.8046 - val_accuracy: 0.8284 - 3s/epoch - 24ms/step\n","Epoch 10/100\n","118/118 - 2s - loss: 0.7433 - accuracy: 0.8603 - val_loss: 0.7716 - val_accuracy: 0.8318 - 2s/epoch - 19ms/step\n","Epoch 11/100\n","118/118 - 2s - loss: 0.7074 - accuracy: 0.8612 - val_loss: 0.7376 - val_accuracy: 0.8310 - 2s/epoch - 17ms/step\n","Epoch 12/100\n","118/118 - 2s - loss: 0.6737 - accuracy: 0.8633 - val_loss: 0.7099 - val_accuracy: 0.8337 - 2s/epoch - 17ms/step\n","Epoch 13/100\n","118/118 - 2s - loss: 0.6449 - accuracy: 0.8660 - val_loss: 0.6856 - val_accuracy: 0.8343 - 2s/epoch - 17ms/step\n","Epoch 14/100\n","118/118 - 2s - loss: 0.6212 - accuracy: 0.8671 - val_loss: 0.6677 - val_accuracy: 0.8311 - 2s/epoch - 17ms/step\n","Epoch 15/100\n","118/118 - 2s - loss: 0.6017 - accuracy: 0.8669 - val_loss: 0.6508 - val_accuracy: 0.8368 - 2s/epoch - 17ms/step\n","Epoch 16/100\n","118/118 - 2s - loss: 0.5820 - accuracy: 0.8684 - val_loss: 0.6365 - val_accuracy: 0.8355 - 2s/epoch - 17ms/step\n","Epoch 17/100\n","118/118 - 2s - loss: 0.5660 - accuracy: 0.8689 - val_loss: 0.6209 - val_accuracy: 0.8384 - 2s/epoch - 17ms/step\n","Epoch 18/100\n","118/118 - 2s - loss: 0.5512 - accuracy: 0.8694 - val_loss: 0.6098 - val_accuracy: 0.8354 - 2s/epoch - 17ms/step\n","Epoch 19/100\n","118/118 - 2s - loss: 0.5374 - accuracy: 0.8702 - val_loss: 0.5996 - val_accuracy: 0.8368 - 2s/epoch - 17ms/step\n","Epoch 20/100\n","118/118 - 2s - loss: 0.5244 - accuracy: 0.8709 - val_loss: 0.5900 - val_accuracy: 0.8332 - 2s/epoch - 17ms/step\n","Epoch 21/100\n","118/118 - 2s - loss: 0.5152 - accuracy: 0.8705 - val_loss: 0.5848 - val_accuracy: 0.8341 - 2s/epoch - 17ms/step\n","Epoch 22/100\n","118/118 - 2s - loss: 0.5069 - accuracy: 0.8717 - val_loss: 0.5744 - val_accuracy: 0.8366 - 2s/epoch - 17ms/step\n","Epoch 23/100\n","118/118 - 2s - loss: 0.4967 - accuracy: 0.8716 - val_loss: 0.5651 - val_accuracy: 0.8402 - 2s/epoch - 17ms/step\n","Epoch 24/100\n","118/118 - 2s - loss: 0.4880 - accuracy: 0.8723 - val_loss: 0.5572 - val_accuracy: 0.8372 - 2s/epoch - 17ms/step\n","Epoch 25/100\n","118/118 - 2s - loss: 0.4810 - accuracy: 0.8719 - val_loss: 0.5569 - val_accuracy: 0.8376 - 2s/epoch - 17ms/step\n","Epoch 26/100\n","118/118 - 2s - loss: 0.4756 - accuracy: 0.8719 - val_loss: 0.5519 - val_accuracy: 0.8387 - 2s/epoch - 17ms/step\n","Epoch 27/100\n","118/118 - 2s - loss: 0.4696 - accuracy: 0.8723 - val_loss: 0.5535 - val_accuracy: 0.8332 - 2s/epoch - 17ms/step\n","Epoch 28/100\n","118/118 - 2s - loss: 0.4648 - accuracy: 0.8725 - val_loss: 0.5420 - val_accuracy: 0.8367 - 2s/epoch - 17ms/step\n","Epoch 29/100\n","118/118 - 2s - loss: 0.4593 - accuracy: 0.8727 - val_loss: 0.5408 - val_accuracy: 0.8379 - 2s/epoch - 17ms/step\n","Epoch 30/100\n","118/118 - 2s - loss: 0.4545 - accuracy: 0.8713 - val_loss: 0.5319 - val_accuracy: 0.8400 - 2s/epoch - 17ms/step\n","Epoch 31/100\n","118/118 - 2s - loss: 0.4460 - accuracy: 0.8741 - val_loss: 0.5295 - val_accuracy: 0.8426 - 2s/epoch - 17ms/step\n","Epoch 32/100\n","118/118 - 2s - loss: 0.4415 - accuracy: 0.8756 - val_loss: 0.5260 - val_accuracy: 0.8391 - 2s/epoch - 17ms/step\n","Epoch 33/100\n","118/118 - 2s - loss: 0.4387 - accuracy: 0.8741 - val_loss: 0.5235 - val_accuracy: 0.8404 - 2s/epoch - 17ms/step\n","Epoch 34/100\n","118/118 - 2s - loss: 0.4348 - accuracy: 0.8745 - val_loss: 0.5278 - val_accuracy: 0.8365 - 2s/epoch - 17ms/step\n","Epoch 35/100\n","118/118 - 2s - loss: 0.4371 - accuracy: 0.8721 - val_loss: 0.5160 - val_accuracy: 0.8395 - 2s/epoch - 17ms/step\n","Epoch 36/100\n","118/118 - 2s - loss: 0.4307 - accuracy: 0.8744 - val_loss: 0.5130 - val_accuracy: 0.8434 - 2s/epoch - 17ms/step\n","Epoch 37/100\n","118/118 - 2s - loss: 0.4270 - accuracy: 0.8750 - val_loss: 0.5154 - val_accuracy: 0.8367 - 2s/epoch - 17ms/step\n","Epoch 38/100\n","118/118 - 2s - loss: 0.4238 - accuracy: 0.8745 - val_loss: 0.5073 - val_accuracy: 0.8413 - 2s/epoch - 17ms/step\n","Epoch 39/100\n","118/118 - 2s - loss: 0.4219 - accuracy: 0.8752 - val_loss: 0.5328 - val_accuracy: 0.8330 - 2s/epoch - 17ms/step\n","Epoch 40/100\n","118/118 - 2s - loss: 0.4219 - accuracy: 0.8752 - val_loss: 0.5068 - val_accuracy: 0.8401 - 2s/epoch - 17ms/step\n","Epoch 41/100\n","118/118 - 2s - loss: 0.4202 - accuracy: 0.8747 - val_loss: 0.5015 - val_accuracy: 0.8441 - 2s/epoch - 17ms/step\n","Epoch 42/100\n","118/118 - 2s - loss: 0.4154 - accuracy: 0.8756 - val_loss: 0.5096 - val_accuracy: 0.8393 - 2s/epoch - 17ms/step\n","Epoch 43/100\n","118/118 - 2s - loss: 0.4122 - accuracy: 0.8763 - val_loss: 0.4991 - val_accuracy: 0.8428 - 2s/epoch - 17ms/step\n","Epoch 44/100\n","118/118 - 2s - loss: 0.4091 - accuracy: 0.8767 - val_loss: 0.5016 - val_accuracy: 0.8411 - 2s/epoch - 18ms/step\n","Epoch 45/100\n","118/118 - 2s - loss: 0.4106 - accuracy: 0.8753 - val_loss: 0.5046 - val_accuracy: 0.8402 - 2s/epoch - 17ms/step\n","Epoch 46/100\n","118/118 - 2s - loss: 0.4070 - accuracy: 0.8764 - val_loss: 0.4967 - val_accuracy: 0.8392 - 2s/epoch - 17ms/step\n","Epoch 47/100\n","118/118 - 2s - loss: 0.4024 - accuracy: 0.8774 - val_loss: 0.5041 - val_accuracy: 0.8400 - 2s/epoch - 17ms/step\n","Epoch 48/100\n","118/118 - 2s - loss: 0.4041 - accuracy: 0.8763 - val_loss: 0.4947 - val_accuracy: 0.8432 - 2s/epoch - 17ms/step\n","Epoch 49/100\n","118/118 - 2s - loss: 0.3982 - accuracy: 0.8787 - val_loss: 0.5013 - val_accuracy: 0.8368 - 2s/epoch - 17ms/step\n","Epoch 50/100\n","118/118 - 2s - loss: 0.4015 - accuracy: 0.8759 - val_loss: 0.4885 - val_accuracy: 0.8438 - 2s/epoch - 17ms/step\n","Epoch 51/100\n","118/118 - 2s - loss: 0.3934 - accuracy: 0.8775 - val_loss: 0.4887 - val_accuracy: 0.8432 - 2s/epoch - 17ms/step\n","Epoch 52/100\n","118/118 - 2s - loss: 0.3933 - accuracy: 0.8776 - val_loss: 0.4899 - val_accuracy: 0.8440 - 2s/epoch - 17ms/step\n","Epoch 53/100\n","118/118 - 2s - loss: 0.3953 - accuracy: 0.8781 - val_loss: 0.5047 - val_accuracy: 0.8384 - 2s/epoch - 17ms/step\n","Epoch 54/100\n","118/118 - 2s - loss: 0.3951 - accuracy: 0.8766 - val_loss: 0.4907 - val_accuracy: 0.8380 - 2s/epoch - 17ms/step\n","Epoch 55/100\n","118/118 - 2s - loss: 0.3928 - accuracy: 0.8772 - val_loss: 0.4972 - val_accuracy: 0.8377 - 2s/epoch - 17ms/step\n","Epoch 56/100\n","118/118 - 2s - loss: 0.4036 - accuracy: 0.8741 - val_loss: 0.4971 - val_accuracy: 0.8378 - 2s/epoch - 17ms/step\n","Epoch 57/100\n","118/118 - 2s - loss: 0.3947 - accuracy: 0.8761 - val_loss: 0.4864 - val_accuracy: 0.8448 - 2s/epoch - 17ms/step\n","Epoch 58/100\n","118/118 - 2s - loss: 0.3934 - accuracy: 0.8770 - val_loss: 0.4948 - val_accuracy: 0.8437 - 2s/epoch - 17ms/step\n","Epoch 59/100\n","118/118 - 2s - loss: 0.3917 - accuracy: 0.8770 - val_loss: 0.4925 - val_accuracy: 0.8401 - 2s/epoch - 17ms/step\n","Epoch 60/100\n","118/118 - 2s - loss: 0.3878 - accuracy: 0.8768 - val_loss: 0.4873 - val_accuracy: 0.8393 - 2s/epoch - 17ms/step\n","Epoch 61/100\n","118/118 - 2s - loss: 0.3876 - accuracy: 0.8782 - val_loss: 0.4911 - val_accuracy: 0.8418 - 2s/epoch - 17ms/step\n","Epoch 62/100\n","118/118 - 2s - loss: 0.3911 - accuracy: 0.8758 - val_loss: 0.4952 - val_accuracy: 0.8419 - 2s/epoch - 17ms/step\n","Epoch 63/100\n","118/118 - 2s - loss: 0.3922 - accuracy: 0.8759 - val_loss: 0.4859 - val_accuracy: 0.8424 - 2s/epoch - 17ms/step\n","Epoch 64/100\n","118/118 - 2s - loss: 0.3886 - accuracy: 0.8770 - val_loss: 0.4849 - val_accuracy: 0.8436 - 2s/epoch - 17ms/step\n","Epoch 65/100\n","118/118 - 2s - loss: 0.3853 - accuracy: 0.8775 - val_loss: 0.4860 - val_accuracy: 0.8413 - 2s/epoch - 17ms/step\n","Epoch 66/100\n","118/118 - 2s - loss: 0.3820 - accuracy: 0.8791 - val_loss: 0.4932 - val_accuracy: 0.8360 - 2s/epoch - 17ms/step\n","Epoch 67/100\n","118/118 - 2s - loss: 0.3823 - accuracy: 0.8768 - val_loss: 0.4819 - val_accuracy: 0.8439 - 2s/epoch - 17ms/step\n","Epoch 68/100\n","118/118 - 2s - loss: 0.3789 - accuracy: 0.8783 - val_loss: 0.4741 - val_accuracy: 0.8474 - 2s/epoch - 17ms/step\n","Epoch 69/100\n","118/118 - 2s - loss: 0.3809 - accuracy: 0.8787 - val_loss: 0.4794 - val_accuracy: 0.8427 - 2s/epoch - 17ms/step\n","Epoch 70/100\n","118/118 - 2s - loss: 0.3813 - accuracy: 0.8774 - val_loss: 0.4794 - val_accuracy: 0.8442 - 2s/epoch - 17ms/step\n","Epoch 71/100\n","118/118 - 2s - loss: 0.3784 - accuracy: 0.8786 - val_loss: 0.4767 - val_accuracy: 0.8441 - 2s/epoch - 16ms/step\n","Epoch 72/100\n","118/118 - 2s - loss: 0.3849 - accuracy: 0.8759 - val_loss: 0.4955 - val_accuracy: 0.8362 - 2s/epoch - 17ms/step\n","Epoch 73/100\n","118/118 - 2s - loss: 0.3787 - accuracy: 0.8786 - val_loss: 0.4771 - val_accuracy: 0.8446 - 2s/epoch - 17ms/step\n","Epoch 74/100\n","118/118 - 2s - loss: 0.3772 - accuracy: 0.8790 - val_loss: 0.4839 - val_accuracy: 0.8414 - 2s/epoch - 17ms/step\n","Epoch 75/100\n","118/118 - 2s - loss: 0.3807 - accuracy: 0.8779 - val_loss: 0.4836 - val_accuracy: 0.8436 - 2s/epoch - 17ms/step\n","Epoch 76/100\n","118/118 - 2s - loss: 0.3844 - accuracy: 0.8757 - val_loss: 0.4890 - val_accuracy: 0.8384 - 2s/epoch - 16ms/step\n","Epoch 77/100\n","118/118 - 2s - loss: 0.3860 - accuracy: 0.8756 - val_loss: 0.4832 - val_accuracy: 0.8458 - 2s/epoch - 17ms/step\n","Epoch 78/100\n","118/118 - 2s - loss: 0.3847 - accuracy: 0.8748 - val_loss: 0.4829 - val_accuracy: 0.8404 - 2s/epoch - 17ms/step\n","Epoch 79/100\n","118/118 - 2s - loss: 0.3810 - accuracy: 0.8769 - val_loss: 0.4813 - val_accuracy: 0.8439 - 2s/epoch - 17ms/step\n","Epoch 80/100\n","118/118 - 2s - loss: 0.3765 - accuracy: 0.8779 - val_loss: 0.4919 - val_accuracy: 0.8413 - 2s/epoch - 17ms/step\n","Epoch 81/100\n","118/118 - 2s - loss: 0.3783 - accuracy: 0.8764 - val_loss: 0.4866 - val_accuracy: 0.8411 - 2s/epoch - 17ms/step\n","Epoch 82/100\n","118/118 - 2s - loss: 0.3818 - accuracy: 0.8747 - val_loss: 0.4726 - val_accuracy: 0.8454 - 2s/epoch - 17ms/step\n","Epoch 83/100\n","118/118 - 2s - loss: 0.3746 - accuracy: 0.8774 - val_loss: 0.4748 - val_accuracy: 0.8451 - 2s/epoch - 17ms/step\n","Epoch 84/100\n","118/118 - 2s - loss: 0.3752 - accuracy: 0.8781 - val_loss: 0.4702 - val_accuracy: 0.8451 - 2s/epoch - 17ms/step\n","Epoch 85/100\n","118/118 - 2s - loss: 0.3761 - accuracy: 0.8788 - val_loss: 0.4751 - val_accuracy: 0.8446 - 2s/epoch - 17ms/step\n","Epoch 86/100\n","118/118 - 2s - loss: 0.3739 - accuracy: 0.8775 - val_loss: 0.4801 - val_accuracy: 0.8413 - 2s/epoch - 17ms/step\n","Epoch 87/100\n","118/118 - 2s - loss: 0.3764 - accuracy: 0.8762 - val_loss: 0.4944 - val_accuracy: 0.8359 - 2s/epoch - 17ms/step\n","Epoch 88/100\n","118/118 - 2s - loss: 0.3837 - accuracy: 0.8743 - val_loss: 0.4776 - val_accuracy: 0.8409 - 2s/epoch - 17ms/step\n","Epoch 89/100\n","118/118 - 2s - loss: 0.3798 - accuracy: 0.8759 - val_loss: 0.4894 - val_accuracy: 0.8349 - 2s/epoch - 17ms/step\n","Epoch 90/100\n","118/118 - 2s - loss: 0.3740 - accuracy: 0.8773 - val_loss: 0.4892 - val_accuracy: 0.8406 - 2s/epoch - 17ms/step\n","Epoch 91/100\n","118/118 - 2s - loss: 0.3707 - accuracy: 0.8774 - val_loss: 0.4883 - val_accuracy: 0.8380 - 2s/epoch - 17ms/step\n","Epoch 92/100\n","118/118 - 2s - loss: 0.3736 - accuracy: 0.8760 - val_loss: 0.4700 - val_accuracy: 0.8423 - 2s/epoch - 17ms/step\n","Epoch 93/100\n","118/118 - 2s - loss: 0.3707 - accuracy: 0.8777 - val_loss: 0.4804 - val_accuracy: 0.8425 - 2s/epoch - 17ms/step\n","Epoch 94/100\n","118/118 - 2s - loss: 0.3803 - accuracy: 0.8751 - val_loss: 0.4811 - val_accuracy: 0.8367 - 2s/epoch - 17ms/step\n","Epoch 95/100\n","118/118 - 2s - loss: 0.3824 - accuracy: 0.8737 - val_loss: 0.4752 - val_accuracy: 0.8440 - 2s/epoch - 17ms/step\n","Epoch 96/100\n","118/118 - 2s - loss: 0.3779 - accuracy: 0.8757 - val_loss: 0.4983 - val_accuracy: 0.8354 - 2s/epoch - 17ms/step\n","Epoch 97/100\n","118/118 - 2s - loss: 0.3875 - accuracy: 0.8731 - val_loss: 0.4888 - val_accuracy: 0.8332 - 2s/epoch - 17ms/step\n","Epoch 98/100\n","118/118 - 2s - loss: 0.3804 - accuracy: 0.8745 - val_loss: 0.4794 - val_accuracy: 0.8425 - 2s/epoch - 17ms/step\n","Epoch 99/100\n","118/118 - 2s - loss: 0.3818 - accuracy: 0.8746 - val_loss: 0.4838 - val_accuracy: 0.8396 - 2s/epoch - 17ms/step\n","Epoch 100/100\n","118/118 - 2s - loss: 0.3788 - accuracy: 0.8768 - val_loss: 0.4705 - val_accuracy: 0.8447 - 2s/epoch - 17ms/step\n"]}],"source":["# train first, second, and third\n","#################### YOUR CODE HERE ####################\n","# Các bạn thực hiện tương tự như Ví dụ train lần 1 \n","# Train lần 5: Tiếp tục train thêm 100 epoch nhưng lần này không fix weight của sub model 1 và sub model 2\n","# (tạo model mới và thêm Input layer, sub model 1, sub model 2, sub model 3, Output layer)\n","third_model = keras.Sequential()\n","third_model.add(keras.Input(shape=(784,)))\n","third_model.add(first)\n","third_model.add(second)\n","third_model.add(third)\n","third_model.add(keras.layers.Dense(10))\n","\n","# unfreeze sub model 1 và sub model 2\n","first.trainable = True\n","second.trainable = True\n","########################################################\n","\n","\n","third_model.compile(optimizer=tf.keras.optimizers.SGD(), \n","              loss=keras.losses.SparseCategoricalCrossentropy(from_logits=True), \n","              metrics=['accuracy'])\n","\n","history = third_model.fit(X_train, y_train, batch_size,\n","                          validation_data=(X_test, y_test),\n","                          epochs=100, verbose=2)"]},{"cell_type":"code","execution_count":16,"id":"HD0MljlBLVPn","metadata":{"id":"HD0MljlBLVPn","executionInfo":{"status":"ok","timestamp":1669969001700,"user_tz":-420,"elapsed":6,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"}}},"outputs":[],"source":[]},{"cell_type":"code","execution_count":17,"id":"QeeOXyzGLVfP","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":296},"executionInfo":{"elapsed":6,"status":"ok","timestamp":1669969001701,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"},"user_tz":-420},"id":"QeeOXyzGLVfP","outputId":"341aff47-ac30-48e4-ac7c-5eba4f2b455d"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.legend.Legend at 0x7f134b4d5c40>"]},"metadata":{},"execution_count":17},{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 1 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}}],"source":["import matplotlib.pyplot as plt\n","\n","plt.plot(history.history['loss'], label='loss')\n","plt.plot(history.history['val_loss'], label='val_loss')\n","plt.xlabel('iteration')\n","plt.ylabel('Loss')\n","plt.legend()"]},{"cell_type":"code","execution_count":17,"id":"tWVWG6bVLVfS","metadata":{"id":"tWVWG6bVLVfS","executionInfo":{"status":"ok","timestamp":1669969001701,"user_tz":-420,"elapsed":4,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"}}},"outputs":[],"source":[]},{"cell_type":"code","execution_count":18,"id":"URKj17ikLVfT","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":300},"executionInfo":{"elapsed":389,"status":"ok","timestamp":1669969002086,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"},"user_tz":-420},"id":"URKj17ikLVfT","outputId":"4a82577c-fbe8-4e02-d52a-6d8a24d03b73"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.legend.Legend at 0x7f134b49cca0>"]},"metadata":{},"execution_count":18},{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 1 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}}],"source":["import matplotlib.pyplot as plt\n","\n","plt.plot(history.history['accuracy'], label='train_accuracy')\n","plt.plot(history.history['val_accuracy'], label='val_accuracy')\n","plt.xlabel('iteration')\n","plt.ylabel('Accuracy')\n","plt.legend()"]},{"cell_type":"markdown","source":["### Train 6"],"metadata":{"id":"27cc22bf"},"id":"27cc22bf"},{"cell_type":"code","execution_count":19,"id":"16c07e2a","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":151363,"status":"ok","timestamp":1669969153446,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"},"user_tz":-420},"id":"16c07e2a","outputId":"90f5e60d-fc40-405a-b47f-a8a0ae0bb76d"},"outputs":[{"output_type":"stream","name":"stdout","text":["Epoch 1/100\n","118/118 - 2s - loss: 2.2653 - accuracy: 0.2558 - val_loss: 2.1892 - val_accuracy: 0.4890 - 2s/epoch - 20ms/step\n","Epoch 2/100\n","118/118 - 1s - loss: 2.1238 - accuracy: 0.6049 - val_loss: 2.0627 - val_accuracy: 0.6744 - 1s/epoch - 12ms/step\n","Epoch 3/100\n","118/118 - 1s - loss: 1.9971 - accuracy: 0.7426 - val_loss: 1.9395 - val_accuracy: 0.7672 - 1s/epoch - 12ms/step\n","Epoch 4/100\n","118/118 - 1s - loss: 1.8726 - accuracy: 0.8174 - val_loss: 1.8179 - val_accuracy: 0.8099 - 1s/epoch - 12ms/step\n","Epoch 5/100\n","118/118 - 1s - loss: 1.7503 - accuracy: 0.8330 - val_loss: 1.6996 - val_accuracy: 0.8031 - 1s/epoch - 12ms/step\n","Epoch 6/100\n","118/118 - 1s - loss: 1.6322 - accuracy: 0.8400 - val_loss: 1.5870 - val_accuracy: 0.8116 - 1s/epoch - 13ms/step\n","Epoch 7/100\n","118/118 - 1s - loss: 1.5203 - accuracy: 0.8480 - val_loss: 1.4814 - val_accuracy: 0.8197 - 1s/epoch - 13ms/step\n","Epoch 8/100\n","118/118 - 1s - loss: 1.4162 - accuracy: 0.8540 - val_loss: 1.3843 - val_accuracy: 0.8247 - 1s/epoch - 12ms/step\n","Epoch 9/100\n","118/118 - 1s - loss: 1.3204 - accuracy: 0.8583 - val_loss: 1.2957 - val_accuracy: 0.8297 - 1s/epoch - 12ms/step\n","Epoch 10/100\n","118/118 - 1s - loss: 1.2331 - accuracy: 0.8624 - val_loss: 1.2155 - val_accuracy: 0.8288 - 1s/epoch - 12ms/step\n","Epoch 11/100\n","118/118 - 1s - loss: 1.1541 - accuracy: 0.8644 - val_loss: 1.1433 - val_accuracy: 0.8335 - 1s/epoch - 12ms/step\n","Epoch 12/100\n","118/118 - 1s - loss: 1.0828 - accuracy: 0.8656 - val_loss: 1.0786 - val_accuracy: 0.8351 - 1s/epoch - 12ms/step\n","Epoch 13/100\n","118/118 - 1s - loss: 1.0186 - accuracy: 0.8689 - val_loss: 1.0203 - val_accuracy: 0.8349 - 1s/epoch - 12ms/step\n","Epoch 14/100\n","118/118 - 1s - loss: 0.9609 - accuracy: 0.8692 - val_loss: 0.9684 - val_accuracy: 0.8358 - 1s/epoch - 12ms/step\n","Epoch 15/100\n","118/118 - 2s - loss: 0.9091 - accuracy: 0.8708 - val_loss: 0.9221 - val_accuracy: 0.8361 - 2s/epoch - 14ms/step\n","Epoch 16/100\n","118/118 - 2s - loss: 0.8626 - accuracy: 0.8719 - val_loss: 0.8808 - val_accuracy: 0.8353 - 2s/epoch - 14ms/step\n","Epoch 17/100\n","118/118 - 2s - loss: 0.8210 - accuracy: 0.8717 - val_loss: 0.8438 - val_accuracy: 0.8378 - 2s/epoch - 14ms/step\n","Epoch 18/100\n","118/118 - 2s - loss: 0.7837 - accuracy: 0.8731 - val_loss: 0.8109 - val_accuracy: 0.8382 - 2s/epoch - 14ms/step\n","Epoch 19/100\n","118/118 - 2s - loss: 0.7502 - accuracy: 0.8741 - val_loss: 0.7814 - val_accuracy: 0.8382 - 2s/epoch - 14ms/step\n","Epoch 20/100\n","118/118 - 2s - loss: 0.7201 - accuracy: 0.8744 - val_loss: 0.7550 - val_accuracy: 0.8381 - 2s/epoch - 14ms/step\n","Epoch 21/100\n","118/118 - 2s - loss: 0.6931 - accuracy: 0.8746 - val_loss: 0.7315 - val_accuracy: 0.8391 - 2s/epoch - 14ms/step\n","Epoch 22/100\n","118/118 - 2s - loss: 0.6687 - accuracy: 0.8749 - val_loss: 0.7103 - val_accuracy: 0.8395 - 2s/epoch - 14ms/step\n","Epoch 23/100\n","118/118 - 2s - loss: 0.6466 - accuracy: 0.8750 - val_loss: 0.6915 - val_accuracy: 0.8423 - 2s/epoch - 13ms/step\n","Epoch 24/100\n","118/118 - 2s - loss: 0.6267 - accuracy: 0.8758 - val_loss: 0.6742 - val_accuracy: 0.8406 - 2s/epoch - 13ms/step\n","Epoch 25/100\n","118/118 - 2s - loss: 0.6086 - accuracy: 0.8760 - val_loss: 0.6587 - val_accuracy: 0.8405 - 2s/epoch - 14ms/step\n","Epoch 26/100\n","118/118 - 2s - loss: 0.5921 - accuracy: 0.8763 - val_loss: 0.6448 - val_accuracy: 0.8403 - 2s/epoch - 13ms/step\n","Epoch 27/100\n","118/118 - 2s - loss: 0.5771 - accuracy: 0.8762 - val_loss: 0.6321 - val_accuracy: 0.8416 - 2s/epoch - 13ms/step\n","Epoch 28/100\n","118/118 - 1s - loss: 0.5634 - accuracy: 0.8762 - val_loss: 0.6206 - val_accuracy: 0.8406 - 1s/epoch - 12ms/step\n","Epoch 29/100\n","118/118 - 1s - loss: 0.5509 - accuracy: 0.8766 - val_loss: 0.6102 - val_accuracy: 0.8422 - 1s/epoch - 12ms/step\n","Epoch 30/100\n","118/118 - 1s - loss: 0.5395 - accuracy: 0.8769 - val_loss: 0.6007 - val_accuracy: 0.8433 - 1s/epoch - 12ms/step\n","Epoch 31/100\n","118/118 - 2s - loss: 0.5289 - accuracy: 0.8772 - val_loss: 0.5920 - val_accuracy: 0.8421 - 2s/epoch - 13ms/step\n","Epoch 32/100\n","118/118 - 2s - loss: 0.5192 - accuracy: 0.8774 - val_loss: 0.5840 - val_accuracy: 0.8420 - 2s/epoch - 14ms/step\n","Epoch 33/100\n","118/118 - 2s - loss: 0.5102 - accuracy: 0.8775 - val_loss: 0.5767 - val_accuracy: 0.8422 - 2s/epoch - 13ms/step\n","Epoch 34/100\n","118/118 - 2s - loss: 0.5019 - accuracy: 0.8775 - val_loss: 0.5700 - val_accuracy: 0.8426 - 2s/epoch - 13ms/step\n","Epoch 35/100\n","118/118 - 2s - loss: 0.4943 - accuracy: 0.8781 - val_loss: 0.5638 - val_accuracy: 0.8425 - 2s/epoch - 13ms/step\n","Epoch 36/100\n","118/118 - 2s - loss: 0.4871 - accuracy: 0.8779 - val_loss: 0.5582 - val_accuracy: 0.8429 - 2s/epoch - 17ms/step\n","Epoch 37/100\n","118/118 - 3s - loss: 0.4805 - accuracy: 0.8781 - val_loss: 0.5528 - val_accuracy: 0.8423 - 3s/epoch - 22ms/step\n","Epoch 38/100\n","118/118 - 1s - loss: 0.4744 - accuracy: 0.8781 - val_loss: 0.5480 - val_accuracy: 0.8428 - 1s/epoch - 12ms/step\n","Epoch 39/100\n","118/118 - 1s - loss: 0.4686 - accuracy: 0.8787 - val_loss: 0.5435 - val_accuracy: 0.8432 - 1s/epoch - 12ms/step\n","Epoch 40/100\n","118/118 - 1s - loss: 0.4633 - accuracy: 0.8787 - val_loss: 0.5395 - val_accuracy: 0.8425 - 1s/epoch - 12ms/step\n","Epoch 41/100\n","118/118 - 1s - loss: 0.4583 - accuracy: 0.8788 - val_loss: 0.5356 - val_accuracy: 0.8425 - 1s/epoch - 12ms/step\n","Epoch 42/100\n","118/118 - 1s - loss: 0.4536 - accuracy: 0.8790 - val_loss: 0.5320 - val_accuracy: 0.8431 - 1s/epoch - 12ms/step\n","Epoch 43/100\n","118/118 - 1s - loss: 0.4493 - accuracy: 0.8789 - val_loss: 0.5287 - val_accuracy: 0.8432 - 1s/epoch - 12ms/step\n","Epoch 44/100\n","118/118 - 2s - loss: 0.4451 - accuracy: 0.8790 - val_loss: 0.5256 - val_accuracy: 0.8429 - 2s/epoch - 13ms/step\n","Epoch 45/100\n","118/118 - 1s - loss: 0.4413 - accuracy: 0.8788 - val_loss: 0.5228 - val_accuracy: 0.8440 - 1s/epoch - 12ms/step\n","Epoch 46/100\n","118/118 - 1s - loss: 0.4376 - accuracy: 0.8792 - val_loss: 0.5201 - val_accuracy: 0.8431 - 1s/epoch - 12ms/step\n","Epoch 47/100\n","118/118 - 1s - loss: 0.4342 - accuracy: 0.8794 - val_loss: 0.5176 - val_accuracy: 0.8434 - 1s/epoch - 12ms/step\n","Epoch 48/100\n","118/118 - 1s - loss: 0.4310 - accuracy: 0.8795 - val_loss: 0.5153 - val_accuracy: 0.8434 - 1s/epoch - 12ms/step\n","Epoch 49/100\n","118/118 - 2s - loss: 0.4280 - accuracy: 0.8793 - val_loss: 0.5130 - val_accuracy: 0.8435 - 2s/epoch - 13ms/step\n","Epoch 50/100\n","118/118 - 2s - loss: 0.4251 - accuracy: 0.8793 - val_loss: 0.5110 - val_accuracy: 0.8439 - 2s/epoch - 13ms/step\n","Epoch 51/100\n","118/118 - 2s - loss: 0.4224 - accuracy: 0.8794 - val_loss: 0.5091 - val_accuracy: 0.8439 - 2s/epoch - 13ms/step\n","Epoch 52/100\n","118/118 - 2s - loss: 0.4199 - accuracy: 0.8796 - val_loss: 0.5074 - val_accuracy: 0.8436 - 2s/epoch - 13ms/step\n","Epoch 53/100\n","118/118 - 2s - loss: 0.4175 - accuracy: 0.8795 - val_loss: 0.5058 - val_accuracy: 0.8435 - 2s/epoch - 14ms/step\n","Epoch 54/100\n","118/118 - 2s - loss: 0.4151 - accuracy: 0.8798 - val_loss: 0.5043 - val_accuracy: 0.8438 - 2s/epoch - 13ms/step\n","Epoch 55/100\n","118/118 - 1s - loss: 0.4130 - accuracy: 0.8795 - val_loss: 0.5027 - val_accuracy: 0.8441 - 1s/epoch - 12ms/step\n","Epoch 56/100\n","118/118 - 1s - loss: 0.4109 - accuracy: 0.8796 - val_loss: 0.5013 - val_accuracy: 0.8438 - 1s/epoch - 12ms/step\n","Epoch 57/100\n","118/118 - 1s - loss: 0.4090 - accuracy: 0.8801 - val_loss: 0.4999 - val_accuracy: 0.8444 - 1s/epoch - 12ms/step\n","Epoch 58/100\n","118/118 - 1s - loss: 0.4071 - accuracy: 0.8795 - val_loss: 0.4987 - val_accuracy: 0.8438 - 1s/epoch - 12ms/step\n","Epoch 59/100\n","118/118 - 1s - loss: 0.4054 - accuracy: 0.8797 - val_loss: 0.4976 - val_accuracy: 0.8441 - 1s/epoch - 12ms/step\n","Epoch 60/100\n","118/118 - 1s - loss: 0.4037 - accuracy: 0.8795 - val_loss: 0.4965 - val_accuracy: 0.8441 - 1s/epoch - 12ms/step\n","Epoch 61/100\n","118/118 - 1s - loss: 0.4021 - accuracy: 0.8800 - val_loss: 0.4955 - val_accuracy: 0.8443 - 1s/epoch - 12ms/step\n","Epoch 62/100\n","118/118 - 1s - loss: 0.4005 - accuracy: 0.8798 - val_loss: 0.4945 - val_accuracy: 0.8442 - 1s/epoch - 12ms/step\n","Epoch 63/100\n","118/118 - 1s - loss: 0.3991 - accuracy: 0.8797 - val_loss: 0.4936 - val_accuracy: 0.8442 - 1s/epoch - 12ms/step\n","Epoch 64/100\n","118/118 - 1s - loss: 0.3977 - accuracy: 0.8799 - val_loss: 0.4928 - val_accuracy: 0.8439 - 1s/epoch - 12ms/step\n","Epoch 65/100\n","118/118 - 1s - loss: 0.3964 - accuracy: 0.8801 - val_loss: 0.4920 - val_accuracy: 0.8444 - 1s/epoch - 13ms/step\n","Epoch 66/100\n","118/118 - 2s - loss: 0.3951 - accuracy: 0.8799 - val_loss: 0.4912 - val_accuracy: 0.8440 - 2s/epoch - 13ms/step\n","Epoch 67/100\n","118/118 - 2s - loss: 0.3938 - accuracy: 0.8799 - val_loss: 0.4905 - val_accuracy: 0.8441 - 2s/epoch - 13ms/step\n","Epoch 68/100\n","118/118 - 2s - loss: 0.3927 - accuracy: 0.8801 - val_loss: 0.4898 - val_accuracy: 0.8446 - 2s/epoch - 13ms/step\n","Epoch 69/100\n","118/118 - 2s - loss: 0.3916 - accuracy: 0.8801 - val_loss: 0.4892 - val_accuracy: 0.8444 - 2s/epoch - 13ms/step\n","Epoch 70/100\n","118/118 - 2s - loss: 0.3905 - accuracy: 0.8800 - val_loss: 0.4886 - val_accuracy: 0.8442 - 2s/epoch - 13ms/step\n","Epoch 71/100\n","118/118 - 2s - loss: 0.3895 - accuracy: 0.8800 - val_loss: 0.4880 - val_accuracy: 0.8444 - 2s/epoch - 13ms/step\n","Epoch 72/100\n","118/118 - 2s - loss: 0.3885 - accuracy: 0.8799 - val_loss: 0.4874 - val_accuracy: 0.8443 - 2s/epoch - 13ms/step\n","Epoch 73/100\n","118/118 - 2s - loss: 0.3876 - accuracy: 0.8800 - val_loss: 0.4869 - val_accuracy: 0.8443 - 2s/epoch - 13ms/step\n","Epoch 74/100\n","118/118 - 2s - loss: 0.3866 - accuracy: 0.8799 - val_loss: 0.4865 - val_accuracy: 0.8443 - 2s/epoch - 13ms/step\n","Epoch 75/100\n","118/118 - 2s - loss: 0.3858 - accuracy: 0.8799 - val_loss: 0.4860 - val_accuracy: 0.8445 - 2s/epoch - 13ms/step\n","Epoch 76/100\n","118/118 - 2s - loss: 0.3850 - accuracy: 0.8798 - val_loss: 0.4855 - val_accuracy: 0.8446 - 2s/epoch - 13ms/step\n","Epoch 77/100\n","118/118 - 2s - loss: 0.3841 - accuracy: 0.8801 - val_loss: 0.4850 - val_accuracy: 0.8448 - 2s/epoch - 13ms/step\n","Epoch 78/100\n","118/118 - 2s - loss: 0.3834 - accuracy: 0.8802 - val_loss: 0.4847 - val_accuracy: 0.8447 - 2s/epoch - 13ms/step\n","Epoch 79/100\n","118/118 - 2s - loss: 0.3826 - accuracy: 0.8801 - val_loss: 0.4844 - val_accuracy: 0.8449 - 2s/epoch - 13ms/step\n","Epoch 80/100\n","118/118 - 2s - loss: 0.3819 - accuracy: 0.8799 - val_loss: 0.4840 - val_accuracy: 0.8449 - 2s/epoch - 13ms/step\n","Epoch 81/100\n","118/118 - 2s - loss: 0.3812 - accuracy: 0.8802 - val_loss: 0.4836 - val_accuracy: 0.8451 - 2s/epoch - 13ms/step\n","Epoch 82/100\n","118/118 - 1s - loss: 0.3805 - accuracy: 0.8801 - val_loss: 0.4833 - val_accuracy: 0.8447 - 1s/epoch - 11ms/step\n","Epoch 83/100\n","118/118 - 1s - loss: 0.3799 - accuracy: 0.8802 - val_loss: 0.4829 - val_accuracy: 0.8448 - 1s/epoch - 12ms/step\n","Epoch 84/100\n","118/118 - 1s - loss: 0.3793 - accuracy: 0.8802 - val_loss: 0.4827 - val_accuracy: 0.8451 - 1s/epoch - 11ms/step\n","Epoch 85/100\n","118/118 - 1s - loss: 0.3787 - accuracy: 0.8802 - val_loss: 0.4824 - val_accuracy: 0.8452 - 1s/epoch - 12ms/step\n","Epoch 86/100\n","118/118 - 1s - loss: 0.3781 - accuracy: 0.8804 - val_loss: 0.4821 - val_accuracy: 0.8448 - 1s/epoch - 11ms/step\n","Epoch 87/100\n","118/118 - 1s - loss: 0.3775 - accuracy: 0.8802 - val_loss: 0.4818 - val_accuracy: 0.8451 - 1s/epoch - 11ms/step\n","Epoch 88/100\n","118/118 - 1s - loss: 0.3770 - accuracy: 0.8801 - val_loss: 0.4816 - val_accuracy: 0.8450 - 1s/epoch - 11ms/step\n","Epoch 89/100\n","118/118 - 1s - loss: 0.3765 - accuracy: 0.8802 - val_loss: 0.4813 - val_accuracy: 0.8453 - 1s/epoch - 11ms/step\n","Epoch 90/100\n","118/118 - 1s - loss: 0.3760 - accuracy: 0.8802 - val_loss: 0.4812 - val_accuracy: 0.8449 - 1s/epoch - 11ms/step\n","Epoch 91/100\n","118/118 - 1s - loss: 0.3755 - accuracy: 0.8803 - val_loss: 0.4809 - val_accuracy: 0.8450 - 1s/epoch - 12ms/step\n","Epoch 92/100\n","118/118 - 1s - loss: 0.3750 - accuracy: 0.8803 - val_loss: 0.4807 - val_accuracy: 0.8453 - 1s/epoch - 12ms/step\n","Epoch 93/100\n","118/118 - 1s - loss: 0.3745 - accuracy: 0.8806 - val_loss: 0.4805 - val_accuracy: 0.8448 - 1s/epoch - 12ms/step\n","Epoch 94/100\n","118/118 - 1s - loss: 0.3741 - accuracy: 0.8804 - val_loss: 0.4803 - val_accuracy: 0.8449 - 1s/epoch - 12ms/step\n","Epoch 95/100\n","118/118 - 1s - loss: 0.3737 - accuracy: 0.8805 - val_loss: 0.4801 - val_accuracy: 0.8452 - 1s/epoch - 12ms/step\n","Epoch 96/100\n","118/118 - 2s - loss: 0.3732 - accuracy: 0.8805 - val_loss: 0.4799 - val_accuracy: 0.8454 - 2s/epoch - 13ms/step\n","Epoch 97/100\n","118/118 - 2s - loss: 0.3728 - accuracy: 0.8806 - val_loss: 0.4798 - val_accuracy: 0.8451 - 2s/epoch - 13ms/step\n","Epoch 98/100\n","118/118 - 2s - loss: 0.3724 - accuracy: 0.8803 - val_loss: 0.4796 - val_accuracy: 0.8457 - 2s/epoch - 13ms/step\n","Epoch 99/100\n","118/118 - 2s - loss: 0.3720 - accuracy: 0.8804 - val_loss: 0.4794 - val_accuracy: 0.8453 - 2s/epoch - 13ms/step\n","Epoch 100/100\n","118/118 - 1s - loss: 0.3717 - accuracy: 0.8804 - val_loss: 0.4794 - val_accuracy: 0.8457 - 1s/epoch - 12ms/step\n"]}],"source":["# fix first, second and third, and train fourth\n","#################### YOUR CODE HERE ####################\n","# Các bạn thực hiện tương tự như Ví dụ train lần 1 \n","# Train lần 6:  Ghép sub mode 1, 2, 3 và 4 lại với nhau. Train với 100 epoch nhưng fix weight của sub model 1, sub model 2 và sub model 3\n","# (tạo model mới và thêm Input layer, sub model 1, sub model 2, sub model 3, sub model 4 , Output layer)\n","fourth_model = keras.Sequential()\n","fourth_model.add(keras.Input(shape=(784,)))\n","fourth_model.add(first)\n","fourth_model.add(second)\n","fourth_model.add(third)\n","fourth_model.add(fourth)\n","fourth_model.add(keras.layers.Dense(10))\n","\n","# freeze sub model 1 ,sub model 2 và sub model 3\n","first.trainable = False\n","second.trainable = False\n","third.trainable = False\n","########################################################\n","\n","\n","fourth_model.compile(optimizer=tf.keras.optimizers.SGD(), \n","              loss=keras.losses.SparseCategoricalCrossentropy(from_logits=True), \n","              metrics=['accuracy'])\n","\n","history = fourth_model.fit(X_train, y_train, batch_size,\n","                          validation_data=(X_test, y_test),\n","                          epochs=100, verbose=2)"]},{"cell_type":"code","execution_count":19,"id":"TGcRSptrLYAa","metadata":{"id":"TGcRSptrLYAa","executionInfo":{"status":"ok","timestamp":1669969153446,"user_tz":-420,"elapsed":24,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"}}},"outputs":[],"source":[]},{"cell_type":"code","execution_count":20,"id":"wq1CBjg2LYVi","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":296},"executionInfo":{"elapsed":23,"status":"ok","timestamp":1669969153447,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"},"user_tz":-420},"id":"wq1CBjg2LYVi","outputId":"93a6db4c-502c-4c0f-ac9b-b64ef78e1fd8"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.legend.Legend at 0x7f1349203070>"]},"metadata":{},"execution_count":20},{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 1 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}}],"source":["import matplotlib.pyplot as plt\n","\n","plt.plot(history.history['loss'], label='loss')\n","plt.plot(history.history['val_loss'], label='val_loss')\n","plt.xlabel('iteration')\n","plt.ylabel('Loss')\n","plt.legend()"]},{"cell_type":"code","execution_count":20,"id":"76GFFjk0LYVn","metadata":{"id":"76GFFjk0LYVn","executionInfo":{"status":"ok","timestamp":1669969153447,"user_tz":-420,"elapsed":19,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"}}},"outputs":[],"source":[]},{"cell_type":"code","execution_count":21,"id":"jVBQFFWwLYVq","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":297},"executionInfo":{"elapsed":651,"status":"ok","timestamp":1669969154079,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"},"user_tz":-420},"id":"jVBQFFWwLYVq","outputId":"8d72d11f-b422-4858-e02b-74cf499cb3e3"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.legend.Legend at 0x7f13491d4a60>"]},"metadata":{},"execution_count":21},{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 1 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}}],"source":["import matplotlib.pyplot as plt\n","\n","plt.plot(history.history['accuracy'], label='train_accuracy')\n","plt.plot(history.history['val_accuracy'], label='val_accuracy')\n","plt.xlabel('iteration')\n","plt.ylabel('Accuracy')\n","plt.legend()"]},{"cell_type":"markdown","source":["### Train 7"],"metadata":{"id":"fa1b86ec"},"id":"fa1b86ec"},{"cell_type":"code","execution_count":22,"id":"1f0f8cea","metadata":{"id":"1f0f8cea","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1669969897165,"user_tz":-420,"elapsed":743091,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"}},"outputId":"8602235b-917a-4886-b486-6d6bd7b6ccc5"},"outputs":[{"output_type":"stream","name":"stdout","text":["Epoch 1/300\n","118/118 - 4s - loss: 2.0101 - accuracy: 0.3737 - val_loss: 1.7157 - val_accuracy: 0.5890 - 4s/epoch - 32ms/step\n","Epoch 2/300\n","118/118 - 2s - loss: 1.5401 - accuracy: 0.6474 - val_loss: 1.4069 - val_accuracy: 0.6830 - 2s/epoch - 20ms/step\n","Epoch 3/300\n","118/118 - 2s - loss: 1.2824 - accuracy: 0.7506 - val_loss: 1.2037 - val_accuracy: 0.7683 - 2s/epoch - 19ms/step\n","Epoch 4/300\n","118/118 - 2s - loss: 1.1011 - accuracy: 0.8258 - val_loss: 1.0557 - val_accuracy: 0.8115 - 2s/epoch - 19ms/step\n","Epoch 5/300\n","118/118 - 2s - loss: 0.9661 - accuracy: 0.8501 - val_loss: 0.9462 - val_accuracy: 0.8237 - 2s/epoch - 20ms/step\n","Epoch 6/300\n","118/118 - 2s - loss: 0.8638 - accuracy: 0.8612 - val_loss: 0.8645 - val_accuracy: 0.8270 - 2s/epoch - 19ms/step\n","Epoch 7/300\n","118/118 - 2s - loss: 0.7851 - accuracy: 0.8663 - val_loss: 0.7994 - val_accuracy: 0.8302 - 2s/epoch - 19ms/step\n","Epoch 8/300\n","118/118 - 2s - loss: 0.7221 - accuracy: 0.8698 - val_loss: 0.7486 - val_accuracy: 0.8326 - 2s/epoch - 19ms/step\n","Epoch 9/300\n","118/118 - 2s - loss: 0.6725 - accuracy: 0.8740 - val_loss: 0.7094 - val_accuracy: 0.8376 - 2s/epoch - 19ms/step\n","Epoch 10/300\n","118/118 - 2s - loss: 0.6313 - accuracy: 0.8772 - val_loss: 0.6773 - val_accuracy: 0.8371 - 2s/epoch - 20ms/step\n","Epoch 11/300\n","118/118 - 2s - loss: 0.5987 - accuracy: 0.8779 - val_loss: 0.6584 - val_accuracy: 0.8337 - 2s/epoch - 19ms/step\n","Epoch 12/300\n","118/118 - 2s - loss: 0.5755 - accuracy: 0.8763 - val_loss: 0.6314 - val_accuracy: 0.8385 - 2s/epoch - 19ms/step\n","Epoch 13/300\n","118/118 - 2s - loss: 0.5491 - accuracy: 0.8803 - val_loss: 0.6196 - val_accuracy: 0.8350 - 2s/epoch - 20ms/step\n","Epoch 14/300\n","118/118 - 2s - loss: 0.5286 - accuracy: 0.8805 - val_loss: 0.5964 - val_accuracy: 0.8406 - 2s/epoch - 19ms/step\n","Epoch 15/300\n","118/118 - 2s - loss: 0.5109 - accuracy: 0.8817 - val_loss: 0.5865 - val_accuracy: 0.8385 - 2s/epoch - 20ms/step\n","Epoch 16/300\n","118/118 - 2s - loss: 0.4987 - accuracy: 0.8811 - val_loss: 0.5761 - val_accuracy: 0.8403 - 2s/epoch - 20ms/step\n","Epoch 17/300\n","118/118 - 2s - loss: 0.4825 - accuracy: 0.8825 - val_loss: 0.5652 - val_accuracy: 0.8401 - 2s/epoch - 19ms/step\n","Epoch 18/300\n","118/118 - 2s - loss: 0.4735 - accuracy: 0.8816 - val_loss: 0.5600 - val_accuracy: 0.8359 - 2s/epoch - 19ms/step\n","Epoch 19/300\n","118/118 - 2s - loss: 0.4630 - accuracy: 0.8820 - val_loss: 0.5497 - val_accuracy: 0.8416 - 2s/epoch - 19ms/step\n","Epoch 20/300\n","118/118 - 2s - loss: 0.4523 - accuracy: 0.8838 - val_loss: 0.5421 - val_accuracy: 0.8434 - 2s/epoch - 19ms/step\n","Epoch 21/300\n","118/118 - 2s - loss: 0.4461 - accuracy: 0.8834 - val_loss: 0.5502 - val_accuracy: 0.8356 - 2s/epoch - 20ms/step\n","Epoch 22/300\n","118/118 - 2s - loss: 0.4458 - accuracy: 0.8807 - val_loss: 0.5341 - val_accuracy: 0.8408 - 2s/epoch - 20ms/step\n","Epoch 23/300\n","118/118 - 2s - loss: 0.4332 - accuracy: 0.8833 - val_loss: 0.5293 - val_accuracy: 0.8414 - 2s/epoch - 20ms/step\n","Epoch 24/300\n","118/118 - 2s - loss: 0.4276 - accuracy: 0.8833 - val_loss: 0.5242 - val_accuracy: 0.8437 - 2s/epoch - 19ms/step\n","Epoch 25/300\n","118/118 - 2s - loss: 0.4238 - accuracy: 0.8828 - val_loss: 0.5246 - val_accuracy: 0.8377 - 2s/epoch - 20ms/step\n","Epoch 26/300\n","118/118 - 2s - loss: 0.4168 - accuracy: 0.8827 - val_loss: 0.5235 - val_accuracy: 0.8386 - 2s/epoch - 20ms/step\n","Epoch 27/300\n","118/118 - 2s - loss: 0.4162 - accuracy: 0.8826 - val_loss: 0.5486 - val_accuracy: 0.8281 - 2s/epoch - 20ms/step\n","Epoch 28/300\n","118/118 - 2s - loss: 0.4139 - accuracy: 0.8819 - val_loss: 0.5117 - val_accuracy: 0.8456 - 2s/epoch - 19ms/step\n","Epoch 29/300\n","118/118 - 2s - loss: 0.4068 - accuracy: 0.8828 - val_loss: 0.5168 - val_accuracy: 0.8431 - 2s/epoch - 19ms/step\n","Epoch 30/300\n","118/118 - 2s - loss: 0.4094 - accuracy: 0.8802 - val_loss: 0.5142 - val_accuracy: 0.8426 - 2s/epoch - 20ms/step\n","Epoch 31/300\n","118/118 - 2s - loss: 0.4000 - accuracy: 0.8841 - val_loss: 0.5091 - val_accuracy: 0.8380 - 2s/epoch - 19ms/step\n","Epoch 32/300\n","118/118 - 2s - loss: 0.3938 - accuracy: 0.8852 - val_loss: 0.5119 - val_accuracy: 0.8432 - 2s/epoch - 18ms/step\n","Epoch 33/300\n","118/118 - 2s - loss: 0.3950 - accuracy: 0.8837 - val_loss: 0.5078 - val_accuracy: 0.8384 - 2s/epoch - 19ms/step\n","Epoch 34/300\n","118/118 - 2s - loss: 0.3970 - accuracy: 0.8824 - val_loss: 0.5130 - val_accuracy: 0.8400 - 2s/epoch - 19ms/step\n","Epoch 35/300\n","118/118 - 2s - loss: 0.4037 - accuracy: 0.8787 - val_loss: 0.5193 - val_accuracy: 0.8349 - 2s/epoch - 19ms/step\n","Epoch 36/300\n","118/118 - 2s - loss: 0.4035 - accuracy: 0.8778 - val_loss: 0.5080 - val_accuracy: 0.8390 - 2s/epoch - 19ms/step\n","Epoch 37/300\n","118/118 - 2s - loss: 0.4022 - accuracy: 0.8782 - val_loss: 0.5176 - val_accuracy: 0.8353 - 2s/epoch - 19ms/step\n","Epoch 38/300\n","118/118 - 2s - loss: 0.3991 - accuracy: 0.8783 - val_loss: 0.5162 - val_accuracy: 0.8386 - 2s/epoch - 20ms/step\n","Epoch 39/300\n","118/118 - 2s - loss: 0.3987 - accuracy: 0.8785 - val_loss: 0.4997 - val_accuracy: 0.8393 - 2s/epoch - 20ms/step\n","Epoch 40/300\n","118/118 - 2s - loss: 0.3909 - accuracy: 0.8795 - val_loss: 0.5284 - val_accuracy: 0.8264 - 2s/epoch - 20ms/step\n","Epoch 41/300\n","118/118 - 2s - loss: 0.3893 - accuracy: 0.8810 - val_loss: 0.5026 - val_accuracy: 0.8385 - 2s/epoch - 20ms/step\n","Epoch 42/300\n","118/118 - 2s - loss: 0.3867 - accuracy: 0.8806 - val_loss: 0.5205 - val_accuracy: 0.8275 - 2s/epoch - 19ms/step\n","Epoch 43/300\n","118/118 - 2s - loss: 0.3952 - accuracy: 0.8770 - val_loss: 0.4941 - val_accuracy: 0.8411 - 2s/epoch - 19ms/step\n","Epoch 44/300\n","118/118 - 2s - loss: 0.3909 - accuracy: 0.8781 - val_loss: 0.4944 - val_accuracy: 0.8426 - 2s/epoch - 18ms/step\n","Epoch 45/300\n","118/118 - 2s - loss: 0.3838 - accuracy: 0.8795 - val_loss: 0.4999 - val_accuracy: 0.8413 - 2s/epoch - 20ms/step\n","Epoch 46/300\n","118/118 - 2s - loss: 0.3924 - accuracy: 0.8760 - val_loss: 0.5196 - val_accuracy: 0.8280 - 2s/epoch - 20ms/step\n","Epoch 47/300\n","118/118 - 2s - loss: 0.4004 - accuracy: 0.8728 - val_loss: 0.5179 - val_accuracy: 0.8347 - 2s/epoch - 20ms/step\n","Epoch 48/300\n","118/118 - 2s - loss: 0.3950 - accuracy: 0.8731 - val_loss: 0.5019 - val_accuracy: 0.8371 - 2s/epoch - 19ms/step\n","Epoch 49/300\n","118/118 - 2s - loss: 0.3892 - accuracy: 0.8771 - val_loss: 0.5078 - val_accuracy: 0.8378 - 2s/epoch - 19ms/step\n","Epoch 50/300\n","118/118 - 2s - loss: 0.3958 - accuracy: 0.8744 - val_loss: 0.4982 - val_accuracy: 0.8390 - 2s/epoch - 20ms/step\n","Epoch 51/300\n","118/118 - 2s - loss: 0.3902 - accuracy: 0.8755 - val_loss: 0.5032 - val_accuracy: 0.8371 - 2s/epoch - 20ms/step\n","Epoch 52/300\n","118/118 - 2s - loss: 0.3837 - accuracy: 0.8776 - val_loss: 0.4928 - val_accuracy: 0.8470 - 2s/epoch - 19ms/step\n","Epoch 53/300\n","118/118 - 2s - loss: 0.3841 - accuracy: 0.8766 - val_loss: 0.5003 - val_accuracy: 0.8378 - 2s/epoch - 19ms/step\n","Epoch 54/300\n","118/118 - 2s - loss: 0.3844 - accuracy: 0.8769 - val_loss: 0.5067 - val_accuracy: 0.8310 - 2s/epoch - 19ms/step\n","Epoch 55/300\n","118/118 - 2s - loss: 0.3919 - accuracy: 0.8739 - val_loss: 0.4959 - val_accuracy: 0.8385 - 2s/epoch - 20ms/step\n","Epoch 56/300\n","118/118 - 2s - loss: 0.3932 - accuracy: 0.8734 - val_loss: 0.5119 - val_accuracy: 0.8359 - 2s/epoch - 19ms/step\n","Epoch 57/300\n","118/118 - 2s - loss: 0.3892 - accuracy: 0.8746 - val_loss: 0.5070 - val_accuracy: 0.8319 - 2s/epoch - 19ms/step\n","Epoch 58/300\n","118/118 - 2s - loss: 0.3885 - accuracy: 0.8737 - val_loss: 0.4954 - val_accuracy: 0.8389 - 2s/epoch - 19ms/step\n","Epoch 59/300\n","118/118 - 2s - loss: 0.3850 - accuracy: 0.8758 - val_loss: 0.4946 - val_accuracy: 0.8383 - 2s/epoch - 20ms/step\n","Epoch 60/300\n","118/118 - 2s - loss: 0.3850 - accuracy: 0.8749 - val_loss: 0.4989 - val_accuracy: 0.8339 - 2s/epoch - 20ms/step\n","Epoch 61/300\n","118/118 - 2s - loss: 0.3898 - accuracy: 0.8740 - val_loss: 0.4940 - val_accuracy: 0.8399 - 2s/epoch - 19ms/step\n","Epoch 62/300\n","118/118 - 2s - loss: 0.3874 - accuracy: 0.8732 - val_loss: 0.4970 - val_accuracy: 0.8391 - 2s/epoch - 19ms/step\n","Epoch 63/300\n","118/118 - 2s - loss: 0.3818 - accuracy: 0.8756 - val_loss: 0.5139 - val_accuracy: 0.8328 - 2s/epoch - 19ms/step\n","Epoch 64/300\n","118/118 - 2s - loss: 0.4003 - accuracy: 0.8683 - val_loss: 0.5256 - val_accuracy: 0.8244 - 2s/epoch - 20ms/step\n","Epoch 65/300\n","118/118 - 2s - loss: 0.3956 - accuracy: 0.8700 - val_loss: 0.5057 - val_accuracy: 0.8327 - 2s/epoch - 19ms/step\n","Epoch 66/300\n","118/118 - 2s - loss: 0.3951 - accuracy: 0.8698 - val_loss: 0.5006 - val_accuracy: 0.8369 - 2s/epoch - 20ms/step\n","Epoch 67/300\n","118/118 - 2s - loss: 0.3976 - accuracy: 0.8682 - val_loss: 0.4866 - val_accuracy: 0.8427 - 2s/epoch - 18ms/step\n","Epoch 68/300\n","118/118 - 2s - loss: 0.3895 - accuracy: 0.8713 - val_loss: 0.5023 - val_accuracy: 0.8360 - 2s/epoch - 18ms/step\n","Epoch 69/300\n","118/118 - 2s - loss: 0.3931 - accuracy: 0.8703 - val_loss: 0.4917 - val_accuracy: 0.8368 - 2s/epoch - 18ms/step\n","Epoch 70/300\n","118/118 - 2s - loss: 0.3826 - accuracy: 0.8738 - val_loss: 0.5059 - val_accuracy: 0.8329 - 2s/epoch - 19ms/step\n","Epoch 71/300\n","118/118 - 2s - loss: 0.3889 - accuracy: 0.8719 - val_loss: 0.4962 - val_accuracy: 0.8358 - 2s/epoch - 19ms/step\n","Epoch 72/300\n","118/118 - 2s - loss: 0.3848 - accuracy: 0.8735 - val_loss: 0.5063 - val_accuracy: 0.8308 - 2s/epoch - 19ms/step\n","Epoch 73/300\n","118/118 - 2s - loss: 0.3890 - accuracy: 0.8716 - val_loss: 0.4873 - val_accuracy: 0.8406 - 2s/epoch - 19ms/step\n","Epoch 74/300\n","118/118 - 2s - loss: 0.3862 - accuracy: 0.8729 - val_loss: 0.5088 - val_accuracy: 0.8331 - 2s/epoch - 20ms/step\n","Epoch 75/300\n","118/118 - 2s - loss: 0.4073 - accuracy: 0.8653 - val_loss: 0.5216 - val_accuracy: 0.8225 - 2s/epoch - 20ms/step\n","Epoch 76/300\n","118/118 - 2s - loss: 0.3997 - accuracy: 0.8676 - val_loss: 0.5014 - val_accuracy: 0.8375 - 2s/epoch - 20ms/step\n","Epoch 77/300\n","118/118 - 2s - loss: 0.4029 - accuracy: 0.8654 - val_loss: 0.5023 - val_accuracy: 0.8342 - 2s/epoch - 19ms/step\n","Epoch 78/300\n","118/118 - 2s - loss: 0.4056 - accuracy: 0.8651 - val_loss: 0.5123 - val_accuracy: 0.8317 - 2s/epoch - 19ms/step\n","Epoch 79/300\n","118/118 - 2s - loss: 0.3984 - accuracy: 0.8679 - val_loss: 0.5018 - val_accuracy: 0.8361 - 2s/epoch - 19ms/step\n","Epoch 80/300\n","118/118 - 2s - loss: 0.3923 - accuracy: 0.8695 - val_loss: 0.5429 - val_accuracy: 0.8260 - 2s/epoch - 19ms/step\n","Epoch 81/300\n","118/118 - 2s - loss: 0.3916 - accuracy: 0.8698 - val_loss: 0.5124 - val_accuracy: 0.8307 - 2s/epoch - 19ms/step\n","Epoch 82/300\n","118/118 - 2s - loss: 0.3926 - accuracy: 0.8686 - val_loss: 0.5132 - val_accuracy: 0.8311 - 2s/epoch - 20ms/step\n","Epoch 83/300\n","118/118 - 2s - loss: 0.3936 - accuracy: 0.8687 - val_loss: 0.5132 - val_accuracy: 0.8286 - 2s/epoch - 20ms/step\n","Epoch 84/300\n","118/118 - 2s - loss: 0.3909 - accuracy: 0.8685 - val_loss: 0.4958 - val_accuracy: 0.8345 - 2s/epoch - 20ms/step\n","Epoch 85/300\n","118/118 - 2s - loss: 0.3854 - accuracy: 0.8709 - val_loss: 0.4873 - val_accuracy: 0.8370 - 2s/epoch - 19ms/step\n","Epoch 86/300\n","118/118 - 2s - loss: 0.3826 - accuracy: 0.8711 - val_loss: 0.5153 - val_accuracy: 0.8294 - 2s/epoch - 19ms/step\n","Epoch 87/300\n","118/118 - 2s - loss: 0.3856 - accuracy: 0.8718 - val_loss: 0.4982 - val_accuracy: 0.8398 - 2s/epoch - 19ms/step\n","Epoch 88/300\n","118/118 - 2s - loss: 0.3903 - accuracy: 0.8691 - val_loss: 0.4923 - val_accuracy: 0.8372 - 2s/epoch - 19ms/step\n","Epoch 89/300\n","118/118 - 2s - loss: 0.3863 - accuracy: 0.8701 - val_loss: 0.4855 - val_accuracy: 0.8369 - 2s/epoch - 20ms/step\n","Epoch 90/300\n","118/118 - 2s - loss: 0.3774 - accuracy: 0.8735 - val_loss: 0.4987 - val_accuracy: 0.8400 - 2s/epoch - 19ms/step\n","Epoch 91/300\n","118/118 - 2s - loss: 0.3803 - accuracy: 0.8723 - val_loss: 0.4752 - val_accuracy: 0.8431 - 2s/epoch - 19ms/step\n","Epoch 92/300\n","118/118 - 2s - loss: 0.3766 - accuracy: 0.8756 - val_loss: 0.4896 - val_accuracy: 0.8340 - 2s/epoch - 19ms/step\n","Epoch 93/300\n","118/118 - 2s - loss: 0.3800 - accuracy: 0.8707 - val_loss: 0.4947 - val_accuracy: 0.8355 - 2s/epoch - 20ms/step\n","Epoch 94/300\n","118/118 - 2s - loss: 0.3806 - accuracy: 0.8725 - val_loss: 0.4880 - val_accuracy: 0.8374 - 2s/epoch - 19ms/step\n","Epoch 95/300\n","118/118 - 2s - loss: 0.3757 - accuracy: 0.8741 - val_loss: 0.4833 - val_accuracy: 0.8389 - 2s/epoch - 19ms/step\n","Epoch 96/300\n","118/118 - 2s - loss: 0.3730 - accuracy: 0.8760 - val_loss: 0.5030 - val_accuracy: 0.8341 - 2s/epoch - 19ms/step\n","Epoch 97/300\n","118/118 - 3s - loss: 0.3777 - accuracy: 0.8723 - val_loss: 0.4948 - val_accuracy: 0.8370 - 3s/epoch - 24ms/step\n","Epoch 98/300\n","118/118 - 3s - loss: 0.3837 - accuracy: 0.8718 - val_loss: 0.4867 - val_accuracy: 0.8412 - 3s/epoch - 28ms/step\n","Epoch 99/300\n","118/118 - 2s - loss: 0.3731 - accuracy: 0.8753 - val_loss: 0.4909 - val_accuracy: 0.8399 - 2s/epoch - 19ms/step\n","Epoch 100/300\n","118/118 - 2s - loss: 0.3718 - accuracy: 0.8765 - val_loss: 0.4968 - val_accuracy: 0.8309 - 2s/epoch - 19ms/step\n","Epoch 101/300\n","118/118 - 2s - loss: 0.3744 - accuracy: 0.8734 - val_loss: 0.4790 - val_accuracy: 0.8432 - 2s/epoch - 19ms/step\n","Epoch 102/300\n","118/118 - 2s - loss: 0.3782 - accuracy: 0.8727 - val_loss: 0.4967 - val_accuracy: 0.8347 - 2s/epoch - 20ms/step\n","Epoch 103/300\n","118/118 - 2s - loss: 0.3749 - accuracy: 0.8737 - val_loss: 0.4837 - val_accuracy: 0.8416 - 2s/epoch - 20ms/step\n","Epoch 104/300\n","118/118 - 2s - loss: 0.3742 - accuracy: 0.8739 - val_loss: 0.4864 - val_accuracy: 0.8383 - 2s/epoch - 20ms/step\n","Epoch 105/300\n","118/118 - 2s - loss: 0.3774 - accuracy: 0.8719 - val_loss: 0.4988 - val_accuracy: 0.8361 - 2s/epoch - 20ms/step\n","Epoch 106/300\n","118/118 - 2s - loss: 0.3753 - accuracy: 0.8738 - val_loss: 0.4890 - val_accuracy: 0.8357 - 2s/epoch - 20ms/step\n","Epoch 107/300\n","118/118 - 2s - loss: 0.3728 - accuracy: 0.8747 - val_loss: 0.4890 - val_accuracy: 0.8356 - 2s/epoch - 19ms/step\n","Epoch 108/300\n","118/118 - 2s - loss: 0.3884 - accuracy: 0.8679 - val_loss: 0.5010 - val_accuracy: 0.8334 - 2s/epoch - 20ms/step\n","Epoch 109/300\n","118/118 - 2s - loss: 0.3733 - accuracy: 0.8754 - val_loss: 0.4884 - val_accuracy: 0.8367 - 2s/epoch - 20ms/step\n","Epoch 110/300\n","118/118 - 2s - loss: 0.3769 - accuracy: 0.8727 - val_loss: 0.5091 - val_accuracy: 0.8292 - 2s/epoch - 19ms/step\n","Epoch 111/300\n","118/118 - 2s - loss: 0.3814 - accuracy: 0.8711 - val_loss: 0.5106 - val_accuracy: 0.8291 - 2s/epoch - 19ms/step\n","Epoch 112/300\n","118/118 - 2s - loss: 0.3853 - accuracy: 0.8701 - val_loss: 0.4867 - val_accuracy: 0.8342 - 2s/epoch - 19ms/step\n","Epoch 113/300\n","118/118 - 2s - loss: 0.3748 - accuracy: 0.8727 - val_loss: 0.5108 - val_accuracy: 0.8289 - 2s/epoch - 19ms/step\n","Epoch 114/300\n","118/118 - 2s - loss: 0.3812 - accuracy: 0.8705 - val_loss: 0.5115 - val_accuracy: 0.8307 - 2s/epoch - 20ms/step\n","Epoch 115/300\n","118/118 - 2s - loss: 0.3851 - accuracy: 0.8692 - val_loss: 0.4878 - val_accuracy: 0.8334 - 2s/epoch - 20ms/step\n","Epoch 116/300\n","118/118 - 2s - loss: 0.3746 - accuracy: 0.8738 - val_loss: 0.5156 - val_accuracy: 0.8204 - 2s/epoch - 20ms/step\n","Epoch 117/300\n","118/118 - 2s - loss: 0.3833 - accuracy: 0.8685 - val_loss: 0.5361 - val_accuracy: 0.8177 - 2s/epoch - 19ms/step\n","Epoch 118/300\n","118/118 - 2s - loss: 0.3870 - accuracy: 0.8687 - val_loss: 0.5209 - val_accuracy: 0.8241 - 2s/epoch - 19ms/step\n","Epoch 119/300\n","118/118 - 2s - loss: 0.3939 - accuracy: 0.8652 - val_loss: 0.4937 - val_accuracy: 0.8370 - 2s/epoch - 19ms/step\n","Epoch 120/300\n","118/118 - 2s - loss: 0.3972 - accuracy: 0.8637 - val_loss: 0.4959 - val_accuracy: 0.8358 - 2s/epoch - 19ms/step\n","Epoch 121/300\n","118/118 - 2s - loss: 0.3921 - accuracy: 0.8674 - val_loss: 0.5039 - val_accuracy: 0.8314 - 2s/epoch - 19ms/step\n","Epoch 122/300\n","118/118 - 2s - loss: 0.3906 - accuracy: 0.8675 - val_loss: 0.5022 - val_accuracy: 0.8302 - 2s/epoch - 19ms/step\n","Epoch 123/300\n","118/118 - 2s - loss: 0.3905 - accuracy: 0.8664 - val_loss: 0.4976 - val_accuracy: 0.8345 - 2s/epoch - 19ms/step\n","Epoch 124/300\n","118/118 - 2s - loss: 0.3965 - accuracy: 0.8653 - val_loss: 0.5082 - val_accuracy: 0.8272 - 2s/epoch - 19ms/step\n","Epoch 125/300\n","118/118 - 2s - loss: 0.3864 - accuracy: 0.8680 - val_loss: 0.5061 - val_accuracy: 0.8298 - 2s/epoch - 20ms/step\n","Epoch 126/300\n","118/118 - 2s - loss: 0.3897 - accuracy: 0.8674 - val_loss: 0.5704 - val_accuracy: 0.7953 - 2s/epoch - 20ms/step\n","Epoch 127/300\n","118/118 - 2s - loss: 0.3953 - accuracy: 0.8658 - val_loss: 0.5077 - val_accuracy: 0.8332 - 2s/epoch - 19ms/step\n","Epoch 128/300\n","118/118 - 2s - loss: 0.3779 - accuracy: 0.8713 - val_loss: 0.5217 - val_accuracy: 0.8218 - 2s/epoch - 19ms/step\n","Epoch 129/300\n","118/118 - 2s - loss: 0.3953 - accuracy: 0.8641 - val_loss: 0.5128 - val_accuracy: 0.8312 - 2s/epoch - 19ms/step\n","Epoch 130/300\n","118/118 - 2s - loss: 0.3793 - accuracy: 0.8694 - val_loss: 0.4977 - val_accuracy: 0.8317 - 2s/epoch - 20ms/step\n","Epoch 131/300\n","118/118 - 2s - loss: 0.3738 - accuracy: 0.8728 - val_loss: 0.4814 - val_accuracy: 0.8362 - 2s/epoch - 20ms/step\n","Epoch 132/300\n","118/118 - 2s - loss: 0.3732 - accuracy: 0.8738 - val_loss: 0.5180 - val_accuracy: 0.8267 - 2s/epoch - 19ms/step\n","Epoch 133/300\n","118/118 - 2s - loss: 0.3747 - accuracy: 0.8725 - val_loss: 0.5077 - val_accuracy: 0.8281 - 2s/epoch - 21ms/step\n","Epoch 134/300\n","118/118 - 2s - loss: 0.3847 - accuracy: 0.8673 - val_loss: 0.4942 - val_accuracy: 0.8376 - 2s/epoch - 20ms/step\n","Epoch 135/300\n","118/118 - 2s - loss: 0.3813 - accuracy: 0.8711 - val_loss: 0.4831 - val_accuracy: 0.8382 - 2s/epoch - 20ms/step\n","Epoch 136/300\n","118/118 - 2s - loss: 0.3818 - accuracy: 0.8692 - val_loss: 0.4832 - val_accuracy: 0.8399 - 2s/epoch - 19ms/step\n","Epoch 137/300\n","118/118 - 2s - loss: 0.3748 - accuracy: 0.8731 - val_loss: 0.4976 - val_accuracy: 0.8335 - 2s/epoch - 20ms/step\n","Epoch 138/300\n","118/118 - 2s - loss: 0.3836 - accuracy: 0.8707 - val_loss: 0.5119 - val_accuracy: 0.8322 - 2s/epoch - 20ms/step\n","Epoch 139/300\n","118/118 - 2s - loss: 0.3854 - accuracy: 0.8669 - val_loss: 0.5169 - val_accuracy: 0.8176 - 2s/epoch - 19ms/step\n","Epoch 140/300\n","118/118 - 2s - loss: 0.3868 - accuracy: 0.8658 - val_loss: 0.4991 - val_accuracy: 0.8318 - 2s/epoch - 19ms/step\n","Epoch 141/300\n","118/118 - 2s - loss: 0.3872 - accuracy: 0.8680 - val_loss: 0.4897 - val_accuracy: 0.8415 - 2s/epoch - 19ms/step\n","Epoch 142/300\n","118/118 - 2s - loss: 0.3832 - accuracy: 0.8705 - val_loss: 0.4981 - val_accuracy: 0.8317 - 2s/epoch - 20ms/step\n","Epoch 143/300\n","118/118 - 2s - loss: 0.3786 - accuracy: 0.8712 - val_loss: 0.4987 - val_accuracy: 0.8356 - 2s/epoch - 19ms/step\n","Epoch 144/300\n","118/118 - 2s - loss: 0.3773 - accuracy: 0.8723 - val_loss: 0.4926 - val_accuracy: 0.8335 - 2s/epoch - 19ms/step\n","Epoch 145/300\n","118/118 - 2s - loss: 0.3925 - accuracy: 0.8660 - val_loss: 0.5157 - val_accuracy: 0.8270 - 2s/epoch - 19ms/step\n","Epoch 146/300\n","118/118 - 2s - loss: 0.3829 - accuracy: 0.8694 - val_loss: 0.4864 - val_accuracy: 0.8391 - 2s/epoch - 19ms/step\n","Epoch 147/300\n","118/118 - 2s - loss: 0.3746 - accuracy: 0.8747 - val_loss: 0.4807 - val_accuracy: 0.8373 - 2s/epoch - 19ms/step\n","Epoch 148/300\n","118/118 - 2s - loss: 0.3803 - accuracy: 0.8699 - val_loss: 0.5078 - val_accuracy: 0.8293 - 2s/epoch - 19ms/step\n","Epoch 149/300\n","118/118 - 2s - loss: 0.3804 - accuracy: 0.8702 - val_loss: 0.4996 - val_accuracy: 0.8285 - 2s/epoch - 19ms/step\n","Epoch 150/300\n","118/118 - 2s - loss: 0.3852 - accuracy: 0.8684 - val_loss: 0.5097 - val_accuracy: 0.8282 - 2s/epoch - 19ms/step\n","Epoch 151/300\n","118/118 - 2s - loss: 0.3879 - accuracy: 0.8669 - val_loss: 0.4918 - val_accuracy: 0.8354 - 2s/epoch - 19ms/step\n","Epoch 152/300\n","118/118 - 2s - loss: 0.3837 - accuracy: 0.8686 - val_loss: 0.4928 - val_accuracy: 0.8324 - 2s/epoch - 19ms/step\n","Epoch 153/300\n","118/118 - 2s - loss: 0.3816 - accuracy: 0.8691 - val_loss: 0.5437 - val_accuracy: 0.8036 - 2s/epoch - 20ms/step\n","Epoch 154/300\n","118/118 - 2s - loss: 0.3823 - accuracy: 0.8684 - val_loss: 0.5055 - val_accuracy: 0.8294 - 2s/epoch - 19ms/step\n","Epoch 155/300\n","118/118 - 2s - loss: 0.3925 - accuracy: 0.8657 - val_loss: 0.5051 - val_accuracy: 0.8327 - 2s/epoch - 19ms/step\n","Epoch 156/300\n","118/118 - 2s - loss: 0.4001 - accuracy: 0.8621 - val_loss: 0.5396 - val_accuracy: 0.8129 - 2s/epoch - 20ms/step\n","Epoch 157/300\n","118/118 - 2s - loss: 0.3947 - accuracy: 0.8648 - val_loss: 0.5638 - val_accuracy: 0.8057 - 2s/epoch - 19ms/step\n","Epoch 158/300\n","118/118 - 2s - loss: 0.4133 - accuracy: 0.8578 - val_loss: 0.5265 - val_accuracy: 0.8295 - 2s/epoch - 19ms/step\n","Epoch 159/300\n","118/118 - 2s - loss: 0.4123 - accuracy: 0.8590 - val_loss: 0.5101 - val_accuracy: 0.8305 - 2s/epoch - 19ms/step\n","Epoch 160/300\n","118/118 - 2s - loss: 0.4026 - accuracy: 0.8628 - val_loss: 0.5194 - val_accuracy: 0.8256 - 2s/epoch - 19ms/step\n","Epoch 161/300\n","118/118 - 2s - loss: 0.3960 - accuracy: 0.8640 - val_loss: 0.5015 - val_accuracy: 0.8331 - 2s/epoch - 19ms/step\n","Epoch 162/300\n","118/118 - 2s - loss: 0.3894 - accuracy: 0.8671 - val_loss: 0.5332 - val_accuracy: 0.8234 - 2s/epoch - 19ms/step\n","Epoch 163/300\n","118/118 - 2s - loss: 0.3940 - accuracy: 0.8651 - val_loss: 0.5518 - val_accuracy: 0.8048 - 2s/epoch - 19ms/step\n","Epoch 164/300\n","118/118 - 2s - loss: 0.3998 - accuracy: 0.8609 - val_loss: 0.5008 - val_accuracy: 0.8316 - 2s/epoch - 19ms/step\n","Epoch 165/300\n","118/118 - 2s - loss: 0.3911 - accuracy: 0.8648 - val_loss: 0.5327 - val_accuracy: 0.8283 - 2s/epoch - 19ms/step\n","Epoch 166/300\n","118/118 - 2s - loss: 0.3954 - accuracy: 0.8626 - val_loss: 0.4926 - val_accuracy: 0.8376 - 2s/epoch - 19ms/step\n","Epoch 167/300\n","118/118 - 2s - loss: 0.3935 - accuracy: 0.8633 - val_loss: 0.4895 - val_accuracy: 0.8376 - 2s/epoch - 20ms/step\n","Epoch 168/300\n","118/118 - 2s - loss: 0.3810 - accuracy: 0.8701 - val_loss: 0.5086 - val_accuracy: 0.8320 - 2s/epoch - 19ms/step\n","Epoch 169/300\n","118/118 - 2s - loss: 0.3823 - accuracy: 0.8687 - val_loss: 0.5024 - val_accuracy: 0.8321 - 2s/epoch - 20ms/step\n","Epoch 170/300\n","118/118 - 2s - loss: 0.3836 - accuracy: 0.8691 - val_loss: 0.5105 - val_accuracy: 0.8295 - 2s/epoch - 20ms/step\n","Epoch 171/300\n","118/118 - 2s - loss: 0.3802 - accuracy: 0.8704 - val_loss: 0.5446 - val_accuracy: 0.8222 - 2s/epoch - 20ms/step\n","Epoch 172/300\n","118/118 - 2s - loss: 0.4118 - accuracy: 0.8582 - val_loss: 0.5417 - val_accuracy: 0.8114 - 2s/epoch - 20ms/step\n","Epoch 173/300\n","118/118 - 2s - loss: 0.4056 - accuracy: 0.8613 - val_loss: 0.5170 - val_accuracy: 0.8342 - 2s/epoch - 20ms/step\n","Epoch 174/300\n","118/118 - 2s - loss: 0.4128 - accuracy: 0.8601 - val_loss: 0.5513 - val_accuracy: 0.8213 - 2s/epoch - 19ms/step\n","Epoch 175/300\n","118/118 - 2s - loss: 0.4084 - accuracy: 0.8608 - val_loss: 0.5330 - val_accuracy: 0.8199 - 2s/epoch - 18ms/step\n","Epoch 176/300\n","118/118 - 2s - loss: 0.4054 - accuracy: 0.8606 - val_loss: 0.4992 - val_accuracy: 0.8378 - 2s/epoch - 19ms/step\n","Epoch 177/300\n","118/118 - 2s - loss: 0.3946 - accuracy: 0.8659 - val_loss: 0.5104 - val_accuracy: 0.8313 - 2s/epoch - 21ms/step\n","Epoch 178/300\n","118/118 - 2s - loss: 0.4059 - accuracy: 0.8598 - val_loss: 0.5250 - val_accuracy: 0.8270 - 2s/epoch - 20ms/step\n","Epoch 179/300\n","118/118 - 2s - loss: 0.3949 - accuracy: 0.8656 - val_loss: 0.5047 - val_accuracy: 0.8313 - 2s/epoch - 21ms/step\n","Epoch 180/300\n","118/118 - 2s - loss: 0.3909 - accuracy: 0.8651 - val_loss: 0.4985 - val_accuracy: 0.8365 - 2s/epoch - 20ms/step\n","Epoch 181/300\n","118/118 - 2s - loss: 0.3873 - accuracy: 0.8670 - val_loss: 0.4983 - val_accuracy: 0.8349 - 2s/epoch - 20ms/step\n","Epoch 182/300\n","118/118 - 2s - loss: 0.3954 - accuracy: 0.8644 - val_loss: 0.5142 - val_accuracy: 0.8247 - 2s/epoch - 20ms/step\n","Epoch 183/300\n","118/118 - 2s - loss: 0.3944 - accuracy: 0.8645 - val_loss: 0.5028 - val_accuracy: 0.8339 - 2s/epoch - 20ms/step\n","Epoch 184/300\n","118/118 - 2s - loss: 0.3897 - accuracy: 0.8670 - val_loss: 0.5178 - val_accuracy: 0.8259 - 2s/epoch - 19ms/step\n","Epoch 185/300\n","118/118 - 2s - loss: 0.3936 - accuracy: 0.8641 - val_loss: 0.5450 - val_accuracy: 0.8069 - 2s/epoch - 20ms/step\n","Epoch 186/300\n","118/118 - 2s - loss: 0.3903 - accuracy: 0.8653 - val_loss: 0.4928 - val_accuracy: 0.8376 - 2s/epoch - 19ms/step\n","Epoch 187/300\n","118/118 - 2s - loss: 0.3793 - accuracy: 0.8704 - val_loss: 0.4892 - val_accuracy: 0.8389 - 2s/epoch - 20ms/step\n","Epoch 188/300\n","118/118 - 2s - loss: 0.3957 - accuracy: 0.8630 - val_loss: 0.5208 - val_accuracy: 0.8253 - 2s/epoch - 19ms/step\n","Epoch 189/300\n","118/118 - 2s - loss: 0.3901 - accuracy: 0.8654 - val_loss: 0.4976 - val_accuracy: 0.8314 - 2s/epoch - 19ms/step\n","Epoch 190/300\n","118/118 - 2s - loss: 0.3871 - accuracy: 0.8675 - val_loss: 0.5058 - val_accuracy: 0.8277 - 2s/epoch - 20ms/step\n","Epoch 191/300\n","118/118 - 2s - loss: 0.3942 - accuracy: 0.8644 - val_loss: 0.4955 - val_accuracy: 0.8340 - 2s/epoch - 19ms/step\n","Epoch 192/300\n","118/118 - 2s - loss: 0.3858 - accuracy: 0.8672 - val_loss: 0.4917 - val_accuracy: 0.8362 - 2s/epoch - 20ms/step\n","Epoch 193/300\n","118/118 - 2s - loss: 0.3889 - accuracy: 0.8671 - val_loss: 0.5009 - val_accuracy: 0.8386 - 2s/epoch - 19ms/step\n","Epoch 194/300\n","118/118 - 2s - loss: 0.3863 - accuracy: 0.8670 - val_loss: 0.4905 - val_accuracy: 0.8389 - 2s/epoch - 20ms/step\n","Epoch 195/300\n","118/118 - 2s - loss: 0.3734 - accuracy: 0.8722 - val_loss: 0.4898 - val_accuracy: 0.8366 - 2s/epoch - 20ms/step\n","Epoch 196/300\n","118/118 - 2s - loss: 0.3690 - accuracy: 0.8730 - val_loss: 0.4900 - val_accuracy: 0.8371 - 2s/epoch - 20ms/step\n","Epoch 197/300\n","118/118 - 2s - loss: 0.3774 - accuracy: 0.8691 - val_loss: 0.4892 - val_accuracy: 0.8383 - 2s/epoch - 20ms/step\n","Epoch 198/300\n","118/118 - 2s - loss: 0.3782 - accuracy: 0.8705 - val_loss: 0.4990 - val_accuracy: 0.8305 - 2s/epoch - 20ms/step\n","Epoch 199/300\n","118/118 - 2s - loss: 0.3806 - accuracy: 0.8695 - val_loss: 0.5045 - val_accuracy: 0.8307 - 2s/epoch - 19ms/step\n","Epoch 200/300\n","118/118 - 2s - loss: 0.3749 - accuracy: 0.8715 - val_loss: 0.4908 - val_accuracy: 0.8374 - 2s/epoch - 20ms/step\n","Epoch 201/300\n","118/118 - 2s - loss: 0.3785 - accuracy: 0.8706 - val_loss: 0.5138 - val_accuracy: 0.8231 - 2s/epoch - 20ms/step\n","Epoch 202/300\n","118/118 - 2s - loss: 0.3743 - accuracy: 0.8718 - val_loss: 0.4864 - val_accuracy: 0.8355 - 2s/epoch - 20ms/step\n","Epoch 203/300\n","118/118 - 2s - loss: 0.3750 - accuracy: 0.8717 - val_loss: 0.5044 - val_accuracy: 0.8311 - 2s/epoch - 20ms/step\n","Epoch 204/300\n","118/118 - 2s - loss: 0.3808 - accuracy: 0.8679 - val_loss: 0.4697 - val_accuracy: 0.8440 - 2s/epoch - 19ms/step\n","Epoch 205/300\n","118/118 - 2s - loss: 0.3695 - accuracy: 0.8730 - val_loss: 0.5025 - val_accuracy: 0.8309 - 2s/epoch - 20ms/step\n","Epoch 206/300\n","118/118 - 2s - loss: 0.3813 - accuracy: 0.8692 - val_loss: 0.5185 - val_accuracy: 0.8271 - 2s/epoch - 19ms/step\n","Epoch 207/300\n","118/118 - 2s - loss: 0.3821 - accuracy: 0.8687 - val_loss: 0.4869 - val_accuracy: 0.8357 - 2s/epoch - 19ms/step\n","Epoch 208/300\n","118/118 - 2s - loss: 0.3689 - accuracy: 0.8736 - val_loss: 0.5017 - val_accuracy: 0.8336 - 2s/epoch - 19ms/step\n","Epoch 209/300\n","118/118 - 2s - loss: 0.3813 - accuracy: 0.8674 - val_loss: 0.5091 - val_accuracy: 0.8272 - 2s/epoch - 19ms/step\n","Epoch 210/300\n","118/118 - 2s - loss: 0.3795 - accuracy: 0.8695 - val_loss: 0.4970 - val_accuracy: 0.8367 - 2s/epoch - 19ms/step\n","Epoch 211/300\n","118/118 - 2s - loss: 0.3783 - accuracy: 0.8705 - val_loss: 0.4920 - val_accuracy: 0.8331 - 2s/epoch - 19ms/step\n","Epoch 212/300\n","118/118 - 2s - loss: 0.3804 - accuracy: 0.8680 - val_loss: 0.5013 - val_accuracy: 0.8326 - 2s/epoch - 19ms/step\n","Epoch 213/300\n","118/118 - 2s - loss: 0.3709 - accuracy: 0.8733 - val_loss: 0.5040 - val_accuracy: 0.8270 - 2s/epoch - 19ms/step\n","Epoch 214/300\n","118/118 - 2s - loss: 0.3758 - accuracy: 0.8700 - val_loss: 0.4909 - val_accuracy: 0.8400 - 2s/epoch - 19ms/step\n","Epoch 215/300\n","118/118 - 2s - loss: 0.3709 - accuracy: 0.8719 - val_loss: 0.5020 - val_accuracy: 0.8277 - 2s/epoch - 19ms/step\n","Epoch 216/300\n","118/118 - 2s - loss: 0.3792 - accuracy: 0.8686 - val_loss: 0.5061 - val_accuracy: 0.8310 - 2s/epoch - 19ms/step\n","Epoch 217/300\n","118/118 - 2s - loss: 0.3739 - accuracy: 0.8707 - val_loss: 0.4986 - val_accuracy: 0.8320 - 2s/epoch - 19ms/step\n","Epoch 218/300\n","118/118 - 2s - loss: 0.3753 - accuracy: 0.8696 - val_loss: 0.5028 - val_accuracy: 0.8313 - 2s/epoch - 20ms/step\n","Epoch 219/300\n","118/118 - 2s - loss: 0.3823 - accuracy: 0.8668 - val_loss: 0.5108 - val_accuracy: 0.8285 - 2s/epoch - 19ms/step\n","Epoch 220/300\n","118/118 - 2s - loss: 0.3736 - accuracy: 0.8714 - val_loss: 0.4886 - val_accuracy: 0.8420 - 2s/epoch - 19ms/step\n","Epoch 221/300\n","118/118 - 2s - loss: 0.3685 - accuracy: 0.8727 - val_loss: 0.5043 - val_accuracy: 0.8310 - 2s/epoch - 19ms/step\n","Epoch 222/300\n","118/118 - 2s - loss: 0.3850 - accuracy: 0.8668 - val_loss: 0.5063 - val_accuracy: 0.8308 - 2s/epoch - 19ms/step\n","Epoch 223/300\n","118/118 - 2s - loss: 0.3766 - accuracy: 0.8704 - val_loss: 0.4958 - val_accuracy: 0.8253 - 2s/epoch - 19ms/step\n","Epoch 224/300\n","118/118 - 2s - loss: 0.3730 - accuracy: 0.8701 - val_loss: 0.5154 - val_accuracy: 0.8302 - 2s/epoch - 20ms/step\n","Epoch 225/300\n","118/118 - 2s - loss: 0.3889 - accuracy: 0.8645 - val_loss: 0.5109 - val_accuracy: 0.8290 - 2s/epoch - 20ms/step\n","Epoch 226/300\n","118/118 - 2s - loss: 0.3840 - accuracy: 0.8665 - val_loss: 0.5213 - val_accuracy: 0.8287 - 2s/epoch - 20ms/step\n","Epoch 227/300\n","118/118 - 2s - loss: 0.3993 - accuracy: 0.8630 - val_loss: 0.5002 - val_accuracy: 0.8334 - 2s/epoch - 19ms/step\n","Epoch 228/300\n","118/118 - 2s - loss: 0.3794 - accuracy: 0.8676 - val_loss: 0.5054 - val_accuracy: 0.8288 - 2s/epoch - 20ms/step\n","Epoch 229/300\n","118/118 - 2s - loss: 0.3867 - accuracy: 0.8630 - val_loss: 0.5020 - val_accuracy: 0.8324 - 2s/epoch - 19ms/step\n","Epoch 230/300\n","118/118 - 2s - loss: 0.3853 - accuracy: 0.8669 - val_loss: 0.5507 - val_accuracy: 0.8040 - 2s/epoch - 19ms/step\n","Epoch 231/300\n","118/118 - 2s - loss: 0.4004 - accuracy: 0.8610 - val_loss: 0.5091 - val_accuracy: 0.8354 - 2s/epoch - 19ms/step\n","Epoch 232/300\n","118/118 - 2s - loss: 0.3794 - accuracy: 0.8694 - val_loss: 0.5225 - val_accuracy: 0.8246 - 2s/epoch - 19ms/step\n","Epoch 233/300\n","118/118 - 2s - loss: 0.3783 - accuracy: 0.8680 - val_loss: 0.5278 - val_accuracy: 0.8165 - 2s/epoch - 19ms/step\n","Epoch 234/300\n","118/118 - 2s - loss: 0.3950 - accuracy: 0.8637 - val_loss: 0.5086 - val_accuracy: 0.8308 - 2s/epoch - 19ms/step\n","Epoch 235/300\n","118/118 - 2s - loss: 0.3828 - accuracy: 0.8648 - val_loss: 0.5098 - val_accuracy: 0.8247 - 2s/epoch - 19ms/step\n","Epoch 236/300\n","118/118 - 3s - loss: 0.3790 - accuracy: 0.8674 - val_loss: 0.5015 - val_accuracy: 0.8349 - 3s/epoch - 28ms/step\n","Epoch 237/300\n","118/118 - 3s - loss: 0.3822 - accuracy: 0.8652 - val_loss: 0.5026 - val_accuracy: 0.8282 - 3s/epoch - 22ms/step\n","Epoch 238/300\n","118/118 - 2s - loss: 0.3795 - accuracy: 0.8689 - val_loss: 0.5019 - val_accuracy: 0.8305 - 2s/epoch - 18ms/step\n","Epoch 239/300\n","118/118 - 2s - loss: 0.3798 - accuracy: 0.8681 - val_loss: 0.4940 - val_accuracy: 0.8323 - 2s/epoch - 18ms/step\n","Epoch 240/300\n","118/118 - 2s - loss: 0.3796 - accuracy: 0.8653 - val_loss: 0.5148 - val_accuracy: 0.8216 - 2s/epoch - 18ms/step\n","Epoch 241/300\n","118/118 - 2s - loss: 0.3773 - accuracy: 0.8684 - val_loss: 0.4942 - val_accuracy: 0.8318 - 2s/epoch - 18ms/step\n","Epoch 242/300\n","118/118 - 2s - loss: 0.3721 - accuracy: 0.8703 - val_loss: 0.5137 - val_accuracy: 0.8268 - 2s/epoch - 19ms/step\n","Epoch 243/300\n","118/118 - 2s - loss: 0.3917 - accuracy: 0.8636 - val_loss: 0.5248 - val_accuracy: 0.8181 - 2s/epoch - 19ms/step\n","Epoch 244/300\n","118/118 - 2s - loss: 0.3959 - accuracy: 0.8638 - val_loss: 0.5040 - val_accuracy: 0.8297 - 2s/epoch - 19ms/step\n","Epoch 245/300\n","118/118 - 2s - loss: 0.3855 - accuracy: 0.8668 - val_loss: 0.5050 - val_accuracy: 0.8324 - 2s/epoch - 19ms/step\n","Epoch 246/300\n","118/118 - 2s - loss: 0.3842 - accuracy: 0.8684 - val_loss: 0.5153 - val_accuracy: 0.8270 - 2s/epoch - 19ms/step\n","Epoch 247/300\n","118/118 - 2s - loss: 0.3824 - accuracy: 0.8673 - val_loss: 0.4971 - val_accuracy: 0.8352 - 2s/epoch - 19ms/step\n","Epoch 248/300\n","118/118 - 2s - loss: 0.3803 - accuracy: 0.8673 - val_loss: 0.5666 - val_accuracy: 0.8046 - 2s/epoch - 19ms/step\n","Epoch 249/300\n","118/118 - 2s - loss: 0.3927 - accuracy: 0.8611 - val_loss: 0.5168 - val_accuracy: 0.8200 - 2s/epoch - 18ms/step\n","Epoch 250/300\n","118/118 - 2s - loss: 0.3793 - accuracy: 0.8669 - val_loss: 0.4889 - val_accuracy: 0.8426 - 2s/epoch - 19ms/step\n","Epoch 251/300\n","118/118 - 2s - loss: 0.3760 - accuracy: 0.8688 - val_loss: 0.4988 - val_accuracy: 0.8330 - 2s/epoch - 19ms/step\n","Epoch 252/300\n","118/118 - 2s - loss: 0.3730 - accuracy: 0.8715 - val_loss: 0.5654 - val_accuracy: 0.8152 - 2s/epoch - 19ms/step\n","Epoch 253/300\n","118/118 - 2s - loss: 0.3974 - accuracy: 0.8610 - val_loss: 0.5072 - val_accuracy: 0.8286 - 2s/epoch - 19ms/step\n","Epoch 254/300\n","118/118 - 2s - loss: 0.3790 - accuracy: 0.8680 - val_loss: 0.4914 - val_accuracy: 0.8345 - 2s/epoch - 18ms/step\n","Epoch 255/300\n","118/118 - 2s - loss: 0.3721 - accuracy: 0.8705 - val_loss: 0.5180 - val_accuracy: 0.8204 - 2s/epoch - 18ms/step\n","Epoch 256/300\n","118/118 - 2s - loss: 0.3774 - accuracy: 0.8683 - val_loss: 0.4895 - val_accuracy: 0.8375 - 2s/epoch - 19ms/step\n","Epoch 257/300\n","118/118 - 2s - loss: 0.3933 - accuracy: 0.8641 - val_loss: 0.5063 - val_accuracy: 0.8303 - 2s/epoch - 19ms/step\n","Epoch 258/300\n","118/118 - 2s - loss: 0.3816 - accuracy: 0.8661 - val_loss: 0.4901 - val_accuracy: 0.8342 - 2s/epoch - 19ms/step\n","Epoch 259/300\n","118/118 - 2s - loss: 0.3793 - accuracy: 0.8681 - val_loss: 0.5794 - val_accuracy: 0.8033 - 2s/epoch - 18ms/step\n","Epoch 260/300\n","118/118 - 2s - loss: 0.3789 - accuracy: 0.8672 - val_loss: 0.5279 - val_accuracy: 0.8204 - 2s/epoch - 19ms/step\n","Epoch 261/300\n","118/118 - 2s - loss: 0.3877 - accuracy: 0.8642 - val_loss: 0.5070 - val_accuracy: 0.8344 - 2s/epoch - 19ms/step\n","Epoch 262/300\n","118/118 - 2s - loss: 0.3752 - accuracy: 0.8715 - val_loss: 0.4817 - val_accuracy: 0.8379 - 2s/epoch - 19ms/step\n","Epoch 263/300\n","118/118 - 2s - loss: 0.3763 - accuracy: 0.8690 - val_loss: 0.5119 - val_accuracy: 0.8276 - 2s/epoch - 19ms/step\n","Epoch 264/300\n","118/118 - 2s - loss: 0.3768 - accuracy: 0.8683 - val_loss: 0.5144 - val_accuracy: 0.8257 - 2s/epoch - 19ms/step\n","Epoch 265/300\n","118/118 - 2s - loss: 0.3758 - accuracy: 0.8689 - val_loss: 0.5202 - val_accuracy: 0.8201 - 2s/epoch - 19ms/step\n","Epoch 266/300\n","118/118 - 2s - loss: 0.3773 - accuracy: 0.8686 - val_loss: 0.5096 - val_accuracy: 0.8266 - 2s/epoch - 19ms/step\n","Epoch 267/300\n","118/118 - 2s - loss: 0.3803 - accuracy: 0.8685 - val_loss: 0.4990 - val_accuracy: 0.8335 - 2s/epoch - 19ms/step\n","Epoch 268/300\n","118/118 - 2s - loss: 0.3691 - accuracy: 0.8724 - val_loss: 0.4850 - val_accuracy: 0.8348 - 2s/epoch - 19ms/step\n","Epoch 269/300\n","118/118 - 2s - loss: 0.3708 - accuracy: 0.8699 - val_loss: 0.5248 - val_accuracy: 0.8075 - 2s/epoch - 19ms/step\n","Epoch 270/300\n","118/118 - 2s - loss: 0.3858 - accuracy: 0.8665 - val_loss: 0.4972 - val_accuracy: 0.8312 - 2s/epoch - 19ms/step\n","Epoch 271/300\n","118/118 - 2s - loss: 0.3759 - accuracy: 0.8684 - val_loss: 0.4942 - val_accuracy: 0.8312 - 2s/epoch - 18ms/step\n","Epoch 272/300\n","118/118 - 2s - loss: 0.3708 - accuracy: 0.8719 - val_loss: 0.4976 - val_accuracy: 0.8361 - 2s/epoch - 19ms/step\n","Epoch 273/300\n","118/118 - 2s - loss: 0.3665 - accuracy: 0.8729 - val_loss: 0.4800 - val_accuracy: 0.8416 - 2s/epoch - 19ms/step\n","Epoch 274/300\n","118/118 - 2s - loss: 0.3691 - accuracy: 0.8725 - val_loss: 0.4927 - val_accuracy: 0.8409 - 2s/epoch - 19ms/step\n","Epoch 275/300\n","118/118 - 2s - loss: 0.3789 - accuracy: 0.8657 - val_loss: 0.4938 - val_accuracy: 0.8333 - 2s/epoch - 19ms/step\n","Epoch 276/300\n","118/118 - 2s - loss: 0.3859 - accuracy: 0.8658 - val_loss: 0.5015 - val_accuracy: 0.8305 - 2s/epoch - 19ms/step\n","Epoch 277/300\n","118/118 - 2s - loss: 0.3763 - accuracy: 0.8677 - val_loss: 0.5033 - val_accuracy: 0.8282 - 2s/epoch - 20ms/step\n","Epoch 278/300\n","118/118 - 2s - loss: 0.3695 - accuracy: 0.8704 - val_loss: 0.5235 - val_accuracy: 0.8251 - 2s/epoch - 19ms/step\n","Epoch 279/300\n","118/118 - 2s - loss: 0.3894 - accuracy: 0.8652 - val_loss: 0.5071 - val_accuracy: 0.8318 - 2s/epoch - 19ms/step\n","Epoch 280/300\n","118/118 - 2s - loss: 0.3685 - accuracy: 0.8731 - val_loss: 0.5029 - val_accuracy: 0.8273 - 2s/epoch - 19ms/step\n","Epoch 281/300\n","118/118 - 2s - loss: 0.3680 - accuracy: 0.8717 - val_loss: 0.4943 - val_accuracy: 0.8325 - 2s/epoch - 19ms/step\n","Epoch 282/300\n","118/118 - 2s - loss: 0.3730 - accuracy: 0.8673 - val_loss: 0.5372 - val_accuracy: 0.8222 - 2s/epoch - 19ms/step\n","Epoch 283/300\n","118/118 - 2s - loss: 0.3685 - accuracy: 0.8706 - val_loss: 0.4911 - val_accuracy: 0.8361 - 2s/epoch - 19ms/step\n","Epoch 284/300\n","118/118 - 2s - loss: 0.3692 - accuracy: 0.8726 - val_loss: 0.5172 - val_accuracy: 0.8236 - 2s/epoch - 19ms/step\n","Epoch 285/300\n","118/118 - 2s - loss: 0.3659 - accuracy: 0.8714 - val_loss: 0.4770 - val_accuracy: 0.8369 - 2s/epoch - 19ms/step\n","Epoch 286/300\n","118/118 - 2s - loss: 0.3652 - accuracy: 0.8729 - val_loss: 0.4860 - val_accuracy: 0.8356 - 2s/epoch - 20ms/step\n","Epoch 287/300\n","118/118 - 2s - loss: 0.3584 - accuracy: 0.8767 - val_loss: 0.4764 - val_accuracy: 0.8434 - 2s/epoch - 19ms/step\n","Epoch 288/300\n","118/118 - 2s - loss: 0.3625 - accuracy: 0.8749 - val_loss: 0.4833 - val_accuracy: 0.8384 - 2s/epoch - 19ms/step\n","Epoch 289/300\n","118/118 - 2s - loss: 0.3647 - accuracy: 0.8715 - val_loss: 0.5201 - val_accuracy: 0.8249 - 2s/epoch - 19ms/step\n","Epoch 290/300\n","118/118 - 2s - loss: 0.3749 - accuracy: 0.8692 - val_loss: 0.4867 - val_accuracy: 0.8402 - 2s/epoch - 19ms/step\n","Epoch 291/300\n","118/118 - 2s - loss: 0.3609 - accuracy: 0.8758 - val_loss: 0.5051 - val_accuracy: 0.8300 - 2s/epoch - 19ms/step\n","Epoch 292/300\n","118/118 - 2s - loss: 0.3662 - accuracy: 0.8717 - val_loss: 0.4913 - val_accuracy: 0.8330 - 2s/epoch - 19ms/step\n","Epoch 293/300\n","118/118 - 2s - loss: 0.3665 - accuracy: 0.8691 - val_loss: 0.4780 - val_accuracy: 0.8432 - 2s/epoch - 19ms/step\n","Epoch 294/300\n","118/118 - 2s - loss: 0.3631 - accuracy: 0.8726 - val_loss: 0.4763 - val_accuracy: 0.8392 - 2s/epoch - 19ms/step\n","Epoch 295/300\n","118/118 - 2s - loss: 0.3545 - accuracy: 0.8760 - val_loss: 0.4967 - val_accuracy: 0.8360 - 2s/epoch - 19ms/step\n","Epoch 296/300\n","118/118 - 2s - loss: 0.3660 - accuracy: 0.8723 - val_loss: 0.5083 - val_accuracy: 0.8270 - 2s/epoch - 19ms/step\n","Epoch 297/300\n","118/118 - 2s - loss: 0.3656 - accuracy: 0.8720 - val_loss: 0.5019 - val_accuracy: 0.8307 - 2s/epoch - 19ms/step\n","Epoch 298/300\n","118/118 - 2s - loss: 0.3601 - accuracy: 0.8751 - val_loss: 0.4932 - val_accuracy: 0.8352 - 2s/epoch - 19ms/step\n","Epoch 299/300\n","118/118 - 2s - loss: 0.3578 - accuracy: 0.8754 - val_loss: 0.5015 - val_accuracy: 0.8369 - 2s/epoch - 19ms/step\n","Epoch 300/300\n","118/118 - 2s - loss: 0.3758 - accuracy: 0.8694 - val_loss: 0.5098 - val_accuracy: 0.8350 - 2s/epoch - 19ms/step\n"]}],"source":["# first, second, third, and fourth\n","#################### YOUR CODE HERE ####################\n","# Các bạn thực hiện tương tự như Ví dụ train lần 1 \n","# Train lần 7:  Tiếp tục train thêm 300 epoch nhưng lần này không fix weight của sub model 1, sub model 2 và sub model 3\n","# (tạo model mới và thêm Input layer, sub model 1, sub model 2, sub model 3, sub model 4 , Output layer)\n","fourth_model = keras.Sequential()\n","fourth_model.add(keras.Input(shape=(784,)))\n","fourth_model.add(first)\n","fourth_model.add(second)\n","fourth_model.add(third)\n","fourth_model.add(fourth)\n","fourth_model.add(keras.layers.Dense(10))\n","\n","# unfreeze sub model 1 ,sub model 2 và sub model 3\n","first.trainable = True\n","second.trainable = True\n","third.trainable = True\n","########################################################\n","\n","fourth_model.compile(optimizer=tf.keras.optimizers.SGD(), \n","              loss=keras.losses.SparseCategoricalCrossentropy(from_logits=True), \n","              metrics=['accuracy'])\n","\n","history = fourth_model.fit(X_train, y_train, batch_size,\n","                          validation_data=(X_test, y_test),\n","                          epochs=300, verbose=2)"]},{"cell_type":"code","execution_count":22,"id":"RlpO5sCeLbVU","metadata":{"id":"RlpO5sCeLbVU","executionInfo":{"status":"ok","timestamp":1669969897166,"user_tz":-420,"elapsed":13,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"}}},"outputs":[],"source":[]},{"cell_type":"code","execution_count":22,"id":"rNOs-wOJLbQQ","metadata":{"id":"rNOs-wOJLbQQ","executionInfo":{"status":"ok","timestamp":1669969897167,"user_tz":-420,"elapsed":13,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"}}},"outputs":[],"source":[]},{"cell_type":"code","execution_count":23,"id":"1a6d5541","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":296},"executionInfo":{"elapsed":14,"status":"ok","timestamp":1669969897168,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"},"user_tz":-420},"id":"1a6d5541","outputId":"7c6a873f-ba22-4dec-ff30-32d349949721"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.legend.Legend at 0x7f134b928be0>"]},"metadata":{},"execution_count":23},{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 1 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}}],"source":["import matplotlib.pyplot as plt\n","\n","plt.plot(history.history['loss'], label='loss')\n","plt.plot(history.history['val_loss'], label='val_loss')\n","plt.xlabel('iteration')\n","plt.ylabel('Loss')\n","plt.legend()"]},{"cell_type":"code","execution_count":23,"id":"1e570b52","metadata":{"id":"1e570b52","executionInfo":{"status":"ok","timestamp":1669969897169,"user_tz":-420,"elapsed":10,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"}}},"outputs":[],"source":[]},{"cell_type":"code","execution_count":24,"id":"bf862fd5","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":296},"executionInfo":{"elapsed":776,"status":"ok","timestamp":1669969897935,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"},"user_tz":-420},"id":"bf862fd5","outputId":"c8143d2a-2ad5-43e6-d6cc-1b416cc4e8b2"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.legend.Legend at 0x7f134d5eadf0>"]},"metadata":{},"execution_count":24},{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 1 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}}],"source":["import matplotlib.pyplot as plt\n","\n","plt.plot(history.history['accuracy'], label='train_accuracy')\n","plt.plot(history.history['val_accuracy'], label='val_accuracy')\n","plt.xlabel('iteration')\n","plt.ylabel('Accuracy')\n","plt.legend()"]},{"cell_type":"code","execution_count":24,"id":"317043ed","metadata":{"id":"317043ed","executionInfo":{"status":"ok","timestamp":1669969897936,"user_tz":-420,"elapsed":10,"user":{"displayName":"Thanh Hùng","userId":"08335145449228425002"}}},"outputs":[],"source":[]}],"metadata":{"colab":{"provenance":[],"toc_visible":true,"collapsed_sections":["01243d95","8623181e"]},"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":"3.8.8"},"gpuClass":"standard"},"nbformat":4,"nbformat_minor":5}