public-projects/knuckletouch
1
0
Fork 0
Code Issues Pull requests Releases Wiki Activity
knuckletouch/python/Step_12_LSTM-WarmStart.ipynb

10108 lines
496 KiB
Text
Raw Permalink Normal View History

Added notebooks and models
2019-08-07 17:57:12 -04:00
{
"cells": [
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"Using TensorFlow backend.\n"
]
}
],
"source": [
"## USE for Multi GPU Systems\n",
"#import os\n",
"#os.environ[\"CUDA_VISIBLE_DEVICES\"]=\"0\"\n",
"\n",
"from keras.models import Sequential, load_model\n",
"from keras.layers import *\n",
"from keras import optimizers\n",
"from keras import utils\n",
"from keras.callbacks import TensorBoard, ModelCheckpoint, ReduceLROnPlateau\n",
"import keras\n",
"\n",
"import numpy as np\n",
"import matplotlib.pyplot as plt\n",
"import pandas as pd\n",
"import math\n",
"\n",
"import tensorflow as tf\n",
"\n",
"# Importing matplotlib to plot images.\n",
"import matplotlib.pyplot as plt\n",
"import numpy as np\n",
"%matplotlib inline\n",
"\n",
"# Importing SK-learn to calculate precision and recall\n",
"import sklearn\n",
"from sklearn import metrics\n",
"from sklearn.model_selection import train_test_split, cross_val_score, LeaveOneGroupOut\n",
"from sklearn.utils import shuffle \n",
"\n",
"# Used for graph export\n",
"from keras import backend as K"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [],
"source": [
"class LoggingTensorBoard(TensorBoard): \n",
"\n",
" def __init__(self, log_dir, settings_str_to_log, **kwargs):\n",
" super(LoggingTensorBoard, self).__init__(log_dir, **kwargs)\n",
"\n",
" self.settings_str = settings_str_to_log\n",
"\n",
" def on_train_begin(self, logs=None):\n",
" TensorBoard.on_train_begin(self, logs=logs)\n",
"\n",
" tensor = tf.convert_to_tensor(self.settings_str)\n",
" summary = tf.summary.text (\"Run_Settings\", tensor)\n",
"\n",
" with tf.Session() as sess:\n",
" s = sess.run(summary)\n",
" self.writer.add_summary(s)"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"[ 1 2 9 6 4 14 17 16 12 3 10 18 5] [13 8 11 15 7]\n"
]
}
],
"source": [
"dfAll = pd.read_pickle(\"DataStudyCollection/df_lstm_norm50.pkl\")\n",
"\n",
"lst = dfAll.userID.unique()\n",
"np.random.seed(42)\n",
"np.random.shuffle(lst)\n",
"test_ids = lst[-5:]\n",
"train_ids = lst[:-5]\n",
"print(train_ids, test_ids)"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"array([ 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17,\n",
" 18])"
]
},
"execution_count": 4,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"dfAll.userID.unique()"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [],
"source": [
"dfAll.TaskID = dfAll.TaskID % 17"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [],
"source": [
"df_train = dfAll[dfAll.userID.isin(train_ids)][['Blobs', 'TaskID']]\n",
"df_test = dfAll[dfAll.userID.isin(test_ids)][['Blobs', 'TaskID']]"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [],
"source": [
"x_train = np.concatenate(df_train.Blobs.values).reshape(-1,50,27,15,1)\n",
"x_test = np.concatenate(df_test.Blobs.values).reshape(-1,50,27,15,1)\n",
"\n",
"y_train = df_train.TaskID.values\n",
"y_test = df_test.TaskID.values\n",
"\n",
"x_train = x_train / 255.0\n",
"x_test = x_test / 255.0\n",
"\n",
"# convert class vectors to binary class matrices (one-hot notation)\n",
"num_classes = len(dfAll.TaskID.unique())\n",
"y_train_one_hot = utils.to_categorical(df_train.TaskID, num_classes)\n",
"y_test_one_hot = utils.to_categorical(df_test.TaskID, num_classes)"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {},
"outputs": [],
"source": [
"# If GPU is not available: \n",
"# GPU_USE = '/cpu:0'\n",
"#config = tf.ConfigProto(device_count = {\"GPU\": 1})\n",
"\n",
"\n",
"# If GPU is available: \n",
"config = tf.ConfigProto()\n",
"config.log_device_placement = True\n",
"config.allow_soft_placement = True\n",
"config.gpu_options.allow_growth=True\n",
"config.gpu_options.allocator_type = 'BFC'\n",
"\n",
"# Limit the maximum memory used\n",
"config.gpu_options.per_process_gpu_memory_fraction = 0.3\n",
"\n",
"# set session config\n",
"tf.keras.backend.set_session(tf.Session(config=config))"
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {
"scrolled": false
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"_________________________________________________________________\n",
"Layer (type) Output Shape Param # \n",
"=================================================================\n",
"time_distributed_10 (TimeDis (None, 50, 27, 15, 64) 640 \n",
"_________________________________________________________________\n",
"time_distributed_11 (TimeDis (None, 50, 27, 15, 32) 18464 \n",
"_________________________________________________________________\n",
"time_distributed_12 (TimeDis (None, 50, 14, 8, 32) 0 \n",
"_________________________________________________________________\n",
"time_distributed_13 (TimeDis (None, 50, 14, 8, 32) 0 \n",
"_________________________________________________________________\n",
"time_distributed_14 (TimeDis (None, 50, 14, 8, 32) 9248 \n",
"_________________________________________________________________\n",
"time_distributed_15 (TimeDis (None, 50, 14, 8, 16) 4624 \n",
"_________________________________________________________________\n",
"time_distributed_16 (TimeDis (None, 50, 7, 4, 16) 0 \n",
"_________________________________________________________________\n",
"time_distributed_17 (TimeDis (None, 50, 7, 4, 16) 0 \n",
"_________________________________________________________________\n",
"time_distributed_18 (TimeDis (None, 50, 448) 0 \n",
"_________________________________________________________________\n",
"lstm_3 (LSTM) (None, 50, 80) 169280 \n",
"_________________________________________________________________\n",
"dropout_7 (Dropout) (None, 50, 80) 0 \n",
"_________________________________________________________________\n",
"lstm_4 (LSTM) (None, 50) 26200 \n",
"_________________________________________________________________\n",
"dropout_8 (Dropout) (None, 50) 0 \n",
"_________________________________________________________________\n",
"dense_2 (Dense) (None, 17) 867 \n",
"=================================================================\n",
"Total params: 229,323\n",
"Trainable params: 229,323\n",
"Non-trainable params: 0\n",
"_________________________________________________________________\n",
"LSTM-v2\n",
"Train on 6624 samples, validate on 2569 samples\n",
"Epoch 1/3000\n",
" - 38s - loss: 0.6056 - acc: 0.8918 - val_loss: 1.0498 - val_acc: 0.8198\n",
"\n",
"Epoch 00001: val_acc improved from -inf to 0.81977, saving model to ./ModelSnapshots/LSTM-v2-001.h5\n",
"Epoch 2/3000\n",
" - 38s - loss: 0.5269 - acc: 0.9111 - val_loss: 0.9867 - val_acc: 0.8342\n",
"\n",
"Epoch 00002: val_acc improved from 0.81977 to 0.83418, saving model to ./ModelSnapshots/LSTM-v2-002.h5\n",
"Epoch 3/3000\n",
" - 40s - loss: 0.4995 - acc: 0.9215 - val_loss: 1.0149 - val_acc: 0.8315\n",
"\n",
"Epoch 00003: val_acc did not improve from 0.83418\n",
"Epoch 4/3000\n",
" - 39s - loss: 0.4941 - acc: 0.9198 - val_loss: 0.9717 - val_acc: 0.8326\n",
"\n",
"Epoch 00004: val_acc did not improve from 0.83418\n",
"Epoch 5/3000\n",
" - 39s - loss: 0.4916 - acc: 0.9238 - val_loss: 0.9620 - val_acc: 0.8392\n",
"\n",
"Epoch 00005: val_acc improved from 0.83418 to 0.83924, saving model to ./ModelSnapshots/LSTM-v2-005.h5\n",
"Epoch 6/3000\n",
" - 40s - loss: 0.4997 - acc: 0.9241 - val_loss: 1.1411 - val_acc: 0.8108\n",
"\n",
"Epoch 00006: val_acc did not improve from 0.83924\n",
"Epoch 7/3000\n",
" - 40s - loss: 0.5363 - acc: 0.9159 - val_loss: 0.9515 - val_acc: 0.8279\n",
"\n",
"Epoch 00007: val_acc did not improve from 0.83924\n",
"Epoch 8/3000\n",
" - 39s - loss: 0.4610 - acc: 0.9289 - val_loss: 0.9546 - val_acc: 0.8381\n",
"\n",
"Epoch 00008: val_acc did not improve from 0.83924\n",
"Epoch 9/3000\n",
" - 39s - loss: 0.4673 - acc: 0.9283 - val_loss: 0.9096 - val_acc: 0.8459\n",
"\n",
"Epoch 00009: val_acc improved from 0.83924 to 0.84585, saving model to ./ModelSnapshots/LSTM-v2-009.h5\n",
"Epoch 10/3000\n",
" - 39s - loss: 0.4555 - acc: 0.9295 - val_loss: 1.0060 - val_acc: 0.8330\n",
"\n",
"Epoch 00010: val_acc did not improve from 0.84585\n",
"Epoch 11/3000\n",
" - 39s - loss: 0.4624 - acc: 0.9296 - val_loss: 0.9722 - val_acc: 0.8318\n",
"\n",
"Epoch 00011: val_acc did not improve from 0.84585\n",
"Epoch 12/3000\n",
" - 39s - loss: 0.4971 - acc: 0.9209 - val_loss: 1.1189 - val_acc: 0.8062\n",
"\n",
"Epoch 00012: val_acc did not improve from 0.84585\n",
"Epoch 13/3000\n",
" - 39s - loss: 0.5239 - acc: 0.9126 - val_loss: 0.9452 - val_acc: 0.8365\n",
"\n",
"Epoch 00013: val_acc did not improve from 0.84585\n",
"Epoch 14/3000\n",
" - 40s - loss: 0.4821 - acc: 0.9224 - val_loss: 0.9675 - val_acc: 0.8369\n",
"\n",
"Epoch 00014: val_acc did not improve from 0.84585\n",
"Epoch 15/3000\n",
" - 39s - loss: 0.4525 - acc: 0.9281 - val_loss: 0.9729 - val_acc: 0.8404\n",
"\n",
"Epoch 00015: val_acc did not improve from 0.84585\n",
"Epoch 16/3000\n",
" - 40s - loss: 0.4651 - acc: 0.9295 - val_loss: 0.9648 - val_acc: 0.8392\n",
"\n",
"Epoch 00016: val_acc did not improve from 0.84585\n",
"Epoch 17/3000\n",
" - 39s - loss: 0.4269 - acc: 0.9366 - val_loss: 0.8995 - val_acc: 0.8466\n",
"\n",
"Epoch 00017: val_acc improved from 0.84585 to 0.84663, saving model to ./ModelSnapshots/LSTM-v2-017.h5\n",
"Epoch 18/3000\n",
" - 39s - loss: 0.4641 - acc: 0.9268 - val_loss: 0.9171 - val_acc: 0.8431\n",
"\n",
"Epoch 00018: val_acc did not improve from 0.84663\n",
"Epoch 19/3000\n",
" - 39s - loss: 0.4250 - acc: 0.9398 - val_loss: 0.9650 - val_acc: 0.8392\n",
"\n",
"Epoch 00019: val_acc did not improve from 0.84663\n",
"Epoch 20/3000\n",
" - 39s - loss: 0.4639 - acc: 0.9330 - val_loss: 0.9631 - val_acc: 0.8404\n",
"\n",
"Epoch 00020: val_acc did not improve from 0.84663\n",
"Epoch 21/3000\n",
" - 39s - loss: 0.4601 - acc: 0.9315 - val_loss: 0.9159 - val_acc: 0.8443\n",
"\n",
"Epoch 00021: val_acc did not improve from 0.84663\n",
"Epoch 22/3000\n",
" - 39s - loss: 0.4361 - acc: 0.9370 - val_loss: 0.8874 - val_acc: 0.8470\n",
"\n",
"Epoch 00022: val_acc improved from 0.84663 to 0.84702, saving model to ./ModelSnapshots/LSTM-v2-022.h5\n",
"Epoch 23/3000\n",
" - 39s - loss: 0.4560 - acc: 0.9304 - val_loss: 0.8797 - val_acc: 0.8494\n",
"\n",
"Epoch 00023: val_acc improved from 0.84702 to 0.84936, saving model to ./ModelSnapshots/LSTM-v2-023.h5\n",
"Epoch 24/3000\n",
" - 39s - loss: 0.4309 - acc: 0.9355 - val_loss: 0.8776 - val_acc: 0.8529\n",
"\n",
"Epoch 00024: val_acc improved from 0.84936 to 0.85286, saving model to ./ModelSnapshots/LSTM-v2-024.h5\n",
"Epoch 25/3000\n",
" - 40s - loss: 0.4402 - acc: 0.9345 - val_loss: 0.9515 - val_acc: 0.8451\n",
"\n",
"Epoch 00025: val_acc did not improve from 0.85286\n",
"Epoch 26/3000\n",
" - 39s - loss: 0.4727 - acc: 0.9309 - val_loss: 0.9700 - val_acc: 0.8256\n",
"\n",
"Epoch 00026: val_acc did not improve from 0.85286\n",
"Epoch 27/3000\n",
" - 39s - loss: 0.4610 - acc: 0.9286 - val_loss: 0.9460 - val_acc: 0.8342\n",
"\n",
"Epoch 00027: val_acc did not improve from 0.85286\n",
"Epoch 28/3000\n",
" - 39s - loss: 0.4874 - acc: 0.9265 - val_loss: 0.9083 - val_acc: 0.8392\n",
"\n",
"Epoch 00028: val_acc did not improve from 0.85286\n",
"Epoch 29/3000\n",
" - 38s - loss: 0.4450 - acc: 0.9321 - val_loss: 0.8898 - val_acc: 0.8525\n",
"\n",
"Epoch 00029: val_acc did not improve from 0.85286\n",
"Epoch 30/3000\n",
" - 40s - loss: 0.4875 - acc: 0.9229 - val_loss: 0.9824 - val_acc: 0.8443\n",
"\n",
"Epoch 00030: val_acc did not improve from 0.85286\n",
"Epoch 31/3000\n",
" - 39s - loss: 0.4338 - acc: 0.9386 - val_loss: 0.9317 - val_acc: 0.8501\n",
"\n",
"Epoch 00031: val_acc did not improve from 0.85286\n",
"Epoch 32/3000\n",
" - 39s - loss: 0.4561 - acc: 0.9327 - val_loss: 0.9210 - val_acc: 0.8435\n",
"\n",
"Epoch 00032: val_acc did not improve from 0.85286\n",
"Epoch 33/3000\n",
" - 39s - loss: 0.4344 - acc: 0.9319 - val_loss: 0.9535 - val_acc: 0.8466\n",
"\n",
"Epoch 00033: val_acc did not improve from 0.85286\n",
"Epoch 34/3000\n",
" - 37s - loss: 0.4195 - acc: 0.9413 - val_loss: 0.8583 - val_acc: 0.8552\n",
"\n",
"Epoch 00034: val_acc improved from 0.85286 to 0.85520, saving model to ./ModelSnapshots/LSTM-v2-034.h5\n",
"Epoch 35/3000\n",
" - 38s - loss: 0.4471 - acc: 0.9343 - val_loss: 0.9999 - val_acc: 0.8373\n",
"\n",
"Epoch 00035: val_acc did not improve from 0.85520\n",
"Epoch 36/3000\n",
" - 40s - loss: 0.4331 - acc: 0.9351 - val_loss: 0.9634 - val_acc: 0.8361\n",
"\n",
"Epoch 00036: val_acc did not improve from 0.85520\n",
"Epoch 37/3000\n",
" - 39s - loss: 0.4321 - acc: 0.9373 - val_loss: 0.9405 - val_acc: 0.8431\n",
"\n",
"Epoch 00037: val_acc did not improve from 0.85520\n",
"Epoch 38/3000\n",
" - 44s - loss: 0.4193 - acc: 0.9402 - val_loss: 0.9312 - val_acc: 0.8486\n",
"\n",
"Epoch 00038: val_acc did not improve from 0.85520\n",
"Epoch 39/3000\n",
" - 40s - loss: 0.4308 - acc: 0.9396 - val_loss: 0.9360 - val_acc: 0.8431\n",
"\n",
"Epoch 00039: val_acc did not improve from 0.85520\n",
"Epoch 40/3000\n",
" - 42s - loss: 0.4237 - acc: 0.9373 - val_loss: 0.8747 - val_acc: 0.8603\n",
"\n",
"Epoch 00040: val_acc improved from 0.85520 to 0.86026, saving model to ./ModelSnapshots/LSTM-v2-040.h5\n",
"Epoch 41/3000\n",
" - 41s - loss: 0.4495 - acc: 0.9327 - val_loss: 0.9276 - val_acc: 0.8447\n",
"\n",
"Epoch 00041: val_acc did not improve from 0.86026\n",
"Epoch 42/3000\n",
" - 39s - loss: 0.4586 - acc: 0.9358 - val_loss: 0.9724 - val_acc: 0.8408\n",
"\n",
"Epoch 00042: val_acc did not improve from 0.86026\n",
"Epoch 43/3000\n",
" - 40s - loss: 0.4676 - acc: 0.9309 - val_loss: 0.9681 - val_acc: 0.8466\n",
"\n",
"Epoch 00043: val_acc did not improve from 0.86026\n",
"Epoch 44/3000\n",
" - 39s - loss: 0.4382 - acc: 0.9364 - val_loss: 0.8846 - val_acc: 0.8540\n",
"\n",
"Epoch 00044: val_acc did not improve from 0.86026\n",
"Epoch 45/3000\n",
" - 39s - loss: 0.4467 - acc: 0.9346 - val_loss: 0.9072 - val_acc: 0.8459\n",
"\n",
"Epoch 00045: val_acc did not improve from 0.86026\n",
"Epoch 46/3000\n",
" - 39s - loss: 0.4176 - acc: 0.9410 - val_loss: 0.9573 - val_acc: 0.8451\n",
"\n",
"Epoch 00046: val_acc did not improve from 0.86026\n",
"Epoch 47/3000\n",
" - 39s - loss: 0.4432 - acc: 0.9327 - val_loss: 0.9924 - val_acc: 0.8350\n",
"\n",
"Epoch 00047: val_acc did not improve from 0.86026\n",
"Epoch 48/3000\n",
" - 39s - loss: 0.4237 - acc: 0.9387 - val_loss: 1.0595 - val_acc: 0.8264\n",
"\n",
"Epoch 00048: val_acc did not improve from 0.86026\n",
"Epoch 49/3000\n",
" - 39s - loss: 0.4519 - acc: 0.9345 - val_loss: 0.9091 - val_acc: 0.8447\n",
"\n",
"Epoch 00049: val_acc did not improve from 0.86026\n",
"Epoch 50/3000\n",
" - 40s - loss: 0.4514 - acc: 0.9331 - val_loss: 0.8948 - val_acc: 0.8556\n",
"\n",
"Epoch 00050: val_acc did not improve from 0.86026\n",
"Epoch 51/3000\n",
" - 44s - loss: 0.4385 - acc: 0.9380 - val_loss: 0.9468 - val_acc: 0.8369\n",
"\n",
"Epoch 00051: val_acc did not improve from 0.86026\n",
"Epoch 52/3000\n",
" - 39s - loss: 0.4371 - acc: 0.9373 - val_loss: 0.9737 - val_acc: 0.8486\n",
"\n",
"Epoch 00052: val_acc did not improve from 0.86026\n",
"Epoch 53/3000\n",
" - 40s - loss: 0.4320 - acc: 0.9411 - val_loss: 0.9623 - val_acc: 0.8412\n",
"\n",
"Epoch 00053: val_acc did not improve from 0.86026\n",
"Epoch 54/3000\n",
" - 40s - loss: 0.4183 - acc: 0.9438 - val_loss: 0.8416 - val_acc: 0.8591\n",
"\n",
"Epoch 00054: val_acc did not improve from 0.86026\n",
"Epoch 55/3000\n",
" - 39s - loss: 0.4246 - acc: 0.9422 - val_loss: 0.9440 - val_acc: 0.8385\n",
"\n",
"Epoch 00055: val_acc did not improve from 0.86026\n",
"Epoch 56/3000\n",
" - 40s - loss: 0.4368 - acc: 0.9387 - val_loss: 0.9290 - val_acc: 0.8420\n",
"\n",
"Epoch 00056: val_acc did not improve from 0.86026\n",
"Epoch 57/3000\n",
" - 39s - loss: 0.4627 - acc: 0.9333 - val_loss: 0.8787 - val_acc: 0.8474\n",
"\n",
"Epoch 00057: val_acc did not improve from 0.86026\n",
"Epoch 58/3000\n",
" - 40s - loss: 0.4149 - acc: 0.9428 - val_loss: 0.9141 - val_acc: 0.8490\n",
"\n",
"Epoch 00058: val_acc did not improve from 0.86026\n",
"Epoch 59/3000\n",
" - 39s - loss: 0.4551 - acc: 0.9372 - val_loss: 0.8431 - val_acc: 0.8560\n",
"\n",
"Epoch 00059: val_acc did not improve from 0.86026\n",
"Epoch 60/3000\n",
" - 39s - loss: 0.4260 - acc: 0.9399 - val_loss: 0.8984 - val_acc: 0.8548\n",
"\n",
"Epoch 00060: val_acc did not improve from 0.86026\n",
"Epoch 61/3000\n",
" - 40s - loss: 0.4569 - acc: 0.9348 - val_loss: 0.8772 - val_acc: 0.8474\n",
"\n",
"Epoch 00061: val_acc did not improve from 0.86026\n",
"Epoch 62/3000\n",
" - 39s - loss: 0.4428 - acc: 0.9405 - val_loss: 0.9370 - val_acc: 0.8431\n",
"\n",
"Epoch 00062: val_acc did not improve from 0.86026\n",
"Epoch 63/3000\n",
" - 39s - loss: 0.4368 - acc: 0.9363 - val_loss: 0.9072 - val_acc: 0.8424\n",
"\n",
"Epoch 00063: val_acc did not improve from 0.86026\n",
"Epoch 64/3000\n",
" - 40s - loss: 0.4608 - acc: 0.9307 - val_loss: 0.8726 - val_acc: 0.8505\n",
"\n",
"Epoch 00064: val_acc did not improve from 0.86026\n",
"Epoch 65/3000\n",
" - 39s - loss: 0.4588 - acc: 0.9364 - val_loss: 0.9302 - val_acc: 0.8478\n",
"\n",
"Epoch 00065: val_acc did not improve from 0.86026\n",
"Epoch 66/3000\n",
" - 40s - loss: 0.4514 - acc: 0.9369 - val_loss: 0.9376 - val_acc: 0.8490\n",
"\n",
"Epoch 00066: val_acc did not improve from 0.86026\n",
"Epoch 67/3000\n",
" - 39s - loss: 0.4393 - acc: 0.9395 - val_loss: 0.9064 - val_acc: 0.8521\n",
"\n",
"Epoch 00067: val_acc did not improve from 0.86026\n",
"Epoch 68/3000\n",
" - 40s - loss: 0.4319 - acc: 0.9389 - val_loss: 0.8993 - val_acc: 0.8490\n",
"\n",
"Epoch 00068: val_acc did not improve from 0.86026\n",
"Epoch 69/3000\n",
" - 39s - loss: 0.4387 - acc: 0.9360 - val_loss: 0.8797 - val_acc: 0.8521\n",
"\n",
"Epoch 00069: val_acc did not improve from 0.86026\n",
"Epoch 70/3000\n",
" - 39s - loss: 0.4307 - acc: 0.9392 - val_loss: 0.9072 - val_acc: 0.8466\n",
"\n",
"Epoch 00070: val_acc did not improve from 0.86026\n",
"Epoch 71/3000\n",
" - 38s - loss: 0.4298 - acc: 0.9386 - val_loss: 0.9000 - val_acc: 0.8564\n",
"\n",
"Epoch 00071: val_acc did not improve from 0.86026\n",
"Epoch 72/3000\n",
" - 39s - loss: 0.4512 - acc: 0.9358 - val_loss: 0.9628 - val_acc: 0.8443\n",
"\n",
"Epoch 00072: val_acc did not improve from 0.86026\n",
"Epoch 73/3000\n",
" - 39s - loss: 0.4398 - acc: 0.9380 - val_loss: 0.9015 - val_acc: 0.8404\n",
"\n",
"Epoch 00073: val_acc did not improve from 0.86026\n",
"Epoch 74/3000\n",
" - 39s - loss: 0.4250 - acc: 0.9405 - val_loss: 0.9228 - val_acc: 0.8416\n",
"\n",
"Epoch 00074: val_acc did not improve from 0.86026\n",
"Epoch 75/3000\n",
" - 39s - loss: 0.4376 - acc: 0.9389 - val_loss: 1.0756 - val_acc: 0.8264\n",
"\n",
"Epoch 00075: val_acc did not improve from 0.86026\n",
"Epoch 76/3000\n",
" - 40s - loss: 0.4561 - acc: 0.9345 - val_loss: 0.9589 - val_acc: 0.8346\n",
"\n",
"Epoch 00076: val_acc did not improve from 0.86026\n",
"Epoch 77/3000\n",
" - 39s - loss: 0.4814 - acc: 0.9318 - val_loss: 0.9344 - val_acc: 0.8435\n",
"\n",
"Epoch 00077: val_acc did not improve from 0.86026\n",
"Epoch 78/3000\n",
" - 39s - loss: 0.4547 - acc: 0.9336 - val_loss: 0.9314 - val_acc: 0.8404\n",
"\n",
"Epoch 00078: val_acc did not improve from 0.86026\n",
"Epoch 79/3000\n",
" - 40s - loss: 0.4177 - acc: 0.9438 - val_loss: 0.9750 - val_acc: 0.8431\n",
"\n",
"Epoch 00079: val_acc did not improve from 0.86026\n",
"Epoch 80/3000\n",
" - 42s - loss: 0.4243 - acc: 0.9408 - val_loss: 0.9280 - val_acc: 0.8424\n",
"\n",
"Epoch 00080: val_acc did not improve from 0.86026\n",
"Epoch 81/3000\n",
" - 41s - loss: 0.4167 - acc: 0.9432 - val_loss: 0.9659 - val_acc: 0.8427\n",
"\n",
"Epoch 00081: val_acc did not improve from 0.86026\n",
"Epoch 82/3000\n",
" - 39s - loss: 0.4214 - acc: 0.9413 - val_loss: 0.9068 - val_acc: 0.8470\n",
"\n",
"Epoch 00082: val_acc did not improve from 0.86026\n",
"Epoch 83/3000\n",
" - 39s - loss: 0.4220 - acc: 0.9389 - val_loss: 0.9267 - val_acc: 0.8424\n",
"\n",
"Epoch 00083: val_acc did not improve from 0.86026\n",
"Epoch 84/3000\n",
" - 39s - loss: 0.4349 - acc: 0.9399 - val_loss: 0.8673 - val_acc: 0.8505\n",
"\n",
"Epoch 00084: val_acc did not improve from 0.86026\n",
"\n",
"Epoch 00084: ReduceLROnPlateau reducing learning rate to 9.499999760009814e-05.\n",
"Epoch 85/3000\n",
" - 39s - loss: 0.4043 - acc: 0.9423 - val_loss: 0.9716 - val_acc: 0.8424\n",
"\n",
"Epoch 00085: val_acc did not improve from 0.86026\n",
"Epoch 86/3000\n",
" - 39s - loss: 0.4033 - acc: 0.9470 - val_loss: 0.8990 - val_acc: 0.8494\n",
"\n",
"Epoch 00086: val_acc did not improve from 0.86026\n",
"Epoch 87/3000\n",
" - 39s - loss: 0.4128 - acc: 0.9420 - val_loss: 0.9176 - val_acc: 0.8497\n",
"\n",
"Epoch 00087: val_acc did not improve from 0.86026\n",
"Epoch 88/3000\n",
" - 39s - loss: 0.4322 - acc: 0.9413 - val_loss: 0.9477 - val_acc: 0.8455\n",
"\n",
"Epoch 00088: val_acc did not improve from 0.86026\n",
"Epoch 89/3000\n",
" - 39s - loss: 0.4287 - acc: 0.9398 - val_loss: 0.9319 - val_acc: 0.8478\n",
"\n",
"Epoch 00089: val_acc did not improve from 0.86026\n",
"Epoch 90/3000\n",
" - 39s - loss: 0.4087 - acc: 0.9429 - val_loss: 0.9390 - val_acc: 0.8381\n",
"\n",
"Epoch 00090: val_acc did not improve from 0.86026\n",
"Epoch 91/3000\n",
" - 39s - loss: 0.4065 - acc: 0.9457 - val_loss: 0.9405 - val_acc: 0.8385\n",
"\n",
"Epoch 00091: val_acc did not improve from 0.86026\n",
"Epoch 92/3000\n",
" - 39s - loss: 0.4121 - acc: 0.9438 - val_loss: 0.9280 - val_acc: 0.8443\n",
"\n",
"Epoch 00092: val_acc did not improve from 0.86026\n",
"Epoch 93/3000\n",
" - 39s - loss: 0.4262 - acc: 0.9389 - val_loss: 0.9276 - val_acc: 0.8342\n",
"\n",
"Epoch 00093: val_acc did not improve from 0.86026\n",
"Epoch 94/3000\n",
" - 39s - loss: 0.4168 - acc: 0.9432 - val_loss: 0.8850 - val_acc: 0.8392\n",
"\n",
"Epoch 00094: val_acc did not improve from 0.86026\n",
"Epoch 95/3000\n",
" - 39s - loss: 0.4275 - acc: 0.9401 - val_loss: 0.9073 - val_acc: 0.8482\n",
"\n",
"Epoch 00095: val_acc did not improve from 0.86026\n",
"Epoch 96/3000\n",
" - 39s - loss: 0.4154 - acc: 0.9410 - val_loss: 0.9839 - val_acc: 0.8299\n",
"\n",
"Epoch 00096: val_acc did not improve from 0.86026\n",
"Epoch 97/3000\n",
" - 40s - loss: 0.4134 - acc: 0.9446 - val_loss: 1.0332 - val_acc: 0.8276\n",
"\n",
"Epoch 00097: val_acc did not improve from 0.86026\n",
"Epoch 98/3000\n",
" - 39s - loss: 0.4648 - acc: 0.9306 - val_loss: 1.0208 - val_acc: 0.8365\n",
"\n",
"Epoch 00098: val_acc did not improve from 0.86026\n",
"Epoch 99/3000\n",
" - 39s - loss: 0.4515 - acc: 0.9369 - val_loss: 0.9568 - val_acc: 0.8435\n",
"\n",
"Epoch 00099: val_acc did not improve from 0.86026\n",
"Epoch 100/3000\n",
" - 39s - loss: 0.4186 - acc: 0.9411 - val_loss: 0.9582 - val_acc: 0.8365\n",
"\n",
"Epoch 00100: val_acc did not improve from 0.86026\n",
"Epoch 101/3000\n",
" - 39s - loss: 0.4065 - acc: 0.9441 - val_loss: 0.9101 - val_acc: 0.8509\n",
"\n",
"Epoch 00101: val_acc did not improve from 0.86026\n",
"Epoch 102/3000\n",
" - 39s - loss: 0.4159 - acc: 0.9429 - val_loss: 0.9969 - val_acc: 0.8388\n",
"\n",
"Epoch 00102: val_acc did not improve from 0.86026\n",
"Epoch 103/3000\n",
" - 39s - loss: 0.4549 - acc: 0.9334 - val_loss: 1.0005 - val_acc: 0.8447\n",
"\n",
"Epoch 00103: val_acc did not improve from 0.86026\n",
"Epoch 104/3000\n",
" - 39s - loss: 0.4101 - acc: 0.9449 - val_loss: 0.7630 - val_acc: 0.8665\n",
"\n",
"Epoch 00104: val_acc improved from 0.86026 to 0.86649, saving model to ./ModelSnapshots/LSTM-v2-104.h5\n",
"Epoch 105/3000\n",
" - 39s - loss: 0.4220 - acc: 0.9405 - val_loss: 0.8933 - val_acc: 0.8525\n",
"\n",
"Epoch 00105: val_acc did not improve from 0.86649\n",
"Epoch 106/3000\n",
" - 38s - loss: 0.4094 - acc: 0.9417 - val_loss: 0.8874 - val_acc: 0.8490\n",
"\n",
"Epoch 00106: val_acc did not improve from 0.86649\n",
"Epoch 107/3000\n",
" - 38s - loss: 0.4191 - acc: 0.9401 - val_loss: 0.8584 - val_acc: 0.8579\n",
"\n",
"Epoch 00107: val_acc did not improve from 0.86649\n",
"Epoch 108/3000\n",
" - 39s - loss: 0.4098 - acc: 0.9446 - val_loss: 0.9456 - val_acc: 0.8424\n",
"\n",
"Epoch 00108: val_acc did not improve from 0.86649\n",
"Epoch 109/3000\n",
" - 39s - loss: 0.4520 - acc: 0.9348 - val_loss: 0.9103 - val_acc: 0.8447\n",
"\n",
"Epoch 00109: val_acc did not improve from 0.86649\n",
"Epoch 110/3000\n",
" - 39s - loss: 0.4052 - acc: 0.9447 - val_loss: 0.8925 - val_acc: 0.8521\n",
"\n",
"Epoch 00110: val_acc did not improve from 0.86649\n",
"Epoch 111/3000\n",
" - 39s - loss: 0.3867 - acc: 0.9467 - val_loss: 0.9218 - val_acc: 0.8459\n",
"\n",
"Epoch 00111: val_acc did not improve from 0.86649\n",
"Epoch 112/3000\n",
" - 39s - loss: 0.4004 - acc: 0.9466 - val_loss: 0.9572 - val_acc: 0.8424\n",
"\n",
"Epoch 00112: val_acc did not improve from 0.86649\n",
"Epoch 113/3000\n",
" - 40s - loss: 0.4357 - acc: 0.9370 - val_loss: 0.9873 - val_acc: 0.8420\n",
"\n",
"Epoch 00113: val_acc did not improve from 0.86649\n",
"Epoch 114/3000\n",
" - 39s - loss: 0.4327 - acc: 0.9399 - val_loss: 0.9470 - val_acc: 0.8443\n",
"\n",
"Epoch 00114: val_acc did not improve from 0.86649\n",
"Epoch 115/3000\n",
" - 39s - loss: 0.4282 - acc: 0.9454 - val_loss: 0.8582 - val_acc: 0.8548\n",
"\n",
"Epoch 00115: val_acc did not improve from 0.86649\n",
"Epoch 116/3000\n",
" - 39s - loss: 0.4082 - acc: 0.9417 - val_loss: 1.0239 - val_acc: 0.8357\n",
"\n",
"Epoch 00116: val_acc did not improve from 0.86649\n",
"Epoch 117/3000\n",
" - 39s - loss: 0.4483 - acc: 0.9348 - val_loss: 0.8706 - val_acc: 0.8505\n",
"\n",
"Epoch 00117: val_acc did not improve from 0.86649\n",
"Epoch 118/3000\n",
" - 40s - loss: 0.4208 - acc: 0.9408 - val_loss: 0.9734 - val_acc: 0.8388\n",
"\n",
"Epoch 00118: val_acc did not improve from 0.86649\n",
"Epoch 119/3000\n",
" - 38s - loss: 0.4357 - acc: 0.9364 - val_loss: 0.9709 - val_acc: 0.8392\n",
"\n",
"Epoch 00119: val_acc did not improve from 0.86649\n",
"Epoch 120/3000\n",
" - 39s - loss: 0.4519 - acc: 0.9331 - val_loss: 0.9216 - val_acc: 0.8435\n",
"\n",
"Epoch 00120: val_acc did not improve from 0.86649\n",
"Epoch 121/3000\n",
" - 39s - loss: 0.4216 - acc: 0.9413 - val_loss: 0.9375 - val_acc: 0.8490\n",
"\n",
"Epoch 00121: val_acc did not improve from 0.86649\n",
"Epoch 122/3000\n",
" - 39s - loss: 0.4187 - acc: 0.9413 - val_loss: 0.8751 - val_acc: 0.8513\n",
"\n",
"Epoch 00122: val_acc did not improve from 0.86649\n",
"Epoch 123/3000\n",
" - 39s - loss: 0.4169 - acc: 0.9469 - val_loss: 0.8852 - val_acc: 0.8459\n",
"\n",
"Epoch 00123: val_acc did not improve from 0.86649\n",
"Epoch 124/3000\n",
" - 39s - loss: 0.4069 - acc: 0.9461 - val_loss: 0.9012 - val_acc: 0.8443\n",
"\n",
"Epoch 00124: val_acc did not improve from 0.86649\n",
"Epoch 125/3000\n",
" - 39s - loss: 0.4158 - acc: 0.9405 - val_loss: 0.8521 - val_acc: 0.8571\n",
"\n",
"Epoch 00125: val_acc did not improve from 0.86649\n",
"Epoch 126/3000\n",
" - 39s - loss: 0.4191 - acc: 0.9414 - val_loss: 0.9001 - val_acc: 0.8521\n",
"\n",
"Epoch 00126: val_acc did not improve from 0.86649\n",
"Epoch 127/3000\n",
" - 39s - loss: 0.3975 - acc: 0.9476 - val_loss: 0.9316 - val_acc: 0.8501\n",
"\n",
"Epoch 00127: val_acc did not improve from 0.86649\n",
"Epoch 128/3000\n",
" - 39s - loss: 0.4252 - acc: 0.9402 - val_loss: 0.8905 - val_acc: 0.8482\n",
"\n",
"Epoch 00128: val_acc did not improve from 0.86649\n",
"Epoch 129/3000\n",
" - 39s - loss: 0.4259 - acc: 0.9392 - val_loss: 0.8737 - val_acc: 0.8560\n",
"\n",
"Epoch 00129: val_acc did not improve from 0.86649\n",
"Epoch 130/3000\n",
" - 39s - loss: 0.4075 - acc: 0.9441 - val_loss: 0.9467 - val_acc: 0.8381\n",
"\n",
"Epoch 00130: val_acc did not improve from 0.86649\n",
"Epoch 131/3000\n",
" - 39s - loss: 0.3948 - acc: 0.9472 - val_loss: 0.8441 - val_acc: 0.8540\n",
"\n",
"Epoch 00131: val_acc did not improve from 0.86649\n",
"Epoch 132/3000\n",
" - 39s - loss: 0.4289 - acc: 0.9410 - val_loss: 0.9384 - val_acc: 0.8396\n",
"\n",
"Epoch 00132: val_acc did not improve from 0.86649\n",
"Epoch 133/3000\n",
" - 39s - loss: 0.4148 - acc: 0.9431 - val_loss: 0.9680 - val_acc: 0.8369\n",
"\n",
"Epoch 00133: val_acc did not improve from 0.86649\n",
"Epoch 134/3000\n",
" - 39s - loss: 0.4544 - acc: 0.9387 - val_loss: 0.9413 - val_acc: 0.8455\n",
"\n",
"Epoch 00134: val_acc did not improve from 0.86649\n",
"\n",
"Epoch 00134: ReduceLROnPlateau reducing learning rate to 9.02499959920533e-05.\n",
"Epoch 135/3000\n",
" - 39s - loss: 0.4370 - acc: 0.9402 - val_loss: 0.9007 - val_acc: 0.8517\n",
"\n",
"Epoch 00135: val_acc did not improve from 0.86649\n",
"Epoch 136/3000\n",
" - 38s - loss: 0.3923 - acc: 0.9472 - val_loss: 0.8570 - val_acc: 0.8525\n",
"\n",
"Epoch 00136: val_acc did not improve from 0.86649\n",
"Epoch 137/3000\n",
" - 39s - loss: 0.3934 - acc: 0.9450 - val_loss: 0.9029 - val_acc: 0.8529\n",
"\n",
"Epoch 00137: val_acc did not improve from 0.86649\n",
"Epoch 138/3000\n",
" - 39s - loss: 0.3801 - acc: 0.9508 - val_loss: 0.9045 - val_acc: 0.8568\n",
"\n",
"Epoch 00138: val_acc did not improve from 0.86649\n",
"Epoch 139/3000\n",
" - 40s - loss: 0.4030 - acc: 0.9454 - val_loss: 0.9157 - val_acc: 0.8427\n",
"\n",
"Epoch 00139: val_acc did not improve from 0.86649\n",
"Epoch 140/3000\n",
" - 39s - loss: 0.3966 - acc: 0.9485 - val_loss: 0.9957 - val_acc: 0.8342\n",
"\n",
"Epoch 00140: val_acc did not improve from 0.86649\n",
"Epoch 141/3000\n",
" - 39s - loss: 0.4135 - acc: 0.9443 - val_loss: 0.9842 - val_acc: 0.8346\n",
"\n",
"Epoch 00141: val_acc did not improve from 0.86649\n",
"Epoch 142/3000\n",
" - 39s - loss: 0.3960 - acc: 0.9490 - val_loss: 0.9163 - val_acc: 0.8497\n",
"\n",
"Epoch 00142: val_acc did not improve from 0.86649\n",
"Epoch 143/3000\n",
" - 39s - loss: 0.4101 - acc: 0.9410 - val_loss: 0.8855 - val_acc: 0.8544\n",
"\n",
"Epoch 00143: val_acc did not improve from 0.86649\n",
"Epoch 144/3000\n",
" - 39s - loss: 0.3965 - acc: 0.9509 - val_loss: 1.0028 - val_acc: 0.8303\n",
"\n",
"Epoch 00144: val_acc did not improve from 0.86649\n",
"Epoch 145/3000\n",
" - 40s - loss: 0.4101 - acc: 0.9413 - val_loss: 0.9372 - val_acc: 0.8400\n",
"\n",
"Epoch 00145: val_acc did not improve from 0.86649\n",
"Epoch 146/3000\n",
" - 39s - loss: 0.4004 - acc: 0.9444 - val_loss: 0.9043 - val_acc: 0.8474\n",
"\n",
"Epoch 00146: val_acc did not improve from 0.86649\n",
"Epoch 147/3000\n",
" - 39s - loss: 0.4090 - acc: 0.9440 - val_loss: 0.8569 - val_acc: 0.8509\n",
"\n",
"Epoch 00147: val_acc did not improve from 0.86649\n",
"Epoch 148/3000\n",
" - 39s - loss: 0.4181 - acc: 0.9431 - val_loss: 0.9812 - val_acc: 0.8334\n",
"\n",
"Epoch 00148: val_acc did not improve from 0.86649\n",
"Epoch 149/3000\n",
" - 39s - loss: 0.4085 - acc: 0.9450 - val_loss: 0.8226 - val_acc: 0.8564\n",
"\n",
"Epoch 00149: val_acc did not improve from 0.86649\n",
"Epoch 150/3000\n",
" - 39s - loss: 0.4201 - acc: 0.9411 - val_loss: 0.9429 - val_acc: 0.8424\n",
"\n",
"Epoch 00150: val_acc did not improve from 0.86649\n",
"Epoch 151/3000\n",
" - 39s - loss: 0.4101 - acc: 0.9428 - val_loss: 0.8741 - val_acc: 0.8486\n",
"\n",
"Epoch 00151: val_acc did not improve from 0.86649\n",
"Epoch 152/3000\n",
" - 39s - loss: 0.4096 - acc: 0.9438 - val_loss: 0.8906 - val_acc: 0.8533\n",
"\n",
"Epoch 00152: val_acc did not improve from 0.86649\n",
"Epoch 153/3000\n",
" - 39s - loss: 0.4016 - acc: 0.9444 - val_loss: 0.9297 - val_acc: 0.8497\n",
"\n",
"Epoch 00153: val_acc did not improve from 0.86649\n",
"Epoch 154/3000\n",
" - 40s - loss: 0.4003 - acc: 0.9473 - val_loss: 1.0061 - val_acc: 0.8424\n",
"\n",
"Epoch 00154: val_acc did not improve from 0.86649\n",
"Epoch 155/3000\n",
" - 39s - loss: 0.4044 - acc: 0.9454 - val_loss: 1.0106 - val_acc: 0.8412\n",
"\n",
"Epoch 00155: val_acc did not improve from 0.86649\n",
"Epoch 156/3000\n",
" - 39s - loss: 0.4122 - acc: 0.9414 - val_loss: 0.9389 - val_acc: 0.8443\n",
"\n",
"Epoch 00156: val_acc did not improve from 0.86649\n",
"Epoch 157/3000\n",
" - 39s - loss: 0.3976 - acc: 0.9446 - val_loss: 0.9240 - val_acc: 0.8509\n",
"\n",
"Epoch 00157: val_acc did not improve from 0.86649\n",
"Epoch 158/3000\n",
" - 39s - loss: 0.3838 - acc: 0.9523 - val_loss: 0.9041 - val_acc: 0.8494\n",
"\n",
"Epoch 00158: val_acc did not improve from 0.86649\n",
"Epoch 159/3000\n",
" - 39s - loss: 0.4119 - acc: 0.9441 - val_loss: 0.9310 - val_acc: 0.8486\n",
"\n",
"Epoch 00159: val_acc did not improve from 0.86649\n",
"Epoch 160/3000\n",
" - 39s - loss: 0.3936 - acc: 0.9464 - val_loss: 0.9601 - val_acc: 0.8431\n",
"\n",
"Epoch 00160: val_acc did not improve from 0.86649\n",
"Epoch 161/3000\n",
" - 39s - loss: 0.4161 - acc: 0.9435 - val_loss: 0.9179 - val_acc: 0.8466\n",
"\n",
"Epoch 00161: val_acc did not improve from 0.86649\n",
"Epoch 162/3000\n",
" - 39s - loss: 0.4000 - acc: 0.9458 - val_loss: 0.9565 - val_acc: 0.8501\n",
"\n",
"Epoch 00162: val_acc did not improve from 0.86649\n",
"Epoch 163/3000\n",
" - 39s - loss: 0.4164 - acc: 0.9419 - val_loss: 1.0898 - val_acc: 0.8272\n",
"\n",
"Epoch 00163: val_acc did not improve from 0.86649\n",
"Epoch 164/3000\n",
" - 39s - loss: 0.4255 - acc: 0.9398 - val_loss: 0.9603 - val_acc: 0.8381\n",
"\n",
"Epoch 00164: val_acc did not improve from 0.86649\n",
"\n",
"Epoch 00164: ReduceLROnPlateau reducing learning rate to 8.573749619245064e-05.\n",
"Epoch 165/3000\n",
" - 39s - loss: 0.4097 - acc: 0.9446 - val_loss: 0.9117 - val_acc: 0.8427\n",
"\n",
"Epoch 00165: val_acc did not improve from 0.86649\n",
"Epoch 166/3000\n",
" - 39s - loss: 0.3816 - acc: 0.9491 - val_loss: 0.9233 - val_acc: 0.8455\n",
"\n",
"Epoch 00166: val_acc did not improve from 0.86649\n",
"Epoch 167/3000\n",
" - 39s - loss: 0.4109 - acc: 0.9426 - val_loss: 0.9021 - val_acc: 0.8517\n",
"\n",
"Epoch 00167: val_acc did not improve from 0.86649\n",
"Epoch 168/3000\n",
" - 40s - loss: 0.4019 - acc: 0.9446 - val_loss: 0.9406 - val_acc: 0.8462\n",
"\n",
"Epoch 00168: val_acc did not improve from 0.86649\n",
"Epoch 169/3000\n",
" - 40s - loss: 0.3951 - acc: 0.9464 - val_loss: 0.9533 - val_acc: 0.8466\n",
"\n",
"Epoch 00169: val_acc did not improve from 0.86649\n",
"Epoch 170/3000\n",
" - 39s - loss: 0.3935 - acc: 0.9485 - val_loss: 0.9551 - val_acc: 0.8427\n",
"\n",
"Epoch 00170: val_acc did not improve from 0.86649\n",
"Epoch 171/3000\n",
" - 39s - loss: 0.3931 - acc: 0.9496 - val_loss: 0.9058 - val_acc: 0.8424\n",
"\n",
"Epoch 00171: val_acc did not improve from 0.86649\n",
"Epoch 172/3000\n",
" - 40s - loss: 0.3885 - acc: 0.9467 - val_loss: 0.9725 - val_acc: 0.8400\n",
"\n",
"Epoch 00172: val_acc did not improve from 0.86649\n",
"Epoch 173/3000\n",
" - 39s - loss: 0.3981 - acc: 0.9481 - val_loss: 0.9014 - val_acc: 0.8501\n",
"\n",
"Epoch 00173: val_acc did not improve from 0.86649\n",
"Epoch 174/3000\n",
" - 40s - loss: 0.4189 - acc: 0.9460 - val_loss: 0.9159 - val_acc: 0.8490\n",
"\n",
"Epoch 00174: val_acc did not improve from 0.86649\n",
"Epoch 175/3000\n",
" - 39s - loss: 0.4159 - acc: 0.9380 - val_loss: 0.9044 - val_acc: 0.8490\n",
"\n",
"Epoch 00175: val_acc did not improve from 0.86649\n",
"Epoch 176/3000\n",
" - 39s - loss: 0.3944 - acc: 0.9482 - val_loss: 0.9412 - val_acc: 0.8385\n",
"\n",
"Epoch 00176: val_acc did not improve from 0.86649\n",
"Epoch 177/3000\n",
" - 39s - loss: 0.3890 - acc: 0.9502 - val_loss: 0.9233 - val_acc: 0.8451\n",
"\n",
"Epoch 00177: val_acc did not improve from 0.86649\n",
"Epoch 178/3000\n",
" - 39s - loss: 0.3900 - acc: 0.9515 - val_loss: 0.8865 - val_acc: 0.8501\n",
"\n",
"Epoch 00178: val_acc did not improve from 0.86649\n",
"Epoch 179/3000\n",
" - 40s - loss: 0.4063 - acc: 0.9455 - val_loss: 0.9673 - val_acc: 0.8462\n",
"\n",
"Epoch 00179: val_acc did not improve from 0.86649\n",
"Epoch 180/3000\n",
" - 39s - loss: 0.3787 - acc: 0.9499 - val_loss: 0.9494 - val_acc: 0.8462\n",
"\n",
"Epoch 00180: val_acc did not improve from 0.86649\n",
"Epoch 181/3000\n",
" - 40s - loss: 0.4018 - acc: 0.9461 - val_loss: 0.9874 - val_acc: 0.8462\n",
"\n",
"Epoch 00181: val_acc did not improve from 0.86649\n",
"Epoch 182/3000\n",
" - 39s - loss: 0.3996 - acc: 0.9493 - val_loss: 0.8876 - val_acc: 0.8513\n",
"\n",
"Epoch 00182: val_acc did not improve from 0.86649\n",
"Epoch 183/3000\n",
" - 40s - loss: 0.4109 - acc: 0.9461 - val_loss: 0.8838 - val_acc: 0.8525\n",
"\n",
"Epoch 00183: val_acc did not improve from 0.86649\n",
"Epoch 184/3000\n",
" - 39s - loss: 0.3980 - acc: 0.9457 - val_loss: 0.8478 - val_acc: 0.8587\n",
"\n",
"Epoch 00184: val_acc did not improve from 0.86649\n",
"Epoch 185/3000\n",
" - 40s - loss: 0.3885 - acc: 0.9467 - val_loss: 0.9210 - val_acc: 0.8478\n",
"\n",
"Epoch 00185: val_acc did not improve from 0.86649\n",
"Epoch 186/3000\n",
" - 40s - loss: 0.3894 - acc: 0.9499 - val_loss: 0.9790 - val_acc: 0.8431\n",
"\n",
"Epoch 00186: val_acc did not improve from 0.86649\n",
"Epoch 187/3000\n",
" - 39s - loss: 0.3871 - acc: 0.9494 - val_loss: 0.9039 - val_acc: 0.8501\n",
"\n",
"Epoch 00187: val_acc did not improve from 0.86649\n",
"Epoch 188/3000\n",
" - 40s - loss: 0.3955 - acc: 0.9470 - val_loss: 0.9084 - val_acc: 0.8501\n",
"\n",
"Epoch 00188: val_acc did not improve from 0.86649\n",
"Epoch 189/3000\n",
" - 40s - loss: 0.4062 - acc: 0.9446 - val_loss: 0.9570 - val_acc: 0.8318\n",
"\n",
"Epoch 00189: val_acc did not improve from 0.86649\n",
"Epoch 190/3000\n",
" - 39s - loss: 0.3884 - acc: 0.9469 - val_loss: 0.9881 - val_acc: 0.8400\n",
"\n",
"Epoch 00190: val_acc did not improve from 0.86649\n",
"Epoch 191/3000\n",
" - 39s - loss: 0.4025 - acc: 0.9472 - val_loss: 0.9180 - val_acc: 0.8505\n",
"\n",
"Epoch 00191: val_acc did not improve from 0.86649\n",
"Epoch 192/3000\n",
" - 39s - loss: 0.3804 - acc: 0.9502 - val_loss: 0.9212 - val_acc: 0.8536\n",
"\n",
"Epoch 00192: val_acc did not improve from 0.86649\n",
"Epoch 193/3000\n",
" - 39s - loss: 0.4011 - acc: 0.9416 - val_loss: 1.0128 - val_acc: 0.8385\n",
"\n",
"Epoch 00193: val_acc did not improve from 0.86649\n",
"Epoch 194/3000\n",
" - 40s - loss: 0.4137 - acc: 0.9426 - val_loss: 0.9164 - val_acc: 0.8435\n",
"\n",
"Epoch 00194: val_acc did not improve from 0.86649\n",
"\n",
"Epoch 00194: ReduceLROnPlateau reducing learning rate to 8.145062311086804e-05.\n",
"Epoch 195/3000\n",
" - 39s - loss: 0.3887 - acc: 0.9482 - val_loss: 0.9217 - val_acc: 0.8505\n",
"\n",
"Epoch 00195: val_acc did not improve from 0.86649\n",
"Epoch 196/3000\n",
" - 40s - loss: 0.3835 - acc: 0.9490 - val_loss: 0.9337 - val_acc: 0.8412\n",
"\n",
"Epoch 00196: val_acc did not improve from 0.86649\n",
"Epoch 197/3000\n",
" - 39s - loss: 0.3901 - acc: 0.9478 - val_loss: 0.9818 - val_acc: 0.8353\n",
"\n",
"Epoch 00197: val_acc did not improve from 0.86649\n",
"Epoch 198/3000\n",
" - 39s - loss: 0.4046 - acc: 0.9461 - val_loss: 1.0099 - val_acc: 0.8326\n",
"\n",
"Epoch 00198: val_acc did not improve from 0.86649\n",
"Epoch 199/3000\n",
" - 39s - loss: 0.4022 - acc: 0.9463 - val_loss: 0.8810 - val_acc: 0.8459\n",
"\n",
"Epoch 00199: val_acc did not improve from 0.86649\n",
"Epoch 200/3000\n",
" - 39s - loss: 0.3884 - acc: 0.9505 - val_loss: 0.8836 - val_acc: 0.8540\n",
"\n",
"Epoch 00200: val_acc did not improve from 0.86649\n",
"Epoch 201/3000\n",
" - 39s - loss: 0.4034 - acc: 0.9461 - val_loss: 0.9049 - val_acc: 0.8400\n",
"\n",
"Epoch 00201: val_acc did not improve from 0.86649\n",
"Epoch 202/3000\n",
" - 39s - loss: 0.3845 - acc: 0.9475 - val_loss: 0.9071 - val_acc: 0.8478\n",
"\n",
"Epoch 00202: val_acc did not improve from 0.86649\n",
"Epoch 203/3000\n",
" - 40s - loss: 0.3795 - acc: 0.9503 - val_loss: 0.9272 - val_acc: 0.8443\n",
"\n",
"Epoch 00203: val_acc did not improve from 0.86649\n",
"Epoch 204/3000\n",
" - 39s - loss: 0.3677 - acc: 0.9521 - val_loss: 0.8768 - val_acc: 0.8447\n",
"\n",
"Epoch 00204: val_acc did not improve from 0.86649\n",
"Epoch 205/3000\n",
" - 40s - loss: 0.3786 - acc: 0.9527 - val_loss: 0.9458 - val_acc: 0.8365\n",
"\n",
"Epoch 00205: val_acc did not improve from 0.86649\n",
"Epoch 206/3000\n",
" - 39s - loss: 0.4036 - acc: 0.9481 - val_loss: 0.8992 - val_acc: 0.8533\n",
"\n",
"Epoch 00206: val_acc did not improve from 0.86649\n",
"Epoch 207/3000\n",
" - 39s - loss: 0.3865 - acc: 0.9467 - val_loss: 0.8995 - val_acc: 0.8439\n",
"\n",
"Epoch 00207: val_acc did not improve from 0.86649\n",
"Epoch 208/3000\n",
" - 39s - loss: 0.3734 - acc: 0.9532 - val_loss: 0.9297 - val_acc: 0.8478\n",
"\n",
"Epoch 00208: val_acc did not improve from 0.86649\n",
"Epoch 209/3000\n",
" - 40s - loss: 0.3811 - acc: 0.9517 - val_loss: 0.8610 - val_acc: 0.8564\n",
"\n",
"Epoch 00209: val_acc did not improve from 0.86649\n",
"Epoch 210/3000\n",
" - 39s - loss: 0.3746 - acc: 0.9517 - val_loss: 0.9199 - val_acc: 0.8431\n",
"\n",
"Epoch 00210: val_acc did not improve from 0.86649\n",
"Epoch 211/3000\n",
" - 39s - loss: 0.3745 - acc: 0.9505 - val_loss: 0.9545 - val_acc: 0.8361\n",
"\n",
"Epoch 00211: val_acc did not improve from 0.86649\n",
"Epoch 212/3000\n",
" - 40s - loss: 0.3858 - acc: 0.9478 - val_loss: 0.8884 - val_acc: 0.8505\n",
"\n",
"Epoch 00212: val_acc did not improve from 0.86649\n",
"Epoch 213/3000\n",
" - 39s - loss: 0.3898 - acc: 0.9491 - val_loss: 0.9675 - val_acc: 0.8466\n",
"\n",
"Epoch 00213: val_acc did not improve from 0.86649\n",
"Epoch 214/3000\n",
" - 39s - loss: 0.4043 - acc: 0.9485 - val_loss: 0.8941 - val_acc: 0.8525\n",
"\n",
"Epoch 00214: val_acc did not improve from 0.86649\n",
"Epoch 215/3000\n",
" - 39s - loss: 0.3720 - acc: 0.9538 - val_loss: 0.9408 - val_acc: 0.8443\n",
"\n",
"Epoch 00215: val_acc did not improve from 0.86649\n",
"Epoch 216/3000\n",
" - 39s - loss: 0.3596 - acc: 0.9541 - val_loss: 0.9415 - val_acc: 0.8462\n",
"\n",
"Epoch 00216: val_acc did not improve from 0.86649\n",
"Epoch 217/3000\n",
" - 39s - loss: 0.3797 - acc: 0.9508 - val_loss: 0.9347 - val_acc: 0.8462\n",
"\n",
"Epoch 00217: val_acc did not improve from 0.86649\n",
"Epoch 218/3000\n",
" - 39s - loss: 0.3925 - acc: 0.9473 - val_loss: 0.8918 - val_acc: 0.8540\n",
"\n",
"Epoch 00218: val_acc did not improve from 0.86649\n",
"Epoch 219/3000\n",
" - 39s - loss: 0.3767 - acc: 0.9488 - val_loss: 0.8936 - val_acc: 0.8529\n",
"\n",
"Epoch 00219: val_acc did not improve from 0.86649\n",
"Epoch 220/3000\n",
" - 40s - loss: 0.3981 - acc: 0.9444 - val_loss: 0.9101 - val_acc: 0.8529\n",
"\n",
"Epoch 00220: val_acc did not improve from 0.86649\n",
"Epoch 221/3000\n",
" - 39s - loss: 0.3935 - acc: 0.9484 - val_loss: 0.9340 - val_acc: 0.8517\n",
"\n",
"Epoch 00221: val_acc did not improve from 0.86649\n",
"Epoch 222/3000\n",
" - 39s - loss: 0.3827 - acc: 0.9454 - val_loss: 0.9111 - val_acc: 0.8505\n",
"\n",
"Epoch 00224: val_acc did not improve from 0.86649\n",
"\n",
"Epoch 00224: ReduceLROnPlateau reducing learning rate to 7.737808919046074e-05.\n",
"Epoch 225/3000\n",
" - 39s - loss: 0.4087 - acc: 0.9426 - val_loss: 0.9580 - val_acc: 0.8396\n",
"\n",
"Epoch 00225: val_acc did not improve from 0.86649\n",
"Epoch 226/3000\n",
" - 40s - loss: 0.3773 - acc: 0.9485 - val_loss: 0.9078 - val_acc: 0.8490\n",
"\n",
"Epoch 00226: val_acc did not improve from 0.86649\n",
"Epoch 227/3000\n",
" - 40s - loss: 0.3660 - acc: 0.9530 - val_loss: 0.8983 - val_acc: 0.8451\n",
"\n",
"Epoch 00227: val_acc did not improve from 0.86649\n",
"Epoch 228/3000\n",
" - 39s - loss: 0.3804 - acc: 0.9490 - val_loss: 0.8204 - val_acc: 0.8618\n",
"\n",
"Epoch 00228: val_acc did not improve from 0.86649\n",
"Epoch 229/3000\n",
" - 39s - loss: 0.3861 - acc: 0.9479 - val_loss: 0.8979 - val_acc: 0.8466\n",
"\n",
"Epoch 00229: val_acc did not improve from 0.86649\n",
"Epoch 230/3000\n",
" - 39s - loss: 0.4036 - acc: 0.9449 - val_loss: 0.8535 - val_acc: 0.8548\n",
"\n",
"Epoch 00230: val_acc did not improve from 0.86649\n",
"Epoch 231/3000\n",
" - 39s - loss: 0.3948 - acc: 0.9464 - val_loss: 0.9545 - val_acc: 0.8501\n",
"\n",
"Epoch 00231: val_acc did not improve from 0.86649\n",
"Epoch 232/3000\n",
" - 39s - loss: 0.3763 - acc: 0.9517 - val_loss: 1.0186 - val_acc: 0.8338\n",
"\n",
"Epoch 00232: val_acc did not improve from 0.86649\n",
"Epoch 233/3000\n",
" - 39s - loss: 0.3607 - acc: 0.9521 - val_loss: 0.9043 - val_acc: 0.8482\n",
"\n",
"Epoch 00233: val_acc did not improve from 0.86649\n",
"Epoch 234/3000\n",
" - 40s - loss: 0.3710 - acc: 0.9505 - val_loss: 0.9583 - val_acc: 0.8466\n",
"\n",
"Epoch 00234: val_acc did not improve from 0.86649\n",
"Epoch 235/3000\n",
" - 39s - loss: 0.3781 - acc: 0.9488 - val_loss: 0.8799 - val_acc: 0.8521\n",
"\n",
"Epoch 00235: val_acc did not improve from 0.86649\n",
"Epoch 236/3000\n",
" - 40s - loss: 0.3754 - acc: 0.9534 - val_loss: 0.9165 - val_acc: 0.8490\n",
"\n",
"Epoch 00236: val_acc did not improve from 0.86649\n",
"Epoch 237/3000\n",
" - 40s - loss: 0.3724 - acc: 0.9540 - val_loss: 0.9543 - val_acc: 0.8388\n",
"\n",
"Epoch 00237: val_acc did not improve from 0.86649\n",
"Epoch 238/3000\n",
" - 40s - loss: 0.3855 - acc: 0.9484 - val_loss: 0.9242 - val_acc: 0.8474\n",
"\n",
"Epoch 00238: val_acc did not improve from 0.86649\n",
"Epoch 239/3000\n",
" - 39s - loss: 0.3637 - acc: 0.9529 - val_loss: 0.9195 - val_acc: 0.8521\n",
"\n",
"Epoch 00239: val_acc did not improve from 0.86649\n",
"Epoch 240/3000\n",
" - 39s - loss: 0.3883 - acc: 0.9500 - val_loss: 0.9836 - val_acc: 0.8357\n",
"\n",
"Epoch 00240: val_acc did not improve from 0.86649\n",
"Epoch 241/3000\n",
" - 40s - loss: 0.3729 - acc: 0.9505 - val_loss: 0.8994 - val_acc: 0.8513\n",
"\n",
"Epoch 00241: val_acc did not improve from 0.86649\n",
"Epoch 242/3000\n",
" - 39s - loss: 0.3858 - acc: 0.9478 - val_loss: 0.8928 - val_acc: 0.8466\n",
"\n",
"Epoch 00242: val_acc did not improve from 0.86649\n",
"Epoch 243/3000\n",
" - 40s - loss: 0.3655 - acc: 0.9547 - val_loss: 0.8937 - val_acc: 0.8533\n",
"\n",
"Epoch 00243: val_acc did not improve from 0.86649\n",
"Epoch 244/3000\n",
" - 39s - loss: 0.3747 - acc: 0.9490 - val_loss: 0.9015 - val_acc: 0.8494\n",
"\n",
"Epoch 00244: val_acc did not improve from 0.86649\n",
"Epoch 245/3000\n",
" - 39s - loss: 0.3713 - acc: 0.9499 - val_loss: 0.9938 - val_acc: 0.8482\n",
"\n",
"Epoch 00245: val_acc did not improve from 0.86649\n",
"Epoch 246/3000\n",
" - 39s - loss: 0.3796 - acc: 0.9500 - val_loss: 0.9559 - val_acc: 0.8431\n",
"\n",
"Epoch 00246: val_acc did not improve from 0.86649\n",
"Epoch 247/3000\n",
" - 40s - loss: 0.3666 - acc: 0.9523 - val_loss: 0.9361 - val_acc: 0.8385\n",
"\n",
"Epoch 00247: val_acc did not improve from 0.86649\n",
"Epoch 248/3000\n",
" - 40s - loss: 0.3775 - acc: 0.9515 - val_loss: 0.8607 - val_acc: 0.8564\n",
"\n",
"Epoch 00248: val_acc did not improve from 0.86649\n",
"Epoch 249/3000\n",
" - 39s - loss: 0.3624 - acc: 0.9541 - val_loss: 0.9073 - val_acc: 0.8486\n",
"\n",
"Epoch 00249: val_acc did not improve from 0.86649\n",
"Epoch 250/3000\n",
" - 39s - loss: 0.3687 - acc: 0.9524 - val_loss: 0.9094 - val_acc: 0.8533\n",
"\n",
"Epoch 00250: val_acc did not improve from 0.86649\n",
"Epoch 251/3000\n",
" - 39s - loss: 0.3681 - acc: 0.9534 - val_loss: 0.9190 - val_acc: 0.8427\n",
"\n",
"Epoch 00251: val_acc did not improve from 0.86649\n",
"Epoch 252/3000\n",
" - 39s - loss: 0.3672 - acc: 0.9508 - val_loss: 0.8954 - val_acc: 0.8525\n",
"\n",
"Epoch 00252: val_acc did not improve from 0.86649\n",
"Epoch 253/3000\n",
" - 40s - loss: 0.3870 - acc: 0.9473 - val_loss: 0.8745 - val_acc: 0.8509\n",
"\n",
"Epoch 00253: val_acc did not improve from 0.86649\n",
"Epoch 254/3000\n",
" - 40s - loss: 0.3811 - acc: 0.9481 - val_loss: 0.9155 - val_acc: 0.8482\n",
"\n",
"Epoch 00254: val_acc did not improve from 0.86649\n",
"\n",
"Epoch 00254: ReduceLROnPlateau reducing learning rate to 7.350918749580159e-05.\n",
"Epoch 255/3000\n",
" - 40s - loss: 0.3656 - acc: 0.9534 - val_loss: 0.8551 - val_acc: 0.8536\n",
"\n",
"Epoch 00255: val_acc did not improve from 0.86649\n",
"Epoch 256/3000\n",
" - 39s - loss: 0.3653 - acc: 0.9514 - val_loss: 0.9610 - val_acc: 0.8338\n",
"\n",
"Epoch 00256: val_acc did not improve from 0.86649\n",
"Epoch 257/3000\n",
" - 39s - loss: 0.3980 - acc: 0.9463 - val_loss: 1.0120 - val_acc: 0.8295\n",
"\n",
"Epoch 00257: val_acc did not improve from 0.86649\n",
"Epoch 258/3000\n",
" - 40s - loss: 0.3869 - acc: 0.9481 - val_loss: 0.9067 - val_acc: 0.8439\n",
"\n",
"Epoch 00258: val_acc did not improve from 0.86649\n",
"Epoch 259/3000\n",
" - 39s - loss: 0.4147 - acc: 0.9423 - val_loss: 0.9569 - val_acc: 0.8416\n",
"\n",
"Epoch 00259: val_acc did not improve from 0.86649\n",
"Epoch 260/3000\n",
" - 39s - loss: 0.3765 - acc: 0.9493 - val_loss: 0.9149 - val_acc: 0.8482\n",
"\n",
"Epoch 00260: val_acc did not improve from 0.86649\n",
"Epoch 261/3000\n",
" - 39s - loss: 0.3723 - acc: 0.9494 - val_loss: 0.9884 - val_acc: 0.8350\n",
"\n",
"Epoch 00261: val_acc did not improve from 0.86649\n",
"Epoch 262/3000\n",
" - 39s - loss: 0.3825 - acc: 0.9463 - val_loss: 0.9140 - val_acc: 0.8482\n",
"\n",
"Epoch 00262: val_acc did not improve from 0.86649\n",
"Epoch 263/3000\n",
" - 40s - loss: 0.3643 - acc: 0.9535 - val_loss: 0.9832 - val_acc: 0.8369\n",
"\n",
"Epoch 00263: val_acc did not improve from 0.86649\n",
"Epoch 264/3000\n",
" - 39s - loss: 0.3596 - acc: 0.9526 - val_loss: 0.9553 - val_acc: 0.8447\n",
"\n",
"Epoch 00264: val_acc did not improve from 0.86649\n",
"Epoch 265/3000\n",
" - 39s - loss: 0.3603 - acc: 0.9532 - val_loss: 0.9267 - val_acc: 0.8420\n",
"\n",
"Epoch 00265: val_acc did not improve from 0.86649\n",
"Epoch 266/3000\n",
" - 39s - loss: 0.3523 - acc: 0.9583 - val_loss: 0.9068 - val_acc: 0.8533\n",
"\n",
"Epoch 00266: val_acc did not improve from 0.86649\n",
"Epoch 267/3000\n",
" - 39s - loss: 0.3496 - acc: 0.9553 - val_loss: 0.8808 - val_acc: 0.8505\n",
"\n",
"Epoch 00267: val_acc did not improve from 0.86649\n",
"Epoch 268/3000\n",
" - 39s - loss: 0.3839 - acc: 0.9479 - val_loss: 0.8474 - val_acc: 0.8591\n",
"\n",
"Epoch 00268: val_acc did not improve from 0.86649\n",
"Epoch 269/3000\n",
" - 39s - loss: 0.3651 - acc: 0.9552 - val_loss: 0.9066 - val_acc: 0.8505\n",
"\n",
"Epoch 00269: val_acc did not improve from 0.86649\n",
"Epoch 270/3000\n",
" - 39s - loss: 0.3781 - acc: 0.9494 - val_loss: 0.9599 - val_acc: 0.8439\n",
"\n",
"Epoch 00270: val_acc did not improve from 0.86649\n",
"Epoch 271/3000\n",
" - 40s - loss: 0.3577 - acc: 0.9529 - val_loss: 0.9236 - val_acc: 0.8474\n",
"\n",
"Epoch 00271: val_acc did not improve from 0.86649\n",
"Epoch 272/3000\n",
" - 40s - loss: 0.3792 - acc: 0.9473 - val_loss: 0.9265 - val_acc: 0.8439\n",
"\n",
"Epoch 00272: val_acc did not improve from 0.86649\n",
"Epoch 273/3000\n",
" - 40s - loss: 0.3674 - acc: 0.9523 - val_loss: 0.9896 - val_acc: 0.8373\n",
"\n",
"Epoch 00273: val_acc did not improve from 0.86649\n",
"Epoch 274/3000\n",
" - 39s - loss: 0.3655 - acc: 0.9526 - val_loss: 0.8524 - val_acc: 0.8548\n",
"\n",
"Epoch 00274: val_acc did not improve from 0.86649\n",
"Epoch 275/3000\n",
" - 39s - loss: 0.3658 - acc: 0.9502 - val_loss: 0.9661 - val_acc: 0.8435\n",
"\n",
"Epoch 00275: val_acc did not improve from 0.86649\n",
"Epoch 276/3000\n",
" - 39s - loss: 0.3713 - acc: 0.9500 - val_loss: 0.8588 - val_acc: 0.8568\n",
"\n",
"Epoch 00276: val_acc did not improve from 0.86649\n",
"Epoch 277/3000\n",
" - 39s - loss: 0.3685 - acc: 0.9540 - val_loss: 0.9974 - val_acc: 0.8408\n",
"\n",
"Epoch 00277: val_acc did not improve from 0.86649\n",
"Epoch 278/3000\n",
" - 39s - loss: 0.3696 - acc: 0.9526 - val_loss: 0.8494 - val_acc: 0.8540\n",
"\n",
"Epoch 00278: val_acc did not improve from 0.86649\n",
"Epoch 279/3000\n",
" - 39s - loss: 0.3585 - acc: 0.9538 - val_loss: 0.9636 - val_acc: 0.8392\n",
"\n",
"Epoch 00279: val_acc did not improve from 0.86649\n",
"Epoch 280/3000\n",
" - 39s - loss: 0.3744 - acc: 0.9490 - val_loss: 0.8726 - val_acc: 0.8587\n",
"\n",
"Epoch 00280: val_acc did not improve from 0.86649\n",
"Epoch 281/3000\n",
" - 40s - loss: 0.3828 - acc: 0.9469 - val_loss: 0.9184 - val_acc: 0.8521\n",
"\n",
"Epoch 00281: val_acc did not improve from 0.86649\n",
"Epoch 282/3000\n",
" - 39s - loss: 0.3731 - acc: 0.9505 - val_loss: 0.8927 - val_acc: 0.8517\n",
"\n",
"Epoch 00282: val_acc did not improve from 0.86649\n",
"Epoch 283/3000\n",
" - 39s - loss: 0.3597 - acc: 0.9524 - val_loss: 0.8174 - val_acc: 0.8610\n",
"\n",
"Epoch 00283: val_acc did not improve from 0.86649\n",
"Epoch 284/3000\n",
" - 39s - loss: 0.3644 - acc: 0.9515 - val_loss: 0.9366 - val_acc: 0.8505\n",
"\n",
"Epoch 00284: val_acc did not improve from 0.86649\n",
"\n",
"Epoch 00284: ReduceLROnPlateau reducing learning rate to 6.983372950344346e-05.\n",
"Epoch 285/3000\n",
" - 38s - loss: 0.3537 - acc: 0.9524 - val_loss: 0.9509 - val_acc: 0.8474\n",
"\n",
"Epoch 00285: val_acc did not improve from 0.86649\n",
"Epoch 286/3000\n",
" - 39s - loss: 0.3539 - acc: 0.9556 - val_loss: 0.8748 - val_acc: 0.8603\n",
"\n",
"Epoch 00286: val_acc did not improve from 0.86649\n",
"Epoch 287/3000\n",
" - 40s - loss: 0.3665 - acc: 0.9512 - val_loss: 0.9176 - val_acc: 0.8509\n",
"\n",
"Epoch 00287: val_acc did not improve from 0.86649\n",
"Epoch 288/3000\n",
" - 39s - loss: 0.3592 - acc: 0.9523 - val_loss: 0.8962 - val_acc: 0.8536\n",
"\n",
"Epoch 00288: val_acc did not improve from 0.86649\n",
"Epoch 289/3000\n",
" - 39s - loss: 0.3485 - acc: 0.9558 - val_loss: 0.8369 - val_acc: 0.8587\n",
"\n",
"Epoch 00289: val_acc did not improve from 0.86649\n",
"Epoch 290/3000\n",
" - 39s - loss: 0.3433 - acc: 0.9537 - val_loss: 0.8916 - val_acc: 0.8575\n",
"\n",
"Epoch 00290: val_acc did not improve from 0.86649\n",
"Epoch 291/3000\n",
" - 39s - loss: 0.3764 - acc: 0.9499 - val_loss: 0.9457 - val_acc: 0.8439\n",
"\n",
"Epoch 00291: val_acc did not improve from 0.86649\n",
"Epoch 292/3000\n",
" - 39s - loss: 0.3546 - acc: 0.9543 - val_loss: 0.9729 - val_acc: 0.8435\n",
"\n",
"Epoch 00292: val_acc did not improve from 0.86649\n",
"Epoch 293/3000\n",
" - 40s - loss: 0.3677 - acc: 0.9505 - val_loss: 0.9273 - val_acc: 0.8494\n",
"\n",
"Epoch 00293: val_acc did not improve from 0.86649\n",
"Epoch 294/3000\n",
" - 39s - loss: 0.3894 - acc: 0.9470 - val_loss: 0.9159 - val_acc: 0.8505\n",
"\n",
"Epoch 00294: val_acc did not improve from 0.86649\n",
"Epoch 295/3000\n",
" - 40s - loss: 0.3607 - acc: 0.9532 - val_loss: 0.9351 - val_acc: 0.8462\n",
"\n",
"Epoch 00295: val_acc did not improve from 0.86649\n",
"Epoch 296/3000\n",
" - 39s - loss: 0.3785 - acc: 0.9485 - val_loss: 0.9046 - val_acc: 0.8513\n",
"\n",
"Epoch 00296: val_acc did not improve from 0.86649\n",
"Epoch 297/3000\n",
" - 39s - loss: 0.3575 - acc: 0.9547 - val_loss: 0.9030 - val_acc: 0.8501\n",
"\n",
"Epoch 00297: val_acc did not improve from 0.86649\n",
"Epoch 298/3000\n",
" - 40s - loss: 0.3679 - acc: 0.9521 - val_loss: 0.8527 - val_acc: 0.8583\n",
"\n",
"Epoch 00298: val_acc did not improve from 0.86649\n",
"Epoch 299/3000\n",
" - 40s - loss: 0.3658 - acc: 0.9527 - val_loss: 0.8763 - val_acc: 0.8560\n",
"\n",
"Epoch 00299: val_acc did not improve from 0.86649\n",
"Epoch 300/3000\n",
" - 39s - loss: 0.3501 - acc: 0.9561 - val_loss: 0.8783 - val_acc: 0.8513\n",
"\n",
"Epoch 00300: val_acc did not improve from 0.86649\n",
"Epoch 301/3000\n",
" - 40s - loss: 0.3594 - acc: 0.9538 - val_loss: 0.9492 - val_acc: 0.8443\n",
"\n",
"Epoch 00301: val_acc did not improve from 0.86649\n",
"Epoch 302/3000\n",
" - 40s - loss: 0.3566 - acc: 0.9541 - val_loss: 0.9000 - val_acc: 0.8568\n",
"\n",
"Epoch 00302: val_acc did not improve from 0.86649\n",
"Epoch 303/3000\n",
" - 39s - loss: 0.3498 - acc: 0.9567 - val_loss: 0.9092 - val_acc: 0.8486\n",
"\n",
"Epoch 00303: val_acc did not improve from 0.86649\n",
"Epoch 304/3000\n",
" - 39s - loss: 0.3646 - acc: 0.9534 - val_loss: 0.9032 - val_acc: 0.8494\n",
"\n",
"Epoch 00304: val_acc did not improve from 0.86649\n",
"Epoch 305/3000\n",
" - 39s - loss: 0.3381 - acc: 0.9588 - val_loss: 0.8957 - val_acc: 0.8571\n",
"\n",
"Epoch 00305: val_acc did not improve from 0.86649\n",
"Epoch 306/3000\n",
" - 39s - loss: 0.3567 - acc: 0.9541 - val_loss: 0.8587 - val_acc: 0.8595\n",
"\n",
"Epoch 00306: val_acc did not improve from 0.86649\n",
"Epoch 307/3000\n",
" - 39s - loss: 0.3696 - acc: 0.9493 - val_loss: 0.9091 - val_acc: 0.8525\n",
"\n",
"Epoch 00307: val_acc did not improve from 0.86649\n",
"Epoch 308/3000\n",
" - 39s - loss: 0.3529 - acc: 0.9530 - val_loss: 0.9425 - val_acc: 0.8478\n",
"\n",
"Epoch 00308: val_acc did not improve from 0.86649\n",
"Epoch 309/3000\n",
" - 39s - loss: 0.3585 - acc: 0.9540 - val_loss: 0.9658 - val_acc: 0.8307\n",
"\n",
"Epoch 00309: val_acc did not improve from 0.86649\n",
"Epoch 310/3000\n",
" - 39s - loss: 0.3701 - acc: 0.9482 - val_loss: 0.9474 - val_acc: 0.8404\n",
"\n",
"Epoch 00310: val_acc did not improve from 0.86649\n",
"Epoch 311/3000\n",
" - 39s - loss: 0.3728 - acc: 0.9485 - val_loss: 0.8468 - val_acc: 0.8560\n",
"\n",
"Epoch 00311: val_acc did not improve from 0.86649\n",
"Epoch 312/3000\n",
" - 40s - loss: 0.3428 - acc: 0.9547 - val_loss: 0.8516 - val_acc: 0.8583\n",
"\n",
"Epoch 00312: val_acc did not improve from 0.86649\n",
"Epoch 313/3000\n",
" - 39s - loss: 0.3676 - acc: 0.9511 - val_loss: 0.9169 - val_acc: 0.8486\n",
"\n",
"Epoch 00313: val_acc did not improve from 0.86649\n",
"Epoch 314/3000\n",
" - 39s - loss: 0.3761 - acc: 0.9464 - val_loss: 0.8943 - val_acc: 0.8517\n",
"\n",
"Epoch 00314: val_acc did not improve from 0.86649\n",
"\n",
"Epoch 00314: ReduceLROnPlateau reducing learning rate to 6.634204510191921e-05.\n",
"Epoch 315/3000\n",
" - 39s - loss: 0.3402 - acc: 0.9567 - val_loss: 0.9325 - val_acc: 0.8377\n",
"\n",
"Epoch 00315: val_acc did not improve from 0.86649\n",
"Epoch 316/3000\n",
" - 39s - loss: 0.3531 - acc: 0.9526 - val_loss: 0.8943 - val_acc: 0.8517\n",
"\n",
"Epoch 00316: val_acc did not improve from 0.86649\n",
"Epoch 317/3000\n",
" - 39s - loss: 0.3568 - acc: 0.9538 - val_loss: 0.9239 - val_acc: 0.8521\n",
"\n",
"Epoch 00317: val_acc did not improve from 0.86649\n",
"Epoch 318/3000\n",
" - 39s - loss: 0.3511 - acc: 0.9546 - val_loss: 0.8757 - val_acc: 0.8575\n",
"\n",
"Epoch 00318: val_acc did not improve from 0.86649\n",
"Epoch 319/3000\n",
" - 39s - loss: 0.3493 - acc: 0.9565 - val_loss: 0.8912 - val_acc: 0.8525\n",
"\n",
"Epoch 00319: val_acc did not improve from 0.86649\n",
"Epoch 320/3000\n",
" - 40s - loss: 0.3661 - acc: 0.9515 - val_loss: 0.9028 - val_acc: 0.8482\n",
"\n",
"Epoch 00320: val_acc did not improve from 0.86649\n",
"Epoch 321/3000\n",
" - 39s - loss: 0.3615 - acc: 0.9559 - val_loss: 0.9068 - val_acc: 0.8560\n",
"\n",
"Epoch 00321: val_acc did not improve from 0.86649\n",
"Epoch 322/3000\n",
" - 39s - loss: 0.3571 - acc: 0.9530 - val_loss: 0.9152 - val_acc: 0.8505\n",
"\n",
"Epoch 00322: val_acc did not improve from 0.86649\n",
"Epoch 323/3000\n",
" - 39s - loss: 0.3619 - acc: 0.9535 - val_loss: 0.8739 - val_acc: 0.8533\n",
"\n",
"Epoch 00323: val_acc did not improve from 0.86649\n",
"Epoch 324/3000\n",
" - 39s - loss: 0.3484 - acc: 0.9538 - val_loss: 0.9118 - val_acc: 0.8462\n",
"\n",
"Epoch 00324: val_acc did not improve from 0.86649\n",
"Epoch 325/3000\n",
" - 39s - loss: 0.3583 - acc: 0.9556 - val_loss: 0.9312 - val_acc: 0.8478\n",
"\n",
"Epoch 00325: val_acc did not improve from 0.86649\n",
"Epoch 326/3000\n",
" - 39s - loss: 0.3683 - acc: 0.9506 - val_loss: 0.9007 - val_acc: 0.8533\n",
"\n",
"Epoch 00326: val_acc did not improve from 0.86649\n",
"Epoch 327/3000\n",
" - 39s - loss: 0.3640 - acc: 0.9529 - val_loss: 0.9076 - val_acc: 0.8525\n",
"\n",
"Epoch 00327: val_acc did not improve from 0.86649\n",
"Epoch 328/3000\n",
" - 39s - loss: 0.3379 - acc: 0.9583 - val_loss: 0.8766 - val_acc: 0.8548\n",
"\n",
"Epoch 00328: val_acc did not improve from 0.86649\n",
"Epoch 329/3000\n",
" - 39s - loss: 0.3538 - acc: 0.9537 - val_loss: 0.9582 - val_acc: 0.8462\n",
"\n",
"Epoch 00329: val_acc did not improve from 0.86649\n",
"Epoch 330/3000\n",
" - 39s - loss: 0.3485 - acc: 0.9538 - val_loss: 0.8958 - val_acc: 0.8521\n",
"\n",
"Epoch 00330: val_acc did not improve from 0.86649\n",
"Epoch 331/3000\n",
" - 40s - loss: 0.3540 - acc: 0.9534 - val_loss: 0.8465 - val_acc: 0.8583\n",
"\n",
"Epoch 00331: val_acc did not improve from 0.86649\n",
"Epoch 332/3000\n",
" - 39s - loss: 0.3498 - acc: 0.9574 - val_loss: 0.9472 - val_acc: 0.8478\n",
"\n",
"Epoch 00332: val_acc did not improve from 0.86649\n",
"Epoch 333/3000\n",
" - 39s - loss: 0.3407 - acc: 0.9570 - val_loss: 0.9152 - val_acc: 0.8497\n",
"\n",
"Epoch 00333: val_acc did not improve from 0.86649\n",
"Epoch 334/3000\n",
" - 39s - loss: 0.3713 - acc: 0.9541 - val_loss: 0.8634 - val_acc: 0.8529\n",
"\n",
"Epoch 00334: val_acc did not improve from 0.86649\n",
"Epoch 335/3000\n",
" - 39s - loss: 0.3541 - acc: 0.9537 - val_loss: 0.9005 - val_acc: 0.8513\n",
"\n",
"Epoch 00335: val_acc did not improve from 0.86649\n",
"Epoch 336/3000\n",
" - 39s - loss: 0.3709 - acc: 0.9506 - val_loss: 0.9184 - val_acc: 0.8392\n",
"\n",
"Epoch 00336: val_acc did not improve from 0.86649\n",
"Epoch 337/3000\n",
" - 39s - loss: 0.3675 - acc: 0.9500 - val_loss: 0.8765 - val_acc: 0.8571\n",
"\n",
"Epoch 00337: val_acc did not improve from 0.86649\n",
"Epoch 338/3000\n",
" - 39s - loss: 0.3534 - acc: 0.9562 - val_loss: 0.9150 - val_acc: 0.8529\n",
"\n",
"Epoch 00338: val_acc did not improve from 0.86649\n",
"Epoch 339/3000\n",
" - 39s - loss: 0.3440 - acc: 0.9561 - val_loss: 0.9922 - val_acc: 0.8400\n",
"\n",
"Epoch 00339: val_acc did not improve from 0.86649\n",
"Epoch 340/3000\n",
" - 39s - loss: 0.3579 - acc: 0.9535 - val_loss: 0.9198 - val_acc: 0.8451\n",
"\n",
"Epoch 00340: val_acc did not improve from 0.86649\n",
"Epoch 341/3000\n",
" - 39s - loss: 0.3605 - acc: 0.9511 - val_loss: 0.8777 - val_acc: 0.8529\n",
"\n",
"Epoch 00341: val_acc did not improve from 0.86649\n",
"Epoch 342/3000\n",
" - 39s - loss: 0.3638 - acc: 0.9511 - val_loss: 0.9121 - val_acc: 0.8509\n",
"\n",
"Epoch 00342: val_acc did not improve from 0.86649\n",
"Epoch 343/3000\n",
" - 39s - loss: 0.3619 - acc: 0.9514 - val_loss: 0.9426 - val_acc: 0.8424\n",
"\n",
"Epoch 00343: val_acc did not improve from 0.86649\n",
"Epoch 344/3000\n",
" - 39s - loss: 0.3542 - acc: 0.9553 - val_loss: 0.9400 - val_acc: 0.8443\n",
"\n",
"Epoch 00344: val_acc did not improve from 0.86649\n",
"\n",
"Epoch 00344: ReduceLROnPlateau reducing learning rate to 6.302494111878331e-05.\n",
"Epoch 345/3000\n",
" - 40s - loss: 0.3759 - acc: 0.9470 - val_loss: 0.8714 - val_acc: 0.8525\n",
"\n",
"Epoch 00345: val_acc did not improve from 0.86649\n",
"Epoch 346/3000\n",
" - 39s - loss: 0.3568 - acc: 0.9523 - val_loss: 0.9523 - val_acc: 0.8447\n",
"\n",
"Epoch 00346: val_acc did not improve from 0.86649\n",
"Epoch 347/3000\n",
" - 39s - loss: 0.3402 - acc: 0.9568 - val_loss: 0.9460 - val_acc: 0.8459\n",
"\n",
"Epoch 00347: val_acc did not improve from 0.86649\n",
"Epoch 348/3000\n",
" - 39s - loss: 0.3569 - acc: 0.9527 - val_loss: 0.8606 - val_acc: 0.8568\n",
"\n",
"Epoch 00348: val_acc did not improve from 0.86649\n",
"Epoch 349/3000\n",
" - 39s - loss: 0.3515 - acc: 0.9555 - val_loss: 0.9580 - val_acc: 0.8486\n",
"\n",
"Epoch 00349: val_acc did not improve from 0.86649\n",
"Epoch 350/3000\n",
" - 39s - loss: 0.3514 - acc: 0.9541 - val_loss: 0.9491 - val_acc: 0.8427\n",
"\n",
"Epoch 00350: val_acc did not improve from 0.86649\n",
"Epoch 351/3000\n",
" - 39s - loss: 0.3516 - acc: 0.9535 - val_loss: 0.9365 - val_acc: 0.8381\n",
"\n",
"Epoch 00351: val_acc did not improve from 0.86649\n",
"Epoch 352/3000\n",
" - 39s - loss: 0.3413 - acc: 0.9561 - val_loss: 0.9349 - val_acc: 0.8513\n",
"\n",
"Epoch 00352: val_acc did not improve from 0.86649\n",
"Epoch 353/3000\n",
" - 39s - loss: 0.3529 - acc: 0.9534 - val_loss: 0.9106 - val_acc: 0.8513\n",
"\n",
"Epoch 00353: val_acc did not improve from 0.86649\n",
"Epoch 354/3000\n",
" - 40s - loss: 0.3388 - acc: 0.9576 - val_loss: 0.8650 - val_acc: 0.8540\n",
"\n",
"Epoch 00354: val_acc did not improve from 0.86649\n",
"Epoch 355/3000\n",
" - 39s - loss: 0.3457 - acc: 0.9576 - val_loss: 0.9217 - val_acc: 0.8459\n",
"\n",
"Epoch 00355: val_acc did not improve from 0.86649\n",
"Epoch 356/3000\n",
" - 39s - loss: 0.3485 - acc: 0.9546 - val_loss: 0.9274 - val_acc: 0.8482\n",
"\n",
"Epoch 00356: val_acc did not improve from 0.86649\n",
"Epoch 357/3000\n",
" - 39s - loss: 0.3410 - acc: 0.9561 - val_loss: 0.9489 - val_acc: 0.8459\n",
"\n",
"Epoch 00357: val_acc did not improve from 0.86649\n",
"Epoch 358/3000\n",
" - 39s - loss: 0.3405 - acc: 0.9576 - val_loss: 0.9048 - val_acc: 0.8536\n",
"\n",
"Epoch 00358: val_acc did not improve from 0.86649\n",
"Epoch 359/3000\n",
" - 39s - loss: 0.3482 - acc: 0.9550 - val_loss: 0.9928 - val_acc: 0.8462\n",
"\n",
"Epoch 00359: val_acc did not improve from 0.86649\n",
"Epoch 360/3000\n",
" - 39s - loss: 0.3555 - acc: 0.9561 - val_loss: 0.9409 - val_acc: 0.8501\n",
"\n",
"Epoch 00360: val_acc did not improve from 0.86649\n",
"Epoch 361/3000\n",
" - 39s - loss: 0.3373 - acc: 0.9562 - val_loss: 0.8830 - val_acc: 0.8529\n",
"\n",
"Epoch 00361: val_acc did not improve from 0.86649\n",
"Epoch 362/3000\n",
" - 39s - loss: 0.3407 - acc: 0.9527 - val_loss: 0.9251 - val_acc: 0.8466\n",
"\n",
"Epoch 00362: val_acc did not improve from 0.86649\n",
"Epoch 363/3000\n",
" - 39s - loss: 0.3546 - acc: 0.9565 - val_loss: 0.8666 - val_acc: 0.8552\n",
"\n",
"Epoch 00363: val_acc did not improve from 0.86649\n",
"Epoch 364/3000\n",
" - 39s - loss: 0.3530 - acc: 0.9526 - val_loss: 0.9651 - val_acc: 0.8462\n",
"\n",
"Epoch 00364: val_acc did not improve from 0.86649\n",
"Epoch 365/3000\n",
" - 39s - loss: 0.3449 - acc: 0.9586 - val_loss: 0.9295 - val_acc: 0.8513\n",
"\n",
"Epoch 00365: val_acc did not improve from 0.86649\n",
"Epoch 366/3000\n",
" - 39s - loss: 0.3288 - acc: 0.9567 - val_loss: 1.0094 - val_acc: 0.8427\n",
"\n",
"Epoch 00366: val_acc did not improve from 0.86649\n",
"Epoch 367/3000\n",
" - 39s - loss: 0.3459 - acc: 0.9553 - val_loss: 0.9412 - val_acc: 0.8459\n",
"\n",
"Epoch 00367: val_acc did not improve from 0.86649\n",
"Epoch 368/3000\n",
" - 39s - loss: 0.3469 - acc: 0.9565 - val_loss: 0.8496 - val_acc: 0.8575\n",
"\n",
"Epoch 00368: val_acc did not improve from 0.86649\n",
"Epoch 369/3000\n",
" - 39s - loss: 0.3490 - acc: 0.9527 - val_loss: 1.0578 - val_acc: 0.8276\n",
"\n",
"Epoch 00369: val_acc did not improve from 0.86649\n",
"Epoch 370/3000\n",
" - 40s - loss: 0.3739 - acc: 0.9506 - val_loss: 0.9138 - val_acc: 0.8497\n",
"\n",
"Epoch 00370: val_acc did not improve from 0.86649\n",
"Epoch 371/3000\n",
" - 39s - loss: 0.3419 - acc: 0.9556 - val_loss: 0.9400 - val_acc: 0.8459\n",
"\n",
"Epoch 00371: val_acc did not improve from 0.86649\n",
"Epoch 372/3000\n",
" - 39s - loss: 0.3464 - acc: 0.9555 - val_loss: 0.9172 - val_acc: 0.8459\n",
"\n",
"Epoch 00372: val_acc did not improve from 0.86649\n",
"Epoch 373/3000\n",
" - 39s - loss: 0.3294 - acc: 0.9595 - val_loss: 0.8920 - val_acc: 0.8509\n",
"\n",
"Epoch 00373: val_acc did not improve from 0.86649\n",
"Epoch 374/3000\n",
" - 39s - loss: 0.3560 - acc: 0.9523 - val_loss: 0.8554 - val_acc: 0.8529\n",
"\n",
"Epoch 00374: val_acc did not improve from 0.86649\n",
"\n",
"Epoch 00374: ReduceLROnPlateau reducing learning rate to 5.9873694408452134e-05.\n",
"Epoch 375/3000\n",
" - 40s - loss: 0.3404 - acc: 0.9543 - val_loss: 0.9453 - val_acc: 0.8509\n",
"\n",
"Epoch 00375: val_acc did not improve from 0.86649\n",
"Epoch 376/3000\n",
" - 39s - loss: 0.3477 - acc: 0.9555 - val_loss: 0.9316 - val_acc: 0.8486\n",
"\n",
"Epoch 00376: val_acc did not improve from 0.86649\n",
"Epoch 377/3000\n",
" - 39s - loss: 0.3447 - acc: 0.9543 - val_loss: 0.8490 - val_acc: 0.8599\n",
"\n",
"Epoch 00377: val_acc did not improve from 0.86649\n",
"Epoch 378/3000\n",
" - 40s - loss: 0.3414 - acc: 0.9564 - val_loss: 0.9153 - val_acc: 0.8509\n",
"\n",
"Epoch 00378: val_acc did not improve from 0.86649\n",
"Epoch 379/3000\n",
" - 39s - loss: 0.3439 - acc: 0.9565 - val_loss: 0.8819 - val_acc: 0.8517\n",
"\n",
"Epoch 00379: val_acc did not improve from 0.86649\n",
"Epoch 380/3000\n",
" - 39s - loss: 0.3306 - acc: 0.9571 - val_loss: 0.8860 - val_acc: 0.8571\n",
"\n",
"Epoch 00380: val_acc did not improve from 0.86649\n",
"Epoch 381/3000\n",
" - 39s - loss: 0.3328 - acc: 0.9592 - val_loss: 0.9399 - val_acc: 0.8494\n",
"\n",
"Epoch 00381: val_acc did not improve from 0.86649\n",
"Epoch 382/3000\n",
" - 39s - loss: 0.3626 - acc: 0.9530 - val_loss: 0.8948 - val_acc: 0.8513\n",
"\n",
"Epoch 00382: val_acc did not improve from 0.86649\n",
"Epoch 383/3000\n",
" - 39s - loss: 0.3342 - acc: 0.9562 - val_loss: 0.9457 - val_acc: 0.8447\n",
"\n",
"Epoch 00383: val_acc did not improve from 0.86649\n",
"Epoch 384/3000\n",
" - 39s - loss: 0.3262 - acc: 0.9594 - val_loss: 0.8822 - val_acc: 0.8529\n",
"\n",
"Epoch 00384: val_acc did not improve from 0.86649\n",
"Epoch 385/3000\n",
" - 39s - loss: 0.3446 - acc: 0.9561 - val_loss: 0.8581 - val_acc: 0.8591\n",
"\n",
"Epoch 00385: val_acc did not improve from 0.86649\n",
"Epoch 386/3000\n",
" - 40s - loss: 0.3536 - acc: 0.9561 - val_loss: 0.9542 - val_acc: 0.8404\n",
"\n",
"Epoch 00386: val_acc did not improve from 0.86649\n",
"Epoch 387/3000\n",
" - 39s - loss: 0.3523 - acc: 0.9553 - val_loss: 0.9583 - val_acc: 0.8334\n",
"\n",
"Epoch 00387: val_acc did not improve from 0.86649\n",
"Epoch 388/3000\n",
" - 40s - loss: 0.3573 - acc: 0.9558 - val_loss: 0.8497 - val_acc: 0.8614\n",
"\n",
"Epoch 00388: val_acc did not improve from 0.86649\n",
"Epoch 389/3000\n",
" - 39s - loss: 0.3448 - acc: 0.9564 - val_loss: 0.9241 - val_acc: 0.8490\n",
"\n",
"Epoch 00390: val_acc did not improve from 0.86649\n",
"Epoch 391/3000\n",
" - 39s - loss: 0.3355 - acc: 0.9601 - val_loss: 0.9487 - val_acc: 0.8459\n",
"\n",
"Epoch 00391: val_acc did not improve from 0.86649\n",
"Epoch 392/3000\n",
" - 39s - loss: 0.3368 - acc: 0.9591 - val_loss: 0.9421 - val_acc: 0.8540\n",
"\n",
"Epoch 00392: val_acc did not improve from 0.86649\n",
"Epoch 393/3000\n",
" - 40s - loss: 0.3445 - acc: 0.9546 - val_loss: 0.9602 - val_acc: 0.8435\n",
"\n",
"Epoch 00393: val_acc did not improve from 0.86649\n",
"Epoch 394/3000\n",
" - 39s - loss: 0.3369 - acc: 0.9582 - val_loss: 1.0126 - val_acc: 0.8404\n",
"\n",
"Epoch 00394: val_acc did not improve from 0.86649\n",
"Epoch 395/3000\n",
" - 39s - loss: 0.3393 - acc: 0.9553 - val_loss: 0.9342 - val_acc: 0.8509\n",
"\n",
"Epoch 00395: val_acc did not improve from 0.86649\n",
"Epoch 396/3000\n",
" - 39s - loss: 0.3227 - acc: 0.9615 - val_loss: 0.8635 - val_acc: 0.8552\n",
"\n",
"Epoch 00396: val_acc did not improve from 0.86649\n",
"Epoch 397/3000\n",
" - 39s - loss: 0.3304 - acc: 0.9580 - val_loss: 0.9210 - val_acc: 0.8509\n",
"\n",
"Epoch 00397: val_acc did not improve from 0.86649\n",
"Epoch 398/3000\n",
" - 39s - loss: 0.3367 - acc: 0.9574 - val_loss: 0.8472 - val_acc: 0.8677\n",
"\n",
"Epoch 00398: val_acc improved from 0.86649 to 0.86765, saving model to ./ModelSnapshots/LSTM-v2-398.h5\n",
"Epoch 399/3000\n",
" - 39s - loss: 0.3272 - acc: 0.9583 - val_loss: 0.8903 - val_acc: 0.8564\n",
"\n",
"Epoch 00399: val_acc did not improve from 0.86765\n",
"Epoch 400/3000\n",
" - 39s - loss: 0.3386 - acc: 0.9577 - val_loss: 0.8627 - val_acc: 0.8591\n",
"\n",
"Epoch 00400: val_acc did not improve from 0.86765\n",
"Epoch 401/3000\n",
" - 40s - loss: 0.3461 - acc: 0.9568 - val_loss: 0.9331 - val_acc: 0.8482\n",
"\n",
"Epoch 00401: val_acc did not improve from 0.86765\n",
"Epoch 402/3000\n",
" - 39s - loss: 0.3497 - acc: 0.9555 - val_loss: 0.8815 - val_acc: 0.8564\n",
"\n",
"Epoch 00402: val_acc did not improve from 0.86765\n",
"Epoch 403/3000\n",
" - 40s - loss: 0.3369 - acc: 0.9562 - val_loss: 0.9459 - val_acc: 0.8536\n",
"\n",
"Epoch 00403: val_acc did not improve from 0.86765\n",
"Epoch 404/3000\n",
" - 39s - loss: 0.3439 - acc: 0.9532 - val_loss: 0.8664 - val_acc: 0.8595\n",
"\n",
"Epoch 00404: val_acc did not improve from 0.86765\n",
"\n",
"Epoch 00404: ReduceLROnPlateau reducing learning rate to 5.68800103792455e-05.\n",
"Epoch 405/3000\n",
" - 39s - loss: 0.3288 - acc: 0.9592 - val_loss: 0.8931 - val_acc: 0.8521\n",
"\n",
"Epoch 00405: val_acc did not improve from 0.86765\n",
"Epoch 406/3000\n",
" - 39s - loss: 0.3252 - acc: 0.9591 - val_loss: 0.9370 - val_acc: 0.8443\n",
"\n",
"Epoch 00406: val_acc did not improve from 0.86765\n",
"Epoch 407/3000\n",
" - 39s - loss: 0.3420 - acc: 0.9562 - val_loss: 0.8960 - val_acc: 0.8548\n",
"\n",
"Epoch 00407: val_acc did not improve from 0.86765\n",
"Epoch 408/3000\n",
" - 39s - loss: 0.3438 - acc: 0.9552 - val_loss: 0.9718 - val_acc: 0.8427\n",
"\n",
"Epoch 00408: val_acc did not improve from 0.86765\n",
"Epoch 409/3000\n",
" - 39s - loss: 0.3356 - acc: 0.9574 - val_loss: 0.9235 - val_acc: 0.8486\n",
"\n",
"Epoch 00409: val_acc did not improve from 0.86765\n",
"Epoch 410/3000\n",
" - 39s - loss: 0.3196 - acc: 0.9603 - val_loss: 0.8746 - val_acc: 0.8614\n",
"\n",
"Epoch 00410: val_acc did not improve from 0.86765\n",
"Epoch 411/3000\n",
" - 39s - loss: 0.3455 - acc: 0.9568 - val_loss: 0.8772 - val_acc: 0.8568\n",
"\n",
"Epoch 00411: val_acc did not improve from 0.86765\n",
"Epoch 412/3000\n",
" - 39s - loss: 0.3620 - acc: 0.9526 - val_loss: 0.8925 - val_acc: 0.8556\n",
"\n",
"Epoch 00412: val_acc did not improve from 0.86765\n",
"Epoch 413/3000\n",
" - 39s - loss: 0.3418 - acc: 0.9570 - val_loss: 0.8578 - val_acc: 0.8606\n",
"\n",
"Epoch 00413: val_acc did not improve from 0.86765\n",
"Epoch 414/3000\n",
" - 39s - loss: 0.3230 - acc: 0.9583 - val_loss: 0.9399 - val_acc: 0.8509\n",
"\n",
"Epoch 00414: val_acc did not improve from 0.86765\n",
"Epoch 415/3000\n",
" - 39s - loss: 0.3225 - acc: 0.9577 - val_loss: 0.9693 - val_acc: 0.8466\n",
"\n",
"Epoch 00415: val_acc did not improve from 0.86765\n",
"Epoch 416/3000\n",
" - 39s - loss: 0.3343 - acc: 0.9574 - val_loss: 0.8807 - val_acc: 0.8525\n",
"\n",
"Epoch 00416: val_acc did not improve from 0.86765\n",
"Epoch 417/3000\n",
" - 39s - loss: 0.3379 - acc: 0.9561 - val_loss: 0.8396 - val_acc: 0.8595\n",
"\n",
"Epoch 00417: val_acc did not improve from 0.86765\n",
"Epoch 418/3000\n",
" - 39s - loss: 0.3500 - acc: 0.9515 - val_loss: 0.8685 - val_acc: 0.8560\n",
"\n",
"Epoch 00418: val_acc did not improve from 0.86765\n",
"Epoch 419/3000\n",
" - 39s - loss: 0.3236 - acc: 0.9594 - val_loss: 0.8988 - val_acc: 0.8548\n",
"\n",
"Epoch 00419: val_acc did not improve from 0.86765\n",
"Epoch 420/3000\n",
" - 39s - loss: 0.3380 - acc: 0.9535 - val_loss: 0.8662 - val_acc: 0.8544\n",
"\n",
"Epoch 00420: val_acc did not improve from 0.86765\n",
"Epoch 421/3000\n",
" - 39s - loss: 0.3282 - acc: 0.9588 - val_loss: 0.9285 - val_acc: 0.8470\n",
"\n",
"Epoch 00421: val_acc did not improve from 0.86765\n",
"Epoch 422/3000\n",
" - 39s - loss: 0.3402 - acc: 0.9573 - val_loss: 0.9211 - val_acc: 0.8517\n",
"\n",
"Epoch 00422: val_acc did not improve from 0.86765\n",
"Epoch 423/3000\n",
" - 39s - loss: 0.3400 - acc: 0.9570 - val_loss: 0.8997 - val_acc: 0.8540\n",
"\n",
"Epoch 00423: val_acc did not improve from 0.86765\n",
"Epoch 424/3000\n",
" - 39s - loss: 0.3168 - acc: 0.9633 - val_loss: 0.8881 - val_acc: 0.8571\n",
"\n",
"Epoch 00424: val_acc did not improve from 0.86765\n",
"Epoch 425/3000\n",
" - 39s - loss: 0.3453 - acc: 0.9556 - val_loss: 0.9445 - val_acc: 0.8431\n",
"\n",
"Epoch 00425: val_acc did not improve from 0.86765\n",
"Epoch 426/3000\n",
" - 39s - loss: 0.3390 - acc: 0.9570 - val_loss: 0.9125 - val_acc: 0.8497\n",
"\n",
"Epoch 00426: val_acc did not improve from 0.86765\n",
"Epoch 427/3000\n",
" - 39s - loss: 0.3397 - acc: 0.9552 - val_loss: 0.8986 - val_acc: 0.8564\n",
"\n",
"Epoch 00427: val_acc did not improve from 0.86765\n",
"Epoch 428/3000\n",
" - 39s - loss: 0.3488 - acc: 0.9553 - val_loss: 0.8960 - val_acc: 0.8509\n",
"\n",
"Epoch 00428: val_acc did not improve from 0.86765\n",
"Epoch 429/3000\n",
" - 39s - loss: 0.3300 - acc: 0.9600 - val_loss: 0.8449 - val_acc: 0.8599\n",
"\n",
"Epoch 00429: val_acc did not improve from 0.86765\n",
"Epoch 430/3000\n",
" - 39s - loss: 0.3296 - acc: 0.9588 - val_loss: 0.9478 - val_acc: 0.8486\n",
"\n",
"Epoch 00430: val_acc did not improve from 0.86765\n",
"Epoch 431/3000\n",
" - 39s - loss: 0.3305 - acc: 0.9589 - val_loss: 0.8713 - val_acc: 0.8552\n",
"\n",
"Epoch 00431: val_acc did not improve from 0.86765\n",
"Epoch 432/3000\n",
" - 39s - loss: 0.3278 - acc: 0.9580 - val_loss: 0.9387 - val_acc: 0.8568\n",
"\n",
"Epoch 00432: val_acc did not improve from 0.86765\n",
"Epoch 433/3000\n",
" - 39s - loss: 0.3489 - acc: 0.9515 - val_loss: 0.9361 - val_acc: 0.8505\n",
"\n",
"Epoch 00433: val_acc did not improve from 0.86765\n",
"Epoch 434/3000\n",
" - 39s - loss: 0.3327 - acc: 0.9562 - val_loss: 0.9641 - val_acc: 0.8478\n",
"\n",
"Epoch 00434: val_acc did not improve from 0.86765\n",
"\n",
"Epoch 00434: ReduceLROnPlateau reducing learning rate to 5.4036009169067255e-05.\n",
"Epoch 435/3000\n",
" - 39s - loss: 0.3291 - acc: 0.9597 - val_loss: 0.9595 - val_acc: 0.8381\n",
"\n",
"Epoch 00435: val_acc did not improve from 0.86765\n",
"Epoch 436/3000\n",
" - 39s - loss: 0.3637 - acc: 0.9553 - val_loss: 0.9519 - val_acc: 0.8466\n",
"\n",
"Epoch 00436: val_acc did not improve from 0.86765\n",
"Epoch 437/3000\n",
" - 39s - loss: 0.3492 - acc: 0.9565 - val_loss: 1.0486 - val_acc: 0.8326\n",
"\n",
"Epoch 00437: val_acc did not improve from 0.86765\n",
"Epoch 438/3000\n",
" - 39s - loss: 0.3500 - acc: 0.9518 - val_loss: 0.9460 - val_acc: 0.8451\n",
"\n",
"Epoch 00438: val_acc did not improve from 0.86765\n",
"Epoch 439/3000\n",
" - 39s - loss: 0.3331 - acc: 0.9579 - val_loss: 0.9051 - val_acc: 0.8501\n",
"\n",
"Epoch 00439: val_acc did not improve from 0.86765\n",
"Epoch 440/3000\n",
" - 39s - loss: 0.3301 - acc: 0.9580 - val_loss: 0.8741 - val_acc: 0.8509\n",
"\n",
"Epoch 00440: val_acc did not improve from 0.86765\n",
"Epoch 441/3000\n",
" - 39s - loss: 0.3285 - acc: 0.9606 - val_loss: 0.9382 - val_acc: 0.8482\n",
"\n",
"Epoch 00441: val_acc did not improve from 0.86765\n",
"Epoch 442/3000\n",
" - 39s - loss: 0.3185 - acc: 0.9626 - val_loss: 0.9443 - val_acc: 0.8466\n",
"\n",
"Epoch 00442: val_acc did not improve from 0.86765\n",
"Epoch 443/3000\n",
" - 39s - loss: 0.3345 - acc: 0.9580 - val_loss: 0.9091 - val_acc: 0.8466\n",
"\n",
"Epoch 00443: val_acc did not improve from 0.86765\n",
"Epoch 444/3000\n",
" - 40s - loss: 0.3237 - acc: 0.9589 - val_loss: 0.8787 - val_acc: 0.8571\n",
"\n",
"Epoch 00444: val_acc did not improve from 0.86765\n",
"Epoch 445/3000\n",
" - 39s - loss: 0.3170 - acc: 0.9595 - val_loss: 0.9144 - val_acc: 0.8501\n",
"\n",
"Epoch 00445: val_acc did not improve from 0.86765\n",
"Epoch 446/3000\n",
" - 39s - loss: 0.3250 - acc: 0.9580 - val_loss: 0.9949 - val_acc: 0.8404\n",
"\n",
"Epoch 00446: val_acc did not improve from 0.86765\n",
"Epoch 447/3000\n",
" - 40s - loss: 0.3482 - acc: 0.9532 - val_loss: 0.9225 - val_acc: 0.8521\n",
"\n",
"Epoch 00447: val_acc did not improve from 0.86765\n",
"Epoch 448/3000\n",
" - 39s - loss: 0.3508 - acc: 0.9574 - val_loss: 0.9944 - val_acc: 0.8431\n",
"\n",
"Epoch 00448: val_acc did not improve from 0.86765\n",
"Epoch 449/3000\n",
" - 39s - loss: 0.3301 - acc: 0.9565 - val_loss: 0.8755 - val_acc: 0.8591\n",
"\n",
"Epoch 00449: val_acc did not improve from 0.86765\n",
"Epoch 450/3000\n",
" - 39s - loss: 0.3210 - acc: 0.9615 - val_loss: 0.9460 - val_acc: 0.8525\n",
"\n",
"Epoch 00450: val_acc did not improve from 0.86765\n",
"Epoch 451/3000\n",
" - 39s - loss: 0.3255 - acc: 0.9562 - val_loss: 0.9479 - val_acc: 0.8478\n",
"\n",
"Epoch 00451: val_acc did not improve from 0.86765\n",
"Epoch 452/3000\n",
" - 40s - loss: 0.3187 - acc: 0.9589 - val_loss: 0.9648 - val_acc: 0.8490\n",
"\n",
"Epoch 00452: val_acc did not improve from 0.86765\n",
"Epoch 453/3000\n",
" - 40s - loss: 0.3272 - acc: 0.9603 - val_loss: 0.9094 - val_acc: 0.8595\n",
"\n",
"Epoch 00453: val_acc did not improve from 0.86765\n",
"Epoch 454/3000\n",
" - 39s - loss: 0.3186 - acc: 0.9583 - val_loss: 0.9130 - val_acc: 0.8509\n",
"\n",
"Epoch 00454: val_acc did not improve from 0.86765\n",
"Epoch 455/3000\n",
" - 39s - loss: 0.3492 - acc: 0.9540 - val_loss: 0.9120 - val_acc: 0.8494\n",
"\n",
"Epoch 00455: val_acc did not improve from 0.86765\n",
"Epoch 456/3000\n",
" - 39s - loss: 0.3335 - acc: 0.9586 - val_loss: 0.9360 - val_acc: 0.8552\n",
"\n",
"Epoch 00456: val_acc did not improve from 0.86765\n",
"Epoch 457/3000\n",
" - 39s - loss: 0.3273 - acc: 0.9568 - val_loss: 0.9653 - val_acc: 0.8435\n",
"\n",
"Epoch 00457: val_acc did not improve from 0.86765\n",
"Epoch 458/3000\n",
" - 40s - loss: 0.3327 - acc: 0.9576 - val_loss: 0.9670 - val_acc: 0.8439\n",
"\n",
"Epoch 00458: val_acc did not improve from 0.86765\n",
"Epoch 459/3000\n",
" - 39s - loss: 0.3384 - acc: 0.9565 - val_loss: 0.9111 - val_acc: 0.8529\n",
"\n",
"Epoch 00459: val_acc did not improve from 0.86765\n",
"Epoch 460/3000\n",
" - 39s - loss: 0.3256 - acc: 0.9582 - val_loss: 0.9494 - val_acc: 0.8447\n",
"\n",
"Epoch 00460: val_acc did not improve from 0.86765\n",
"Epoch 461/3000\n",
" - 39s - loss: 0.3388 - acc: 0.9567 - val_loss: 0.8919 - val_acc: 0.8521\n",
"\n",
"Epoch 00461: val_acc did not improve from 0.86765\n",
"Epoch 462/3000\n",
" - 39s - loss: 0.3470 - acc: 0.9543 - val_loss: 0.8662 - val_acc: 0.8568\n",
"\n",
"Epoch 00462: val_acc did not improve from 0.86765\n",
"Epoch 463/3000\n",
" - 40s - loss: 0.3261 - acc: 0.9583 - val_loss: 0.8317 - val_acc: 0.8556\n",
"\n",
"Epoch 00463: val_acc did not improve from 0.86765\n",
"Epoch 464/3000\n",
" - 39s - loss: 0.3419 - acc: 0.9561 - val_loss: 1.0010 - val_acc: 0.8416\n",
"\n",
"Epoch 00464: val_acc did not improve from 0.86765\n",
"\n",
"Epoch 00464: ReduceLROnPlateau reducing learning rate to 5.133420836500591e-05.\n",
"Epoch 465/3000\n",
" - 39s - loss: 0.3346 - acc: 0.9576 - val_loss: 0.8434 - val_acc: 0.8525\n",
"\n",
"Epoch 00465: val_acc did not improve from 0.86765\n",
"Epoch 466/3000\n",
" - 39s - loss: 0.3281 - acc: 0.9611 - val_loss: 0.8669 - val_acc: 0.8568\n",
"\n",
"Epoch 00466: val_acc did not improve from 0.86765\n",
"Epoch 467/3000\n",
" - 39s - loss: 0.3298 - acc: 0.9591 - val_loss: 0.9269 - val_acc: 0.8501\n",
"\n",
"Epoch 00467: val_acc did not improve from 0.86765\n",
"Epoch 468/3000\n",
" - 39s - loss: 0.3375 - acc: 0.9564 - val_loss: 0.9649 - val_acc: 0.8478\n",
"\n",
"Epoch 00468: val_acc did not improve from 0.86765\n",
"Epoch 469/3000\n",
" - 40s - loss: 0.3302 - acc: 0.9604 - val_loss: 0.9569 - val_acc: 0.8447\n",
"\n",
"Epoch 00469: val_acc did not improve from 0.86765\n",
"Epoch 470/3000\n",
" - 39s - loss: 0.3267 - acc: 0.9589 - val_loss: 0.9243 - val_acc: 0.8459\n",
"\n",
"Epoch 00470: val_acc did not improve from 0.86765\n",
"Epoch 471/3000\n",
" - 39s - loss: 0.3168 - acc: 0.9603 - val_loss: 0.9311 - val_acc: 0.8490\n",
"\n",
"Epoch 00471: val_acc did not improve from 0.86765\n",
"Epoch 472/3000\n",
" - 39s - loss: 0.3068 - acc: 0.9645 - val_loss: 0.8702 - val_acc: 0.8568\n",
"\n",
"Epoch 00472: val_acc did not improve from 0.86765\n",
"Epoch 473/3000\n",
" - 39s - loss: 0.3163 - acc: 0.9595 - val_loss: 0.9633 - val_acc: 0.8474\n",
"\n",
"Epoch 00473: val_acc did not improve from 0.86765\n",
"Epoch 474/3000\n",
" - 39s - loss: 0.3149 - acc: 0.9624 - val_loss: 0.9056 - val_acc: 0.8497\n",
"\n",
"Epoch 00474: val_acc did not improve from 0.86765\n",
"Epoch 475/3000\n",
" - 39s - loss: 0.3333 - acc: 0.9568 - val_loss: 0.8720 - val_acc: 0.8568\n",
"\n",
"Epoch 00475: val_acc did not improve from 0.86765\n",
"Epoch 476/3000\n",
" - 39s - loss: 0.3150 - acc: 0.9623 - val_loss: 0.9025 - val_acc: 0.8513\n",
"\n",
"Epoch 00476: val_acc did not improve from 0.86765\n",
"Epoch 477/3000\n",
" - 39s - loss: 0.3332 - acc: 0.9570 - val_loss: 0.9184 - val_acc: 0.8513\n",
"\n",
"Epoch 00477: val_acc did not improve from 0.86765\n",
"Epoch 478/3000\n",
" - 39s - loss: 0.3236 - acc: 0.9585 - val_loss: 0.8487 - val_acc: 0.8595\n",
"\n",
"Epoch 00478: val_acc did not improve from 0.86765\n",
"Epoch 479/3000\n",
" - 39s - loss: 0.3201 - acc: 0.9592 - val_loss: 0.8885 - val_acc: 0.8521\n",
"\n",
"Epoch 00479: val_acc did not improve from 0.86765\n",
"Epoch 480/3000\n",
" - 39s - loss: 0.3304 - acc: 0.9562 - val_loss: 0.9052 - val_acc: 0.8509\n",
"\n",
"Epoch 00480: val_acc did not improve from 0.86765\n",
"Epoch 481/3000\n",
" - 39s - loss: 0.3175 - acc: 0.9612 - val_loss: 0.8701 - val_acc: 0.8533\n",
"\n",
"Epoch 00481: val_acc did not improve from 0.86765\n",
"Epoch 482/3000\n",
" - 40s - loss: 0.3281 - acc: 0.9573 - val_loss: 0.9248 - val_acc: 0.8486\n",
"\n",
"Epoch 00482: val_acc did not improve from 0.86765\n",
"Epoch 483/3000\n",
" - 40s - loss: 0.3284 - acc: 0.9576 - val_loss: 0.9771 - val_acc: 0.8501\n",
"\n",
"Epoch 00483: val_acc did not improve from 0.86765\n",
"Epoch 484/3000\n",
" - 39s - loss: 0.3366 - acc: 0.9570 - val_loss: 0.9807 - val_acc: 0.8447\n",
"\n",
"Epoch 00484: val_acc did not improve from 0.86765\n",
"Epoch 485/3000\n",
" - 39s - loss: 0.3184 - acc: 0.9614 - val_loss: 0.9132 - val_acc: 0.8529\n",
"\n",
"Epoch 00485: val_acc did not improve from 0.86765\n",
"Epoch 486/3000\n",
" - 39s - loss: 0.3115 - acc: 0.9617 - val_loss: 0.9153 - val_acc: 0.8505\n",
"\n",
"Epoch 00486: val_acc did not improve from 0.86765\n",
"Epoch 487/3000\n",
" - 39s - loss: 0.3381 - acc: 0.9532 - val_loss: 0.9092 - val_acc: 0.8548\n",
"\n",
"Epoch 00487: val_acc did not improve from 0.86765\n",
"Epoch 488/3000\n",
" - 39s - loss: 0.3372 - acc: 0.9574 - val_loss: 0.9252 - val_acc: 0.8490\n",
"\n",
"Epoch 00488: val_acc did not improve from 0.86765\n",
"Epoch 489/3000\n",
" - 39s - loss: 0.3209 - acc: 0.9597 - val_loss: 0.9447 - val_acc: 0.8427\n",
"\n",
"Epoch 00489: val_acc did not improve from 0.86765\n",
"Epoch 490/3000\n",
" - 39s - loss: 0.3298 - acc: 0.9591 - val_loss: 0.8678 - val_acc: 0.8579\n",
"\n",
"Epoch 00490: val_acc did not improve from 0.86765\n",
"Epoch 491/3000\n",
" - 39s - loss: 0.3257 - acc: 0.9594 - val_loss: 0.8909 - val_acc: 0.8521\n",
"\n",
"Epoch 00491: val_acc did not improve from 0.86765\n",
"Epoch 492/3000\n",
" - 39s - loss: 0.3173 - acc: 0.9597 - val_loss: 0.9157 - val_acc: 0.8536\n",
"\n",
"Epoch 00492: val_acc did not improve from 0.86765\n",
"Epoch 493/3000\n",
" - 39s - loss: 0.3303 - acc: 0.9611 - val_loss: 0.9396 - val_acc: 0.8509\n",
"\n",
"Epoch 00493: val_acc did not improve from 0.86765\n",
"Epoch 494/3000\n",
" - 39s - loss: 0.3222 - acc: 0.9609 - val_loss: 0.9693 - val_acc: 0.8470\n",
"\n",
"Epoch 00494: val_acc did not improve from 0.86765\n",
"\n",
"Epoch 00494: ReduceLROnPlateau reducing learning rate to 4.876749881077558e-05.\n",
"Epoch 495/3000\n",
" - 40s - loss: 0.3281 - acc: 0.9583 - val_loss: 0.9320 - val_acc: 0.8505\n",
"\n",
"Epoch 00495: val_acc did not improve from 0.86765\n",
"Epoch 496/3000\n",
" - 39s - loss: 0.3316 - acc: 0.9574 - val_loss: 0.9316 - val_acc: 0.8533\n",
"\n",
"Epoch 00496: val_acc did not improve from 0.86765\n",
"Epoch 497/3000\n",
" - 39s - loss: 0.3135 - acc: 0.9614 - val_loss: 0.8739 - val_acc: 0.8509\n",
"\n",
"Epoch 00497: val_acc did not improve from 0.86765\n",
"Epoch 498/3000\n",
" - 39s - loss: 0.3212 - acc: 0.9586 - val_loss: 0.8895 - val_acc: 0.8552\n",
"\n",
"Epoch 00498: val_acc did not improve from 0.86765\n",
"Epoch 499/3000\n",
" - 39s - loss: 0.3216 - acc: 0.9595 - val_loss: 0.8799 - val_acc: 0.8568\n",
"\n",
"Epoch 00499: val_acc did not improve from 0.86765\n",
"Epoch 500/3000\n",
" - 39s - loss: 0.3230 - acc: 0.9583 - val_loss: 0.8783 - val_acc: 0.8568\n",
"\n",
"Epoch 00500: val_acc did not improve from 0.86765\n",
"Epoch 501/3000\n",
" - 39s - loss: 0.3177 - acc: 0.9606 - val_loss: 0.9315 - val_acc: 0.8525\n",
"\n",
"Epoch 00501: val_acc did not improve from 0.86765\n",
"Epoch 502/3000\n",
" - 39s - loss: 0.3369 - acc: 0.9550 - val_loss: 0.9903 - val_acc: 0.8505\n",
"\n",
"Epoch 00502: val_acc did not improve from 0.86765\n",
"Epoch 503/3000\n",
" - 39s - loss: 0.3324 - acc: 0.9603 - val_loss: 0.8952 - val_acc: 0.8599\n",
"\n",
"Epoch 00503: val_acc did not improve from 0.86765\n",
"Epoch 504/3000\n",
" - 40s - loss: 0.3390 - acc: 0.9556 - val_loss: 0.9544 - val_acc: 0.8494\n",
"\n",
"Epoch 00504: val_acc did not improve from 0.86765\n",
"Epoch 505/3000\n",
" - 40s - loss: 0.3353 - acc: 0.9564 - val_loss: 0.9124 - val_acc: 0.8517\n",
"\n",
"Epoch 00505: val_acc did not improve from 0.86765\n",
"Epoch 506/3000\n",
" - 39s - loss: 0.3210 - acc: 0.9594 - val_loss: 0.8594 - val_acc: 0.8603\n",
"\n",
"Epoch 00506: val_acc did not improve from 0.86765\n",
"Epoch 507/3000\n",
" - 39s - loss: 0.3219 - acc: 0.9573 - val_loss: 0.8915 - val_acc: 0.8556\n",
"\n",
"Epoch 00507: val_acc did not improve from 0.86765\n",
"Epoch 508/3000\n",
" - 39s - loss: 0.3247 - acc: 0.9588 - val_loss: 0.8953 - val_acc: 0.8536\n",
"\n",
"Epoch 00508: val_acc did not improve from 0.86765\n",
"Epoch 509/3000\n",
" - 39s - loss: 0.3191 - acc: 0.9607 - val_loss: 0.8783 - val_acc: 0.8556\n",
"\n",
"Epoch 00509: val_acc did not improve from 0.86765\n",
"Epoch 510/3000\n",
" - 39s - loss: 0.3234 - acc: 0.9574 - val_loss: 0.9082 - val_acc: 0.8521\n",
"\n",
"Epoch 00510: val_acc did not improve from 0.86765\n",
"Epoch 511/3000\n",
" - 39s - loss: 0.3298 - acc: 0.9562 - val_loss: 0.8984 - val_acc: 0.8564\n",
"\n",
"Epoch 00511: val_acc did not improve from 0.86765\n",
"Epoch 512/3000\n",
" - 39s - loss: 0.3041 - acc: 0.9641 - val_loss: 0.8977 - val_acc: 0.8505\n",
"\n",
"Epoch 00512: val_acc did not improve from 0.86765\n",
"Epoch 513/3000\n",
" - 39s - loss: 0.3270 - acc: 0.9555 - val_loss: 0.8402 - val_acc: 0.8587\n",
"\n",
"Epoch 00513: val_acc did not improve from 0.86765\n",
"Epoch 514/3000\n",
" - 40s - loss: 0.3115 - acc: 0.9606 - val_loss: 0.8794 - val_acc: 0.8606\n",
"\n",
"Epoch 00514: val_acc did not improve from 0.86765\n",
"Epoch 515/3000\n",
" - 39s - loss: 0.3252 - acc: 0.9562 - val_loss: 0.9451 - val_acc: 0.8455\n",
"\n",
"Epoch 00515: val_acc did not improve from 0.86765\n",
"Epoch 516/3000\n",
" - 39s - loss: 0.3219 - acc: 0.9597 - val_loss: 0.8295 - val_acc: 0.8657\n",
"\n",
"Epoch 00516: val_acc did not improve from 0.86765\n",
"Epoch 517/3000\n",
" - 39s - loss: 0.3225 - acc: 0.9577 - val_loss: 0.8308 - val_acc: 0.8638\n",
"\n",
"Epoch 00517: val_acc did not improve from 0.86765\n",
"Epoch 518/3000\n",
" - 39s - loss: 0.3120 - acc: 0.9614 - val_loss: 0.8791 - val_acc: 0.8564\n",
"\n",
"Epoch 00518: val_acc did not improve from 0.86765\n",
"Epoch 519/3000\n",
" - 39s - loss: 0.3207 - acc: 0.9588 - val_loss: 0.8841 - val_acc: 0.8564\n",
"\n",
"Epoch 00519: val_acc did not improve from 0.86765\n",
"Epoch 520/3000\n",
" - 39s - loss: 0.3166 - acc: 0.9603 - val_loss: 0.8913 - val_acc: 0.8548\n",
"\n",
"Epoch 00520: val_acc did not improve from 0.86765\n",
"Epoch 521/3000\n",
" - 39s - loss: 0.3218 - acc: 0.9604 - val_loss: 0.8800 - val_acc: 0.8540\n",
"\n",
"Epoch 00521: val_acc did not improve from 0.86765\n",
"Epoch 522/3000\n",
" - 39s - loss: 0.3278 - acc: 0.9586 - val_loss: 0.8885 - val_acc: 0.8455\n",
"\n",
"Epoch 00522: val_acc did not improve from 0.86765\n",
"Epoch 523/3000\n",
" - 40s - loss: 0.3203 - acc: 0.9577 - val_loss: 0.8893 - val_acc: 0.8548\n",
"\n",
"Epoch 00523: val_acc did not improve from 0.86765\n",
"Epoch 524/3000\n",
" - 39s - loss: 0.3078 - acc: 0.9635 - val_loss: 0.8748 - val_acc: 0.8583\n",
"\n",
"Epoch 00524: val_acc did not improve from 0.86765\n",
"\n",
"Epoch 00524: ReduceLROnPlateau reducing learning rate to 4.63291238702368e-05.\n",
"Epoch 525/3000\n",
" - 40s - loss: 0.3199 - acc: 0.9588 - val_loss: 0.9706 - val_acc: 0.8447\n",
"\n",
"Epoch 00525: val_acc did not improve from 0.86765\n",
"Epoch 526/3000\n",
" - 40s - loss: 0.3191 - acc: 0.9583 - val_loss: 0.8217 - val_acc: 0.8591\n",
"\n",
"Epoch 00526: val_acc did not improve from 0.86765\n",
"Epoch 527/3000\n",
" - 39s - loss: 0.3051 - acc: 0.9626 - val_loss: 0.9206 - val_acc: 0.8505\n",
"\n",
"Epoch 00527: val_acc did not improve from 0.86765\n",
"Epoch 528/3000\n",
" - 39s - loss: 0.3100 - acc: 0.9618 - val_loss: 0.9187 - val_acc: 0.8513\n",
"\n",
"Epoch 00528: val_acc did not improve from 0.86765\n",
"Epoch 529/3000\n",
" - 39s - loss: 0.3182 - acc: 0.9600 - val_loss: 0.8847 - val_acc: 0.8587\n",
"\n",
"Epoch 00529: val_acc did not improve from 0.86765\n",
"Epoch 530/3000\n",
" - 40s - loss: 0.3149 - acc: 0.9598 - val_loss: 0.8858 - val_acc: 0.8459\n",
"\n",
"Epoch 00530: val_acc did not improve from 0.86765\n",
"Epoch 531/3000\n",
" - 39s - loss: 0.3266 - acc: 0.9588 - val_loss: 0.8866 - val_acc: 0.8443\n",
"\n",
"Epoch 00531: val_acc did not improve from 0.86765\n",
"Epoch 532/3000\n",
" - 39s - loss: 0.3166 - acc: 0.9598 - val_loss: 0.8903 - val_acc: 0.8544\n",
"\n",
"Epoch 00532: val_acc did not improve from 0.86765\n",
"Epoch 533/3000\n",
" - 39s - loss: 0.3091 - acc: 0.9632 - val_loss: 0.8742 - val_acc: 0.8560\n",
"\n",
"Epoch 00533: val_acc did not improve from 0.86765\n",
"Epoch 534/3000\n",
" - 39s - loss: 0.3330 - acc: 0.9564 - val_loss: 0.9164 - val_acc: 0.8478\n",
"\n",
"Epoch 00534: val_acc did not improve from 0.86765\n",
"Epoch 535/3000\n",
" - 40s - loss: 0.3164 - acc: 0.9606 - val_loss: 0.9253 - val_acc: 0.8490\n",
"\n",
"Epoch 00535: val_acc did not improve from 0.86765\n",
"Epoch 536/3000\n",
" - 39s - loss: 0.3114 - acc: 0.9585 - val_loss: 0.8726 - val_acc: 0.8568\n",
"\n",
"Epoch 00536: val_acc did not improve from 0.86765\n",
"Epoch 537/3000\n",
" - 39s - loss: 0.3251 - acc: 0.9600 - val_loss: 0.9935 - val_acc: 0.8427\n",
"\n",
"Epoch 00537: val_acc did not improve from 0.86765\n",
"Epoch 538/3000\n",
" - 39s - loss: 0.3159 - acc: 0.9623 - val_loss: 0.9470 - val_acc: 0.8482\n",
"\n",
"Epoch 00538: val_acc did not improve from 0.86765\n",
"Epoch 539/3000\n",
" - 39s - loss: 0.3156 - acc: 0.9591 - val_loss: 0.8687 - val_acc: 0.8630\n",
"\n",
"Epoch 00539: val_acc did not improve from 0.86765\n",
"Epoch 540/3000\n",
" - 39s - loss: 0.3105 - acc: 0.9612 - val_loss: 0.9334 - val_acc: 0.8447\n",
"\n",
"Epoch 00540: val_acc did not improve from 0.86765\n",
"Epoch 541/3000\n",
" - 40s - loss: 0.3124 - acc: 0.9598 - val_loss: 0.9087 - val_acc: 0.8529\n",
"\n",
"Epoch 00541: val_acc did not improve from 0.86765\n",
"Epoch 542/3000\n",
" - 39s - loss: 0.3080 - acc: 0.9615 - val_loss: 0.8623 - val_acc: 0.8622\n",
"\n",
"Epoch 00542: val_acc did not improve from 0.86765\n",
"Epoch 543/3000\n",
" - 39s - loss: 0.3236 - acc: 0.9583 - val_loss: 0.8804 - val_acc: 0.8591\n",
"\n",
"Epoch 00543: val_acc did not improve from 0.86765\n",
"Epoch 544/3000\n",
" - 40s - loss: 0.3359 - acc: 0.9550 - val_loss: 0.8948 - val_acc: 0.8556\n",
"\n",
"Epoch 00544: val_acc did not improve from 0.86765\n",
"Epoch 545/3000\n",
" - 39s - loss: 0.3287 - acc: 0.9595 - val_loss: 0.8948 - val_acc: 0.8540\n",
"\n",
"Epoch 00545: val_acc did not improve from 0.86765\n",
"Epoch 546/3000\n",
" - 39s - loss: 0.3135 - acc: 0.9629 - val_loss: 0.9767 - val_acc: 0.8466\n",
"\n",
"Epoch 00546: val_acc did not improve from 0.86765\n",
"Epoch 547/3000\n",
" - 39s - loss: 0.3047 - acc: 0.9607 - val_loss: 0.9010 - val_acc: 0.8521\n",
"\n",
"Epoch 00547: val_acc did not improve from 0.86765\n",
"Epoch 548/3000\n",
" - 39s - loss: 0.3219 - acc: 0.9603 - val_loss: 0.9056 - val_acc: 0.8564\n",
"\n",
"Epoch 00548: val_acc did not improve from 0.86765\n",
"Epoch 549/3000\n",
" - 40s - loss: 0.3148 - acc: 0.9621 - val_loss: 0.8763 - val_acc: 0.8533\n",
"\n",
"Epoch 00549: val_acc did not improve from 0.86765\n",
"Epoch 550/3000\n",
" - 39s - loss: 0.3018 - acc: 0.9630 - val_loss: 0.9044 - val_acc: 0.8443\n",
"\n",
"Epoch 00550: val_acc did not improve from 0.86765\n",
"Epoch 551/3000\n",
" - 39s - loss: 0.3134 - acc: 0.9621 - val_loss: 0.9063 - val_acc: 0.8513\n",
"\n",
"Epoch 00551: val_acc did not improve from 0.86765\n",
"Epoch 552/3000\n",
" - 39s - loss: 0.3164 - acc: 0.9630 - val_loss: 0.8837 - val_acc: 0.8505\n",
"\n",
"Epoch 00552: val_acc did not improve from 0.86765\n",
"Epoch 553/3000\n",
" - 39s - loss: 0.3270 - acc: 0.9586 - val_loss: 0.8592 - val_acc: 0.8575\n",
"\n",
"Epoch 00553: val_acc did not improve from 0.86765\n",
"Epoch 554/3000\n",
" - 39s - loss: 0.3205 - acc: 0.9588 - val_loss: 0.9231 - val_acc: 0.8517\n",
"\n",
"Epoch 00554: val_acc did not improve from 0.86765\n",
"\n",
"Epoch 00554: ReduceLROnPlateau reducing learning rate to 4.4012669059156905e-05.\n",
"Epoch 555/3000\n",
" - 39s - loss: 0.3312 - acc: 0.9577 - val_loss: 0.8746 - val_acc: 0.8556\n",
"\n",
"Epoch 00555: val_acc did not improve from 0.86765\n",
"Epoch 556/3000\n",
" - 39s - loss: 0.3002 - acc: 0.9606 - val_loss: 0.8975 - val_acc: 0.8579\n",
"\n",
"Epoch 00556: val_acc did not improve from 0.86765\n",
"Epoch 557/3000\n",
" - 39s - loss: 0.3142 - acc: 0.9614 - val_loss: 0.8877 - val_acc: 0.8564\n",
"\n",
"Epoch 00557: val_acc did not improve from 0.86765\n",
"Epoch 558/3000\n",
" - 39s - loss: 0.3120 - acc: 0.9601 - val_loss: 0.9035 - val_acc: 0.8556\n",
"\n",
"Epoch 00558: val_acc did not improve from 0.86765\n",
"Epoch 559/3000\n",
" - 39s - loss: 0.3106 - acc: 0.9635 - val_loss: 0.8927 - val_acc: 0.8459\n",
"\n",
"Epoch 00559: val_acc did not improve from 0.86765\n",
"Epoch 560/3000\n",
" - 39s - loss: 0.3132 - acc: 0.9592 - val_loss: 0.8772 - val_acc: 0.8509\n",
"\n",
"Epoch 00560: val_acc did not improve from 0.86765\n",
"Epoch 561/3000\n",
" - 39s - loss: 0.3115 - acc: 0.9624 - val_loss: 0.9050 - val_acc: 0.8513\n",
"\n",
"Epoch 00561: val_acc did not improve from 0.86765\n",
"Epoch 562/3000\n",
" - 39s - loss: 0.3087 - acc: 0.9621 - val_loss: 0.8773 - val_acc: 0.8583\n",
"\n",
"Epoch 00564: val_acc did not improve from 0.86765\n",
"Epoch 565/3000\n",
" - 39s - loss: 0.3089 - acc: 0.9632 - val_loss: 0.9398 - val_acc: 0.8548\n",
"\n",
"Epoch 00565: val_acc did not improve from 0.86765\n",
"Epoch 566/3000\n",
" - 40s - loss: 0.3065 - acc: 0.9612 - val_loss: 0.9517 - val_acc: 0.8482\n",
"\n",
"Epoch 00566: val_acc did not improve from 0.86765\n",
"Epoch 567/3000\n",
" - 40s - loss: 0.3049 - acc: 0.9630 - val_loss: 0.9277 - val_acc: 0.8470\n",
"\n",
"Epoch 00567: val_acc did not improve from 0.86765\n",
"Epoch 568/3000\n",
" - 40s - loss: 0.3161 - acc: 0.9567 - val_loss: 0.8848 - val_acc: 0.8591\n",
"\n",
"Epoch 00568: val_acc did not improve from 0.86765\n",
"Epoch 569/3000\n",
" - 39s - loss: 0.3028 - acc: 0.9624 - val_loss: 0.9450 - val_acc: 0.8533\n",
"\n",
"Epoch 00569: val_acc did not improve from 0.86765\n",
"Epoch 570/3000\n",
" - 39s - loss: 0.3059 - acc: 0.9601 - val_loss: 0.9400 - val_acc: 0.8494\n",
"\n",
"Epoch 00570: val_acc did not improve from 0.86765\n",
"Epoch 571/3000\n",
" - 39s - loss: 0.3174 - acc: 0.9626 - val_loss: 0.9213 - val_acc: 0.8521\n",
"\n",
"Epoch 00571: val_acc did not improve from 0.86765\n",
"Epoch 572/3000\n",
" - 39s - loss: 0.3332 - acc: 0.9565 - val_loss: 0.9168 - val_acc: 0.8478\n",
"\n",
"Epoch 00572: val_acc did not improve from 0.86765\n",
"Epoch 573/3000\n",
" - 39s - loss: 0.3111 - acc: 0.9617 - val_loss: 0.9227 - val_acc: 0.8490\n",
"\n",
"Epoch 00573: val_acc did not improve from 0.86765\n",
"Epoch 574/3000\n",
" - 39s - loss: 0.3135 - acc: 0.9582 - val_loss: 0.9085 - val_acc: 0.8505\n",
"\n",
"Epoch 00574: val_acc did not improve from 0.86765\n",
"Epoch 575/3000\n",
" - 39s - loss: 0.3082 - acc: 0.9589 - val_loss: 0.9385 - val_acc: 0.8494\n",
"\n",
"Epoch 00575: val_acc did not improve from 0.86765\n",
"Epoch 576/3000\n",
" - 39s - loss: 0.3147 - acc: 0.9614 - val_loss: 0.8738 - val_acc: 0.8575\n",
"\n",
"Epoch 00576: val_acc did not improve from 0.86765\n",
"Epoch 577/3000\n",
" - 39s - loss: 0.3130 - acc: 0.9606 - val_loss: 0.9327 - val_acc: 0.8497\n",
"\n",
"Epoch 00577: val_acc did not improve from 0.86765\n",
"Epoch 578/3000\n",
" - 39s - loss: 0.3073 - acc: 0.9598 - val_loss: 0.9454 - val_acc: 0.8439\n",
"\n",
"Epoch 00578: val_acc did not improve from 0.86765\n",
"Epoch 579/3000\n",
" - 39s - loss: 0.3095 - acc: 0.9617 - val_loss: 0.9060 - val_acc: 0.8556\n",
"\n",
"Epoch 00579: val_acc did not improve from 0.86765\n",
"Epoch 580/3000\n",
" - 39s - loss: 0.3184 - acc: 0.9598 - val_loss: 0.8865 - val_acc: 0.8552\n",
"\n",
"Epoch 00580: val_acc did not improve from 0.86765\n",
"Epoch 581/3000\n",
" - 39s - loss: 0.3052 - acc: 0.9624 - val_loss: 0.8943 - val_acc: 0.8579\n",
"\n",
"Epoch 00581: val_acc did not improve from 0.86765\n",
"Epoch 582/3000\n",
" - 39s - loss: 0.3107 - acc: 0.9618 - val_loss: 0.9659 - val_acc: 0.8501\n",
"\n",
"Epoch 00582: val_acc did not improve from 0.86765\n",
"Epoch 583/3000\n",
" - 39s - loss: 0.3174 - acc: 0.9611 - val_loss: 0.9223 - val_acc: 0.8505\n",
"\n",
"Epoch 00583: val_acc did not improve from 0.86765\n",
"Epoch 584/3000\n",
" - 39s - loss: 0.2990 - acc: 0.9624 - val_loss: 0.9150 - val_acc: 0.8536\n",
"\n",
"Epoch 00584: val_acc did not improve from 0.86765\n",
"\n",
"Epoch 00584: ReduceLROnPlateau reducing learning rate to 4.181203439657111e-05.\n",
"Epoch 585/3000\n",
" - 39s - loss: 0.3184 - acc: 0.9577 - val_loss: 0.9063 - val_acc: 0.8536\n",
"\n",
"Epoch 00585: val_acc did not improve from 0.86765\n",
"Epoch 586/3000\n",
" - 39s - loss: 0.3116 - acc: 0.9614 - val_loss: 0.8674 - val_acc: 0.8544\n",
"\n",
"Epoch 00586: val_acc did not improve from 0.86765\n",
"Epoch 587/3000\n",
" - 39s - loss: 0.3115 - acc: 0.9629 - val_loss: 0.8713 - val_acc: 0.8564\n",
"\n",
"Epoch 00587: val_acc did not improve from 0.86765\n",
"Epoch 588/3000\n",
" - 39s - loss: 0.3092 - acc: 0.9627 - val_loss: 0.9005 - val_acc: 0.8568\n",
"\n",
"Epoch 00588: val_acc did not improve from 0.86765\n",
"Epoch 589/3000\n",
" - 40s - loss: 0.3278 - acc: 0.9556 - val_loss: 0.8666 - val_acc: 0.8560\n",
"\n",
"Epoch 00589: val_acc did not improve from 0.86765\n",
"Epoch 590/3000\n",
" - 39s - loss: 0.3093 - acc: 0.9611 - val_loss: 0.8830 - val_acc: 0.8564\n",
"\n",
"Epoch 00590: val_acc did not improve from 0.86765\n",
"Epoch 591/3000\n",
" - 39s - loss: 0.3041 - acc: 0.9595 - val_loss: 0.8750 - val_acc: 0.8533\n",
"\n",
"Epoch 00591: val_acc did not improve from 0.86765\n",
"Epoch 592/3000\n",
" - 39s - loss: 0.3028 - acc: 0.9633 - val_loss: 0.8818 - val_acc: 0.8544\n",
"\n",
"Epoch 00592: val_acc did not improve from 0.86765\n",
"Epoch 593/3000\n",
" - 39s - loss: 0.3133 - acc: 0.9576 - val_loss: 0.9473 - val_acc: 0.8505\n",
"\n",
"Epoch 00593: val_acc did not improve from 0.86765\n",
"Epoch 594/3000\n",
" - 40s - loss: 0.3150 - acc: 0.9597 - val_loss: 0.9213 - val_acc: 0.8513\n",
"\n",
"Epoch 00594: val_acc did not improve from 0.86765\n",
"Epoch 595/3000\n",
" - 39s - loss: 0.3025 - acc: 0.9630 - val_loss: 0.9128 - val_acc: 0.8478\n",
"\n",
"Epoch 00595: val_acc did not improve from 0.86765\n",
"Epoch 596/3000\n",
" - 39s - loss: 0.2899 - acc: 0.9645 - val_loss: 0.9197 - val_acc: 0.8533\n",
"\n",
"Epoch 00596: val_acc did not improve from 0.86765\n",
"Epoch 597/3000\n",
" - 39s - loss: 0.2943 - acc: 0.9633 - val_loss: 0.9544 - val_acc: 0.8533\n",
"\n",
"Epoch 00597: val_acc did not improve from 0.86765\n",
"Epoch 598/3000\n",
" - 38s - loss: 0.3093 - acc: 0.9623 - val_loss: 0.9325 - val_acc: 0.8513\n",
"\n",
"Epoch 00598: val_acc did not improve from 0.86765\n",
"Epoch 599/3000\n",
" - 39s - loss: 0.3032 - acc: 0.9626 - val_loss: 0.9353 - val_acc: 0.8533\n",
"\n",
"Epoch 00599: val_acc did not improve from 0.86765\n",
"Epoch 600/3000\n",
" - 39s - loss: 0.2977 - acc: 0.9621 - val_loss: 0.9069 - val_acc: 0.8548\n",
"\n",
"Epoch 00600: val_acc did not improve from 0.86765\n",
"Epoch 601/3000\n",
" - 39s - loss: 0.3101 - acc: 0.9614 - val_loss: 0.9313 - val_acc: 0.8513\n",
"\n",
"Epoch 00601: val_acc did not improve from 0.86765\n",
"Epoch 602/3000\n",
" - 39s - loss: 0.3042 - acc: 0.9639 - val_loss: 0.9277 - val_acc: 0.8517\n",
"\n",
"Epoch 00602: val_acc did not improve from 0.86765\n",
"Epoch 603/3000\n",
" - 39s - loss: 0.3144 - acc: 0.9589 - val_loss: 0.8583 - val_acc: 0.8626\n",
"\n",
"Epoch 00603: val_acc did not improve from 0.86765\n",
"Epoch 604/3000\n",
" - 40s - loss: 0.3224 - acc: 0.9591 - val_loss: 0.8898 - val_acc: 0.8521\n",
"\n",
"Epoch 00604: val_acc did not improve from 0.86765\n",
"Epoch 605/3000\n",
" - 39s - loss: 0.3099 - acc: 0.9611 - val_loss: 0.8822 - val_acc: 0.8641\n",
"\n",
"Epoch 00605: val_acc did not improve from 0.86765\n",
"Epoch 606/3000\n",
" - 40s - loss: 0.3005 - acc: 0.9627 - val_loss: 0.9255 - val_acc: 0.8536\n",
"\n",
"Epoch 00606: val_acc did not improve from 0.86765\n",
"Epoch 607/3000\n",
" - 39s - loss: 0.3039 - acc: 0.9620 - val_loss: 0.8470 - val_acc: 0.8610\n",
"\n",
"Epoch 00607: val_acc did not improve from 0.86765\n",
"Epoch 608/3000\n",
" - 39s - loss: 0.3012 - acc: 0.9611 - val_loss: 0.8784 - val_acc: 0.8583\n",
"\n",
"Epoch 00608: val_acc did not improve from 0.86765\n",
"Epoch 609/3000\n",
" - 39s - loss: 0.2958 - acc: 0.9641 - val_loss: 0.9520 - val_acc: 0.8513\n",
"\n",
"Epoch 00609: val_acc did not improve from 0.86765\n",
"Epoch 610/3000\n",
" - 39s - loss: 0.3149 - acc: 0.9615 - val_loss: 0.8923 - val_acc: 0.8548\n",
"\n",
"Epoch 00610: val_acc did not improve from 0.86765\n",
"Epoch 611/3000\n",
" - 39s - loss: 0.3170 - acc: 0.9623 - val_loss: 0.9307 - val_acc: 0.8513\n",
"\n",
"Epoch 00611: val_acc did not improve from 0.86765\n",
"Epoch 612/3000\n",
" - 39s - loss: 0.3073 - acc: 0.9653 - val_loss: 0.8659 - val_acc: 0.8599\n",
"\n",
"Epoch 00612: val_acc did not improve from 0.86765\n",
"Epoch 613/3000\n",
" - 40s - loss: 0.2899 - acc: 0.9662 - val_loss: 0.9281 - val_acc: 0.8552\n",
"\n",
"Epoch 00613: val_acc did not improve from 0.86765\n",
"Epoch 614/3000\n",
" - 39s - loss: 0.3131 - acc: 0.9591 - val_loss: 0.8495 - val_acc: 0.8595\n",
"\n",
"Epoch 00614: val_acc did not improve from 0.86765\n",
"\n",
"Epoch 00614: ReduceLROnPlateau reducing learning rate to 3.9721430948702614e-05.\n",
"Epoch 615/3000\n",
" - 39s - loss: 0.2953 - acc: 0.9629 - val_loss: 0.9233 - val_acc: 0.8501\n",
"\n",
"Epoch 00615: val_acc did not improve from 0.86765\n",
"Epoch 616/3000\n",
" - 39s - loss: 0.3051 - acc: 0.9621 - val_loss: 0.9077 - val_acc: 0.8564\n",
"\n",
"Epoch 00616: val_acc did not improve from 0.86765\n",
"Epoch 617/3000\n",
" - 40s - loss: 0.2863 - acc: 0.9662 - val_loss: 0.8860 - val_acc: 0.8571\n",
"\n",
"Epoch 00617: val_acc did not improve from 0.86765\n",
"Epoch 618/3000\n",
" - 39s - loss: 0.3088 - acc: 0.9597 - val_loss: 0.8963 - val_acc: 0.8540\n",
"\n",
"Epoch 00618: val_acc did not improve from 0.86765\n",
"Epoch 619/3000\n",
" - 39s - loss: 0.3013 - acc: 0.9650 - val_loss: 0.9202 - val_acc: 0.8501\n",
"\n",
"Epoch 00619: val_acc did not improve from 0.86765\n",
"Epoch 620/3000\n",
" - 39s - loss: 0.3097 - acc: 0.9604 - val_loss: 0.9244 - val_acc: 0.8505\n",
"\n",
"Epoch 00620: val_acc did not improve from 0.86765\n",
"Epoch 621/3000\n",
" - 39s - loss: 0.3048 - acc: 0.9626 - val_loss: 0.9033 - val_acc: 0.8517\n",
"\n",
"Epoch 00621: val_acc did not improve from 0.86765\n",
"Epoch 622/3000\n",
" - 39s - loss: 0.2948 - acc: 0.9648 - val_loss: 0.8670 - val_acc: 0.8579\n",
"\n",
"Epoch 00622: val_acc did not improve from 0.86765\n",
"Epoch 623/3000\n",
" - 39s - loss: 0.2986 - acc: 0.9647 - val_loss: 0.8519 - val_acc: 0.8591\n",
"\n",
"Epoch 00623: val_acc did not improve from 0.86765\n",
"Epoch 624/3000\n",
" - 39s - loss: 0.3045 - acc: 0.9632 - val_loss: 0.9180 - val_acc: 0.8560\n",
"\n",
"Epoch 00624: val_acc did not improve from 0.86765\n",
"Epoch 625/3000\n",
" - 39s - loss: 0.3078 - acc: 0.9595 - val_loss: 0.9361 - val_acc: 0.8521\n",
"\n",
"Epoch 00625: val_acc did not improve from 0.86765\n",
"Epoch 626/3000\n",
" - 39s - loss: 0.3033 - acc: 0.9620 - val_loss: 0.8913 - val_acc: 0.8548\n",
"\n",
"Epoch 00626: val_acc did not improve from 0.86765\n",
"Epoch 627/3000\n",
" - 39s - loss: 0.2974 - acc: 0.9641 - val_loss: 0.8778 - val_acc: 0.8575\n",
"\n",
"Epoch 00627: val_acc did not improve from 0.86765\n",
"Epoch 628/3000\n",
" - 39s - loss: 0.3035 - acc: 0.9609 - val_loss: 0.8390 - val_acc: 0.8606\n",
"\n",
"Epoch 00628: val_acc did not improve from 0.86765\n",
"Epoch 629/3000\n",
" - 40s - loss: 0.2961 - acc: 0.9639 - val_loss: 0.9099 - val_acc: 0.8556\n",
"\n",
"Epoch 00629: val_acc did not improve from 0.86765\n",
"Epoch 630/3000\n",
" - 39s - loss: 0.3082 - acc: 0.9594 - val_loss: 0.9248 - val_acc: 0.8517\n",
"\n",
"Epoch 00630: val_acc did not improve from 0.86765\n",
"Epoch 631/3000\n",
" - 40s - loss: 0.2951 - acc: 0.9651 - val_loss: 0.8578 - val_acc: 0.8571\n",
"\n",
"Epoch 00631: val_acc did not improve from 0.86765\n",
"Epoch 632/3000\n",
" - 39s - loss: 0.2976 - acc: 0.9635 - val_loss: 0.9248 - val_acc: 0.8501\n",
"\n",
"Epoch 00632: val_acc did not improve from 0.86765\n",
"Epoch 633/3000\n",
" - 39s - loss: 0.2952 - acc: 0.9662 - val_loss: 0.9488 - val_acc: 0.8486\n",
"\n",
"Epoch 00633: val_acc did not improve from 0.86765\n",
"Epoch 634/3000\n",
" - 39s - loss: 0.2984 - acc: 0.9651 - val_loss: 0.8497 - val_acc: 0.8595\n",
"\n",
"Epoch 00634: val_acc did not improve from 0.86765\n",
"Epoch 635/3000\n",
" - 39s - loss: 0.2962 - acc: 0.9627 - val_loss: 0.9404 - val_acc: 0.8490\n",
"\n",
"Epoch 00635: val_acc did not improve from 0.86765\n",
"Epoch 636/3000\n",
" - 39s - loss: 0.3032 - acc: 0.9617 - val_loss: 0.8722 - val_acc: 0.8583\n",
"\n",
"Epoch 00636: val_acc did not improve from 0.86765\n",
"Epoch 637/3000\n",
" - 39s - loss: 0.3138 - acc: 0.9592 - val_loss: 0.8681 - val_acc: 0.8533\n",
"\n",
"Epoch 00637: val_acc did not improve from 0.86765\n",
"Epoch 638/3000\n",
" - 39s - loss: 0.2964 - acc: 0.9638 - val_loss: 0.9499 - val_acc: 0.8513\n",
"\n",
"Epoch 00638: val_acc did not improve from 0.86765\n",
"Epoch 639/3000\n",
" - 39s - loss: 0.3060 - acc: 0.9618 - val_loss: 0.9046 - val_acc: 0.8536\n",
"\n",
"Epoch 00639: val_acc did not improve from 0.86765\n",
"Epoch 640/3000\n",
" - 39s - loss: 0.2985 - acc: 0.9601 - val_loss: 0.9265 - val_acc: 0.8552\n",
"\n",
"Epoch 00640: val_acc did not improve from 0.86765\n",
"Epoch 641/3000\n",
" - 39s - loss: 0.3079 - acc: 0.9600 - val_loss: 0.8866 - val_acc: 0.8579\n",
"\n",
"Epoch 00641: val_acc did not improve from 0.86765\n",
"Epoch 642/3000\n",
" - 39s - loss: 0.3059 - acc: 0.9606 - val_loss: 0.8764 - val_acc: 0.8595\n",
"\n",
"Epoch 00642: val_acc did not improve from 0.86765\n",
"Epoch 643/3000\n",
" - 39s - loss: 0.2926 - acc: 0.9636 - val_loss: 0.9022 - val_acc: 0.8521\n",
"\n",
"Epoch 00643: val_acc did not improve from 0.86765\n",
"Epoch 644/3000\n",
" - 39s - loss: 0.2894 - acc: 0.9656 - val_loss: 0.8993 - val_acc: 0.8544\n",
"\n",
"Epoch 00644: val_acc did not improve from 0.86765\n",
"\n",
"Epoch 00644: ReduceLROnPlateau reducing learning rate to 3.773536009248346e-05.\n",
"Epoch 645/3000\n",
" - 39s - loss: 0.2983 - acc: 0.9624 - val_loss: 0.8809 - val_acc: 0.8618\n",
"\n",
"Epoch 00645: val_acc did not improve from 0.86765\n",
"Epoch 646/3000\n",
" - 40s - loss: 0.2960 - acc: 0.9615 - val_loss: 0.8956 - val_acc: 0.8591\n",
"\n",
"Epoch 00646: val_acc did not improve from 0.86765\n",
"Epoch 647/3000\n",
" - 39s - loss: 0.3005 - acc: 0.9621 - val_loss: 0.9148 - val_acc: 0.8529\n",
"\n",
"Epoch 00647: val_acc did not improve from 0.86765\n",
"Epoch 648/3000\n",
" - 39s - loss: 0.3006 - acc: 0.9632 - val_loss: 0.9311 - val_acc: 0.8497\n",
"\n",
"Epoch 00648: val_acc did not improve from 0.86765\n",
"Epoch 649/3000\n",
" - 39s - loss: 0.3029 - acc: 0.9617 - val_loss: 0.9337 - val_acc: 0.8536\n",
"\n",
"Epoch 00649: val_acc did not improve from 0.86765\n",
"Epoch 650/3000\n",
" - 39s - loss: 0.3052 - acc: 0.9623 - val_loss: 0.8779 - val_acc: 0.8509\n",
"\n",
"Epoch 00650: val_acc did not improve from 0.86765\n",
"Epoch 651/3000\n",
" - 39s - loss: 0.2960 - acc: 0.9621 - val_loss: 0.9227 - val_acc: 0.8497\n",
"\n",
"Epoch 00651: val_acc did not improve from 0.86765\n",
"Epoch 652/3000\n",
" - 39s - loss: 0.3001 - acc: 0.9630 - val_loss: 0.9128 - val_acc: 0.8571\n",
"\n",
"Epoch 00652: val_acc did not improve from 0.86765\n",
"Epoch 653/3000\n",
" - 40s - loss: 0.2973 - acc: 0.9632 - val_loss: 0.9267 - val_acc: 0.8490\n",
"\n",
"Epoch 00653: val_acc did not improve from 0.86765\n",
"Epoch 654/3000\n",
" - 39s - loss: 0.2871 - acc: 0.9660 - val_loss: 0.9043 - val_acc: 0.8587\n",
"\n",
"Epoch 00654: val_acc did not improve from 0.86765\n",
"Epoch 655/3000\n",
" - 39s - loss: 0.3059 - acc: 0.9630 - val_loss: 0.9146 - val_acc: 0.8521\n",
"\n",
"Epoch 00655: val_acc did not improve from 0.86765\n",
"Epoch 656/3000\n",
" - 39s - loss: 0.2993 - acc: 0.9600 - val_loss: 0.9579 - val_acc: 0.8482\n",
"\n",
"Epoch 00656: val_acc did not improve from 0.86765\n",
"Epoch 657/3000\n",
" - 40s - loss: 0.3047 - acc: 0.9623 - val_loss: 0.8910 - val_acc: 0.8583\n",
"\n",
"Epoch 00657: val_acc did not improve from 0.86765\n",
"Epoch 658/3000\n",
" - 39s - loss: 0.3010 - acc: 0.9624 - val_loss: 0.9186 - val_acc: 0.8564\n",
"\n",
"Epoch 00658: val_acc did not improve from 0.86765\n",
"Epoch 659/3000\n",
" - 39s - loss: 0.3168 - acc: 0.9617 - val_loss: 0.8789 - val_acc: 0.8603\n",
"\n",
"Epoch 00659: val_acc did not improve from 0.86765\n",
"Epoch 660/3000\n",
" - 39s - loss: 0.2925 - acc: 0.9642 - val_loss: 0.9028 - val_acc: 0.8568\n",
"\n",
"Epoch 00660: val_acc did not improve from 0.86765\n",
"Epoch 661/3000\n",
" - 39s - loss: 0.2944 - acc: 0.9653 - val_loss: 0.9155 - val_acc: 0.8529\n",
"\n",
"Epoch 00661: val_acc did not improve from 0.86765\n",
"Epoch 662/3000\n",
" - 39s - loss: 0.2973 - acc: 0.9639 - val_loss: 0.8542 - val_acc: 0.8645\n",
"\n",
"Epoch 00662: val_acc did not improve from 0.86765\n",
"Epoch 663/3000\n",
" - 39s - loss: 0.3077 - acc: 0.9603 - val_loss: 0.8811 - val_acc: 0.8595\n",
"\n",
"Epoch 00663: val_acc did not improve from 0.86765\n",
"Epoch 664/3000\n",
" - 39s - loss: 0.2877 - acc: 0.9650 - val_loss: 0.8856 - val_acc: 0.8556\n",
"\n",
"Epoch 00664: val_acc did not improve from 0.86765\n",
"Epoch 665/3000\n",
" - 39s - loss: 0.3138 - acc: 0.9604 - val_loss: 0.8760 - val_acc: 0.8583\n",
"\n",
"Epoch 00665: val_acc did not improve from 0.86765\n",
"Epoch 666/3000\n",
" - 39s - loss: 0.2929 - acc: 0.9629 - val_loss: 0.8754 - val_acc: 0.8579\n",
"\n",
"Epoch 00666: val_acc did not improve from 0.86765\n",
"Epoch 667/3000\n",
" - 39s - loss: 0.3044 - acc: 0.9629 - val_loss: 0.8867 - val_acc: 0.8521\n",
"\n",
"Epoch 00667: val_acc did not improve from 0.86765\n",
"Epoch 668/3000\n",
" - 39s - loss: 0.2866 - acc: 0.9645 - val_loss: 0.8662 - val_acc: 0.8610\n",
"\n",
"Epoch 00668: val_acc did not improve from 0.86765\n",
"Epoch 669/3000\n",
" - 39s - loss: 0.3043 - acc: 0.9601 - val_loss: 0.9173 - val_acc: 0.8513\n",
"\n",
"Epoch 00669: val_acc did not improve from 0.86765\n",
"Epoch 670/3000\n",
" - 39s - loss: 0.2931 - acc: 0.9630 - val_loss: 0.9022 - val_acc: 0.8603\n",
"\n",
"Epoch 00670: val_acc did not improve from 0.86765\n",
"Epoch 671/3000\n",
" - 39s - loss: 0.2879 - acc: 0.9641 - val_loss: 0.9246 - val_acc: 0.8552\n",
"\n",
"Epoch 00671: val_acc did not improve from 0.86765\n",
"Epoch 672/3000\n",
" - 39s - loss: 0.2970 - acc: 0.9650 - val_loss: 0.9134 - val_acc: 0.8548\n",
"\n",
"Epoch 00672: val_acc did not improve from 0.86765\n",
"Epoch 673/3000\n",
" - 39s - loss: 0.3130 - acc: 0.9598 - val_loss: 0.9538 - val_acc: 0.8416\n",
"\n",
"Epoch 00673: val_acc did not improve from 0.86765\n",
"Epoch 674/3000\n",
" - 39s - loss: 0.2877 - acc: 0.9656 - val_loss: 0.9165 - val_acc: 0.8486\n",
"\n",
"Epoch 00674: val_acc did not improve from 0.86765\n",
"\n",
"Epoch 00674: ReduceLROnPlateau reducing learning rate to 3.584859277907526e-05.\n",
"Epoch 675/3000\n",
" - 39s - loss: 0.3036 - acc: 0.9598 - val_loss: 0.9019 - val_acc: 0.8529\n",
"\n",
"Epoch 00675: val_acc did not improve from 0.86765\n",
"Epoch 676/3000\n",
" - 39s - loss: 0.2976 - acc: 0.9627 - val_loss: 0.9121 - val_acc: 0.8513\n",
"\n",
"Epoch 00676: val_acc did not improve from 0.86765\n",
"Epoch 677/3000\n",
" - 39s - loss: 0.2968 - acc: 0.9601 - val_loss: 0.9235 - val_acc: 0.8494\n",
"\n",
"Epoch 00677: val_acc did not improve from 0.86765\n",
"Epoch 678/3000\n",
" - 40s - loss: 0.2980 - acc: 0.9641 - val_loss: 0.9238 - val_acc: 0.8447\n",
"\n",
"Epoch 00678: val_acc did not improve from 0.86765\n",
"Epoch 679/3000\n",
" - 39s - loss: 0.3045 - acc: 0.9641 - val_loss: 0.9372 - val_acc: 0.8447\n",
"\n",
"Epoch 00679: val_acc did not improve from 0.86765\n",
"Epoch 680/3000\n",
" - 39s - loss: 0.2936 - acc: 0.9653 - val_loss: 0.9538 - val_acc: 0.8466\n",
"\n",
"Epoch 00680: val_acc did not improve from 0.86765\n",
"Epoch 681/3000\n",
" - 40s - loss: 0.2835 - acc: 0.9674 - val_loss: 0.8736 - val_acc: 0.8603\n",
"\n",
"Epoch 00681: val_acc did not improve from 0.86765\n",
"Epoch 682/3000\n",
" - 39s - loss: 0.2869 - acc: 0.9647 - val_loss: 0.9875 - val_acc: 0.8447\n",
"\n",
"Epoch 00682: val_acc did not improve from 0.86765\n",
"Epoch 683/3000\n",
" - 39s - loss: 0.3113 - acc: 0.9603 - val_loss: 0.8752 - val_acc: 0.8536\n",
"\n",
"Epoch 00683: val_acc did not improve from 0.86765\n",
"Epoch 684/3000\n",
" - 39s - loss: 0.2964 - acc: 0.9657 - val_loss: 0.8664 - val_acc: 0.8579\n",
"\n",
"Epoch 00684: val_acc did not improve from 0.86765\n",
"Epoch 685/3000\n",
" - 40s - loss: 0.2932 - acc: 0.9638 - val_loss: 0.8569 - val_acc: 0.8583\n",
"\n",
"Epoch 00685: val_acc did not improve from 0.86765\n",
"Epoch 686/3000\n",
" - 39s - loss: 0.3064 - acc: 0.9621 - val_loss: 0.9977 - val_acc: 0.8455\n",
"\n",
"Epoch 00686: val_acc did not improve from 0.86765\n",
"Epoch 687/3000\n",
" - 39s - loss: 0.3080 - acc: 0.9612 - val_loss: 0.9792 - val_acc: 0.8459\n",
"\n",
"Epoch 00687: val_acc did not improve from 0.86765\n",
"Epoch 688/3000\n",
" - 39s - loss: 0.2962 - acc: 0.9662 - val_loss: 0.9532 - val_acc: 0.8533\n",
"\n",
"Epoch 00688: val_acc did not improve from 0.86765\n",
"Epoch 689/3000\n",
" - 39s - loss: 0.2952 - acc: 0.9656 - val_loss: 0.9066 - val_acc: 0.8556\n",
"\n",
"Epoch 00689: val_acc did not improve from 0.86765\n",
"Epoch 690/3000\n",
" - 39s - loss: 0.2913 - acc: 0.9633 - val_loss: 0.8976 - val_acc: 0.8564\n",
"\n",
"Epoch 00690: val_acc did not improve from 0.86765\n",
"Epoch 691/3000\n",
" - 39s - loss: 0.2921 - acc: 0.9626 - val_loss: 0.9137 - val_acc: 0.8556\n",
"\n",
"Epoch 00691: val_acc did not improve from 0.86765\n",
"Epoch 692/3000\n",
" - 39s - loss: 0.2933 - acc: 0.9644 - val_loss: 0.8747 - val_acc: 0.8579\n",
"\n",
"Epoch 00692: val_acc did not improve from 0.86765\n",
"Epoch 693/3000\n",
" - 39s - loss: 0.2911 - acc: 0.9657 - val_loss: 0.9050 - val_acc: 0.8540\n",
"\n",
"Epoch 00693: val_acc did not improve from 0.86765\n",
"Epoch 694/3000\n",
" - 39s - loss: 0.2948 - acc: 0.9650 - val_loss: 0.8901 - val_acc: 0.8568\n",
"\n",
"Epoch 00694: val_acc did not improve from 0.86765\n",
"Epoch 695/3000\n",
" - 39s - loss: 0.2893 - acc: 0.9642 - val_loss: 0.9297 - val_acc: 0.8509\n",
"\n",
"Epoch 00695: val_acc did not improve from 0.86765\n",
"Epoch 696/3000\n",
" - 39s - loss: 0.2954 - acc: 0.9630 - val_loss: 0.9187 - val_acc: 0.8513\n",
"\n",
"Epoch 00696: val_acc did not improve from 0.86765\n",
"Epoch 697/3000\n",
" - 39s - loss: 0.2788 - acc: 0.9677 - val_loss: 0.9193 - val_acc: 0.8525\n",
"\n",
"Epoch 00697: val_acc did not improve from 0.86765\n",
"Epoch 698/3000\n",
" - 40s - loss: 0.2841 - acc: 0.9654 - val_loss: 0.9162 - val_acc: 0.8540\n",
"\n",
"Epoch 00698: val_acc did not improve from 0.86765\n",
"Epoch 699/3000\n",
" - 40s - loss: 0.2931 - acc: 0.9635 - val_loss: 0.9340 - val_acc: 0.8533\n",
"\n",
"Epoch 00699: val_acc did not improve from 0.86765\n",
"Epoch 700/3000\n",
" - 39s - loss: 0.2893 - acc: 0.9651 - val_loss: 0.8903 - val_acc: 0.8583\n",
"\n",
"Epoch 00700: val_acc did not improve from 0.86765\n",
"Epoch 701/3000\n",
" - 40s - loss: 0.3039 - acc: 0.9606 - val_loss: 0.9014 - val_acc: 0.8548\n",
"\n",
"Epoch 00701: val_acc did not improve from 0.86765\n",
"Epoch 702/3000\n",
" - 39s - loss: 0.2943 - acc: 0.9648 - val_loss: 0.9103 - val_acc: 0.8513\n",
"\n",
"Epoch 00702: val_acc did not improve from 0.86765\n",
"Epoch 703/3000\n",
" - 39s - loss: 0.2827 - acc: 0.9663 - val_loss: 0.9195 - val_acc: 0.8513\n",
"\n",
"Epoch 00703: val_acc did not improve from 0.86765\n",
"Epoch 704/3000\n",
" - 39s - loss: 0.2919 - acc: 0.9632 - val_loss: 0.9178 - val_acc: 0.8513\n",
"\n",
"Epoch 00704: val_acc did not improve from 0.86765\n",
"\n",
"Epoch 00704: ReduceLROnPlateau reducing learning rate to 3.4056162621709515e-05.\n",
"Epoch 705/3000\n",
" - 39s - loss: 0.2836 - acc: 0.9678 - val_loss: 0.9168 - val_acc: 0.8548\n",
"\n",
"Epoch 00705: val_acc did not improve from 0.86765\n",
"Epoch 706/3000\n",
" - 39s - loss: 0.2895 - acc: 0.9648 - val_loss: 0.9295 - val_acc: 0.8455\n",
"\n",
"Epoch 00706: val_acc did not improve from 0.86765\n",
"Epoch 707/3000\n",
" - 40s - loss: 0.2835 - acc: 0.9654 - val_loss: 0.9059 - val_acc: 0.8517\n",
"\n",
"Epoch 00707: val_acc did not improve from 0.86765\n",
"Epoch 708/3000\n",
" - 40s - loss: 0.2936 - acc: 0.9641 - val_loss: 0.8736 - val_acc: 0.8571\n",
"\n",
"Epoch 00708: val_acc did not improve from 0.86765\n",
"Epoch 709/3000\n",
" - 39s - loss: 0.2918 - acc: 0.9656 - val_loss: 0.9068 - val_acc: 0.8552\n",
"\n",
"Epoch 00709: val_acc did not improve from 0.86765\n",
"Epoch 710/3000\n",
" - 39s - loss: 0.2994 - acc: 0.9609 - val_loss: 0.9029 - val_acc: 0.8556\n",
"\n",
"Epoch 00710: val_acc did not improve from 0.86765\n",
"Epoch 711/3000\n",
" - 39s - loss: 0.2889 - acc: 0.9642 - val_loss: 0.8765 - val_acc: 0.8575\n",
"\n",
"Epoch 00711: val_acc did not improve from 0.86765\n",
"Epoch 712/3000\n",
" - 39s - loss: 0.2862 - acc: 0.9663 - val_loss: 0.8791 - val_acc: 0.8575\n",
"\n",
"Epoch 00712: val_acc did not improve from 0.86765\n",
"Epoch 713/3000\n",
" - 39s - loss: 0.2796 - acc: 0.9677 - val_loss: 0.9159 - val_acc: 0.8564\n",
"\n",
"Epoch 00713: val_acc did not improve from 0.86765\n",
"Epoch 714/3000\n",
" - 39s - loss: 0.2880 - acc: 0.9633 - val_loss: 0.8778 - val_acc: 0.8587\n",
"\n",
"Epoch 00714: val_acc did not improve from 0.86765\n",
"Epoch 715/3000\n",
" - 39s - loss: 0.3003 - acc: 0.9618 - val_loss: 0.9154 - val_acc: 0.8536\n",
"\n",
"Epoch 00715: val_acc did not improve from 0.86765\n",
"Epoch 716/3000\n",
" - 39s - loss: 0.2912 - acc: 0.9620 - val_loss: 0.9004 - val_acc: 0.8529\n",
"\n",
"Epoch 00716: val_acc did not improve from 0.86765\n",
"Epoch 717/3000\n",
" - 39s - loss: 0.2935 - acc: 0.9645 - val_loss: 0.9077 - val_acc: 0.8521\n",
"\n",
"Epoch 00717: val_acc did not improve from 0.86765\n",
"Epoch 718/3000\n",
" - 39s - loss: 0.2992 - acc: 0.9620 - val_loss: 0.9111 - val_acc: 0.8544\n",
"\n",
"Epoch 00718: val_acc did not improve from 0.86765\n",
"Epoch 719/3000\n",
" - 39s - loss: 0.2827 - acc: 0.9681 - val_loss: 0.8965 - val_acc: 0.8583\n",
"\n",
"Epoch 00719: val_acc did not improve from 0.86765\n",
"Epoch 720/3000\n",
" - 39s - loss: 0.2885 - acc: 0.9629 - val_loss: 0.8949 - val_acc: 0.8544\n",
"\n",
"Epoch 00720: val_acc did not improve from 0.86765\n",
"Epoch 721/3000\n",
" - 39s - loss: 0.2937 - acc: 0.9650 - val_loss: 0.8743 - val_acc: 0.8610\n",
"\n",
"Epoch 00721: val_acc did not improve from 0.86765\n",
"Epoch 722/3000\n",
" - 39s - loss: 0.2873 - acc: 0.9647 - val_loss: 0.9313 - val_acc: 0.8560\n",
"\n",
"Epoch 00722: val_acc did not improve from 0.86765\n",
"Epoch 723/3000\n",
" - 39s - loss: 0.3143 - acc: 0.9579 - val_loss: 0.8754 - val_acc: 0.8548\n",
"\n",
"Epoch 00723: val_acc did not improve from 0.86765\n",
"Epoch 724/3000\n",
" - 39s - loss: 0.2988 - acc: 0.9609 - val_loss: 0.8939 - val_acc: 0.8548\n",
"\n",
"Epoch 00724: val_acc did not improve from 0.86765\n",
"Epoch 725/3000\n",
" - 39s - loss: 0.2950 - acc: 0.9623 - val_loss: 0.8939 - val_acc: 0.8560\n",
"\n",
"Epoch 00725: val_acc did not improve from 0.86765\n",
"Epoch 726/3000\n",
" - 39s - loss: 0.2831 - acc: 0.9680 - val_loss: 0.9072 - val_acc: 0.8544\n",
"\n",
"Epoch 00726: val_acc did not improve from 0.86765\n",
"Epoch 727/3000\n",
" - 39s - loss: 0.2914 - acc: 0.9660 - val_loss: 0.9478 - val_acc: 0.8478\n",
"\n",
"Epoch 00727: val_acc did not improve from 0.86765\n",
"Epoch 728/3000\n",
" - 39s - loss: 0.2890 - acc: 0.9651 - val_loss: 0.8874 - val_acc: 0.8564\n",
"\n",
"Epoch 00728: val_acc did not improve from 0.86765\n",
"Epoch 729/3000\n",
" - 39s - loss: 0.2919 - acc: 0.9644 - val_loss: 0.9314 - val_acc: 0.8521\n",
"\n",
"Epoch 00729: val_acc did not improve from 0.86765\n",
"Epoch 730/3000\n",
" - 39s - loss: 0.2971 - acc: 0.9609 - val_loss: 0.8941 - val_acc: 0.8513\n",
"\n",
"Epoch 00730: val_acc did not improve from 0.86765\n",
"Epoch 731/3000\n",
" - 39s - loss: 0.2879 - acc: 0.9644 - val_loss: 0.9337 - val_acc: 0.8486\n",
"\n",
"Epoch 00731: val_acc did not improve from 0.86765\n",
"Epoch 732/3000\n",
" - 39s - loss: 0.2856 - acc: 0.9647 - val_loss: 0.9427 - val_acc: 0.8486\n",
"\n",
"Epoch 00732: val_acc did not improve from 0.86765\n",
"Epoch 733/3000\n",
" - 39s - loss: 0.2878 - acc: 0.9639 - val_loss: 0.8807 - val_acc: 0.8536\n",
"\n",
"Epoch 00733: val_acc did not improve from 0.86765\n",
"Epoch 734/3000\n",
" - 39s - loss: 0.2843 - acc: 0.9659 - val_loss: 0.9139 - val_acc: 0.8571\n",
"\n",
"Epoch 00734: val_acc did not improve from 0.86765\n",
"\n",
"Epoch 00734: ReduceLROnPlateau reducing learning rate to 3.2353355527448004e-05.\n",
"Epoch 735/3000\n",
" - 39s - loss: 0.2950 - acc: 0.9642 - val_loss: 0.9057 - val_acc: 0.8540\n",
"\n",
"Epoch 00735: val_acc did not improve from 0.86765\n",
"Epoch 736/3000\n",
" - 39s - loss: 0.2941 - acc: 0.9650 - val_loss: 0.9524 - val_acc: 0.8494\n",
"\n",
"Epoch 00736: val_acc did not improve from 0.86765\n",
"Epoch 737/3000\n",
" - 39s - loss: 0.2927 - acc: 0.9642 - val_loss: 0.8909 - val_acc: 0.8529\n",
"\n",
"Epoch 00737: val_acc did not improve from 0.86765\n",
"Epoch 738/3000\n",
" - 39s - loss: 0.2851 - acc: 0.9650 - val_loss: 0.9237 - val_acc: 0.8474\n",
"\n",
"Epoch 00738: val_acc did not improve from 0.86765\n",
"Epoch 739/3000\n",
" - 39s - loss: 0.3011 - acc: 0.9635 - val_loss: 0.9081 - val_acc: 0.8490\n",
"\n",
"Epoch 00739: val_acc did not improve from 0.86765\n",
"Epoch 740/3000\n",
" - 39s - loss: 0.3006 - acc: 0.9612 - val_loss: 0.8786 - val_acc: 0.8540\n",
"\n",
"Epoch 00740: val_acc did not improve from 0.86765\n",
"Epoch 741/3000\n",
" - 39s - loss: 0.2911 - acc: 0.9650 - val_loss: 0.9304 - val_acc: 0.8466\n",
"\n",
"Epoch 00741: val_acc did not improve from 0.86765\n",
"Epoch 742/3000\n",
" - 40s - loss: 0.2882 - acc: 0.9657 - val_loss: 0.9545 - val_acc: 0.8478\n",
"\n",
"Epoch 00742: val_acc did not improve from 0.86765\n",
"Epoch 743/3000\n",
" - 39s - loss: 0.2863 - acc: 0.9632 - val_loss: 0.9024 - val_acc: 0.8525\n",
"\n",
"Epoch 00743: val_acc did not improve from 0.86765\n",
"Epoch 744/3000\n",
" - 39s - loss: 0.2837 - acc: 0.9647 - val_loss: 0.9288 - val_acc: 0.8525\n",
"\n",
"Epoch 00744: val_acc did not improve from 0.86765\n",
"Epoch 745/3000\n",
" - 39s - loss: 0.2848 - acc: 0.9666 - val_loss: 0.9740 - val_acc: 0.8478\n",
"\n",
"Epoch 00745: val_acc did not improve from 0.86765\n",
"Epoch 746/3000\n",
" - 39s - loss: 0.2994 - acc: 0.9647 - val_loss: 0.9326 - val_acc: 0.8521\n",
"\n",
"Epoch 00746: val_acc did not improve from 0.86765\n",
"Epoch 747/3000\n",
" - 39s - loss: 0.2904 - acc: 0.9644 - val_loss: 0.9005 - val_acc: 0.8544\n",
"\n",
"Epoch 00747: val_acc did not improve from 0.86765\n",
"Epoch 748/3000\n",
" - 39s - loss: 0.2829 - acc: 0.9650 - val_loss: 0.9139 - val_acc: 0.8513\n",
"\n",
"Epoch 00748: val_acc did not improve from 0.86765\n",
"Epoch 749/3000\n",
" - 39s - loss: 0.2870 - acc: 0.9635 - val_loss: 0.9250 - val_acc: 0.8505\n",
"\n",
"Epoch 00749: val_acc did not improve from 0.86765\n",
"Epoch 750/3000\n",
" - 39s - loss: 0.2867 - acc: 0.9639 - val_loss: 0.8846 - val_acc: 0.8575\n",
"\n",
"Epoch 00750: val_acc did not improve from 0.86765\n",
"Epoch 751/3000\n",
" - 39s - loss: 0.2881 - acc: 0.9642 - val_loss: 0.9549 - val_acc: 0.8462\n",
"\n",
"Epoch 00751: val_acc did not improve from 0.86765\n",
"Epoch 752/3000\n",
" - 40s - loss: 0.3015 - acc: 0.9627 - val_loss: 0.9473 - val_acc: 0.8455\n",
"\n",
"Epoch 00752: val_acc did not improve from 0.86765\n",
"Epoch 753/3000\n",
" - 39s - loss: 0.2887 - acc: 0.9639 - val_loss: 0.9583 - val_acc: 0.8455\n",
"\n",
"Epoch 00753: val_acc did not improve from 0.86765\n",
"Epoch 754/3000\n",
" - 39s - loss: 0.2826 - acc: 0.9654 - val_loss: 0.9187 - val_acc: 0.8552\n",
"\n",
"Epoch 00754: val_acc did not improve from 0.86765\n",
"Epoch 755/3000\n",
" - 39s - loss: 0.2901 - acc: 0.9633 - val_loss: 0.9473 - val_acc: 0.8536\n",
"\n",
"Epoch 00755: val_acc did not improve from 0.86765\n",
"Epoch 756/3000\n",
" - 39s - loss: 0.2856 - acc: 0.9657 - val_loss: 0.9160 - val_acc: 0.8544\n",
"\n",
"Epoch 00756: val_acc did not improve from 0.86765\n",
"Epoch 757/3000\n",
" - 39s - loss: 0.2852 - acc: 0.9648 - val_loss: 0.8796 - val_acc: 0.8560\n",
"\n",
"Epoch 00757: val_acc did not improve from 0.86765\n",
"Epoch 758/3000\n",
" - 39s - loss: 0.2768 - acc: 0.9672 - val_loss: 0.9374 - val_acc: 0.8552\n",
"\n",
"Epoch 00758: val_acc did not improve from 0.86765\n",
"Epoch 759/3000\n",
" - 39s - loss: 0.2989 - acc: 0.9626 - val_loss: 0.9388 - val_acc: 0.8540\n",
"\n",
"Epoch 00759: val_acc did not improve from 0.86765\n",
"Epoch 760/3000\n",
" - 39s - loss: 0.2904 - acc: 0.9630 - val_loss: 0.9740 - val_acc: 0.8474\n",
"\n",
"Epoch 00760: val_acc did not improve from 0.86765\n",
"Epoch 761/3000\n",
" - 40s - loss: 0.2885 - acc: 0.9642 - val_loss: 0.9901 - val_acc: 0.8431\n",
"\n",
"Epoch 00761: val_acc did not improve from 0.86765\n",
"Epoch 762/3000\n",
" - 39s - loss: 0.2759 - acc: 0.9662 - val_loss: 0.9343 - val_acc: 0.8513\n",
"\n",
"Epoch 00762: val_acc did not improve from 0.86765\n",
"Epoch 763/3000\n",
" - 39s - loss: 0.2832 - acc: 0.9642 - val_loss: 0.9455 - val_acc: 0.8529\n",
"\n",
"Epoch 00763: val_acc did not improve from 0.86765\n",
"Epoch 764/3000\n",
" - 39s - loss: 0.2840 - acc: 0.9635 - val_loss: 0.9037 - val_acc: 0.8525\n",
"\n",
"Epoch 00764: val_acc did not improve from 0.86765\n",
"\n",
"Epoch 00764: ReduceLROnPlateau reducing learning rate to 3.0735688960703554e-05.\n",
"Epoch 765/3000\n",
" - 39s - loss: 0.2901 - acc: 0.9629 - val_loss: 0.9094 - val_acc: 0.8497\n",
"\n",
"Epoch 00765: val_acc did not improve from 0.86765\n",
"Epoch 766/3000\n",
" - 39s - loss: 0.2887 - acc: 0.9629 - val_loss: 0.9508 - val_acc: 0.8494\n",
"\n",
"Epoch 00766: val_acc did not improve from 0.86765\n",
"Epoch 767/3000\n",
" - 39s - loss: 0.2766 - acc: 0.9669 - val_loss: 0.9100 - val_acc: 0.8552\n",
"\n",
"Epoch 00767: val_acc did not improve from 0.86765\n",
"Epoch 768/3000\n",
" - 39s - loss: 0.2883 - acc: 0.9639 - val_loss: 0.8800 - val_acc: 0.8595\n",
"\n",
"Epoch 00768: val_acc did not improve from 0.86765\n",
"Epoch 769/3000\n",
" - 40s - loss: 0.2919 - acc: 0.9633 - val_loss: 0.8850 - val_acc: 0.8568\n",
"\n",
"Epoch 00769: val_acc did not improve from 0.86765\n",
"Epoch 770/3000\n",
" - 39s - loss: 0.2937 - acc: 0.9633 - val_loss: 0.9037 - val_acc: 0.8548\n",
"\n",
"Epoch 00770: val_acc did not improve from 0.86765\n",
"Epoch 771/3000\n",
" - 39s - loss: 0.2788 - acc: 0.9659 - val_loss: 0.9109 - val_acc: 0.8579\n",
"\n",
"Epoch 00771: val_acc did not improve from 0.86765\n",
"Epoch 772/3000\n",
" - 39s - loss: 0.2797 - acc: 0.9666 - val_loss: 0.9358 - val_acc: 0.8505\n",
"\n",
"Epoch 00772: val_acc did not improve from 0.86765\n",
"Epoch 773/3000\n",
" - 39s - loss: 0.2813 - acc: 0.9662 - val_loss: 0.9673 - val_acc: 0.8490\n",
"\n",
"Epoch 00773: val_acc did not improve from 0.86765\n",
"Epoch 774/3000\n",
" - 39s - loss: 0.2857 - acc: 0.9645 - val_loss: 0.9170 - val_acc: 0.8525\n",
"\n",
"Epoch 00774: val_acc did not improve from 0.86765\n",
"Epoch 775/3000\n",
" - 39s - loss: 0.2743 - acc: 0.9645 - val_loss: 0.9429 - val_acc: 0.8536\n",
"\n",
"Epoch 00775: val_acc did not improve from 0.86765\n",
"Epoch 776/3000\n",
" - 39s - loss: 0.2823 - acc: 0.9632 - val_loss: 0.9266 - val_acc: 0.8583\n",
"\n",
"Epoch 00776: val_acc did not improve from 0.86765\n",
"Epoch 777/3000\n",
" - 39s - loss: 0.3024 - acc: 0.9600 - val_loss: 0.9550 - val_acc: 0.8548\n",
"\n",
"Epoch 00777: val_acc did not improve from 0.86765\n",
"Epoch 778/3000\n",
" - 39s - loss: 0.2889 - acc: 0.9632 - val_loss: 0.9114 - val_acc: 0.8579\n",
"\n",
"Epoch 00778: val_acc did not improve from 0.86765\n",
"Epoch 779/3000\n",
" - 39s - loss: 0.2823 - acc: 0.9648 - val_loss: 0.8880 - val_acc: 0.8610\n",
"\n",
"Epoch 00779: val_acc did not improve from 0.86765\n",
"Epoch 780/3000\n",
" - 39s - loss: 0.2770 - acc: 0.9681 - val_loss: 0.9010 - val_acc: 0.8579\n",
"\n",
"Epoch 00780: val_acc did not improve from 0.86765\n",
"Epoch 781/3000\n",
" - 39s - loss: 0.2822 - acc: 0.9651 - val_loss: 0.9331 - val_acc: 0.8509\n",
"\n",
"Epoch 00781: val_acc did not improve from 0.86765\n",
"Epoch 782/3000\n",
" - 39s - loss: 0.2738 - acc: 0.9688 - val_loss: 0.9178 - val_acc: 0.8544\n",
"\n",
"Epoch 00782: val_acc did not improve from 0.86765\n",
"Epoch 783/3000\n",
" - 40s - loss: 0.2764 - acc: 0.9684 - val_loss: 0.8969 - val_acc: 0.8560\n",
"\n",
"Epoch 00783: val_acc did not improve from 0.86765\n",
"Epoch 784/3000\n",
" - 39s - loss: 0.2783 - acc: 0.9669 - val_loss: 0.9362 - val_acc: 0.8556\n",
"\n",
"Epoch 00784: val_acc did not improve from 0.86765\n",
"Epoch 785/3000\n",
" - 39s - loss: 0.2815 - acc: 0.9660 - val_loss: 0.9326 - val_acc: 0.8474\n",
"\n",
"Epoch 00785: val_acc did not improve from 0.86765\n",
"Epoch 786/3000\n",
" - 39s - loss: 0.2779 - acc: 0.9656 - val_loss: 0.9801 - val_acc: 0.8497\n",
"\n",
"Epoch 00786: val_acc did not improve from 0.86765\n",
"Epoch 787/3000\n",
" - 39s - loss: 0.2955 - acc: 0.9638 - val_loss: 0.9417 - val_acc: 0.8505\n",
"\n",
"Epoch 00787: val_acc did not improve from 0.86765\n",
"Epoch 788/3000\n",
" - 39s - loss: 0.2904 - acc: 0.9635 - val_loss: 0.9325 - val_acc: 0.8505\n",
"\n",
"Epoch 00788: val_acc did not improve from 0.86765\n",
"Epoch 789/3000\n",
" - 40s - loss: 0.2880 - acc: 0.9651 - val_loss: 0.9173 - val_acc: 0.8513\n",
"\n",
"Epoch 00789: val_acc did not improve from 0.86765\n",
"Epoch 790/3000\n",
" - 39s - loss: 0.2861 - acc: 0.9642 - val_loss: 0.8816 - val_acc: 0.8591\n",
"\n",
"Epoch 00790: val_acc did not improve from 0.86765\n",
"Epoch 791/3000\n",
" - 40s - loss: 0.2830 - acc: 0.9644 - val_loss: 0.8885 - val_acc: 0.8638\n",
"\n",
"Epoch 00791: val_acc did not improve from 0.86765\n",
"Epoch 792/3000\n",
" - 39s - loss: 0.2820 - acc: 0.9660 - val_loss: 0.9431 - val_acc: 0.8529\n",
"\n",
"Epoch 00792: val_acc did not improve from 0.86765\n",
"Epoch 793/3000\n",
" - 39s - loss: 0.2788 - acc: 0.9680 - val_loss: 0.9015 - val_acc: 0.8556\n",
"\n",
"Epoch 00793: val_acc did not improve from 0.86765\n",
"Epoch 794/3000\n",
" - 39s - loss: 0.2761 - acc: 0.9668 - val_loss: 0.9139 - val_acc: 0.8513\n",
"\n",
"Epoch 00794: val_acc did not improve from 0.86765\n",
"\n",
"Epoch 00794: ReduceLROnPlateau reducing learning rate to 2.9198905031080356e-05.\n",
"Epoch 795/3000\n",
" - 40s - loss: 0.2831 - acc: 0.9654 - val_loss: 0.9262 - val_acc: 0.8571\n",
"\n",
"Epoch 00795: val_acc did not improve from 0.86765\n",
"Epoch 796/3000\n",
" - 39s - loss: 0.2777 - acc: 0.9659 - val_loss: 0.9238 - val_acc: 0.8536\n",
"\n",
"Epoch 00796: val_acc did not improve from 0.86765\n",
"Epoch 797/3000\n",
" - 39s - loss: 0.2825 - acc: 0.9683 - val_loss: 0.9148 - val_acc: 0.8548\n",
"\n",
"Epoch 00797: val_acc did not improve from 0.86765\n",
"Epoch 798/3000\n",
" - 39s - loss: 0.2807 - acc: 0.9662 - val_loss: 0.9590 - val_acc: 0.8497\n",
"\n",
"Epoch 00798: val_acc did not improve from 0.86765\n",
"Epoch 799/3000\n",
" - 40s - loss: 0.2864 - acc: 0.9662 - val_loss: 0.9451 - val_acc: 0.8536\n",
"\n",
"Epoch 00799: val_acc did not improve from 0.86765\n",
"Epoch 800/3000\n",
" - 39s - loss: 0.2845 - acc: 0.9648 - val_loss: 0.9250 - val_acc: 0.8533\n",
"\n",
"Epoch 00800: val_acc did not improve from 0.86765\n",
"Epoch 801/3000\n",
" - 39s - loss: 0.2791 - acc: 0.9671 - val_loss: 0.9420 - val_acc: 0.8552\n",
"\n",
"Epoch 00801: val_acc did not improve from 0.86765\n",
"Epoch 802/3000\n",
" - 39s - loss: 0.2731 - acc: 0.9680 - val_loss: 0.9543 - val_acc: 0.8494\n",
"\n",
"Epoch 00802: val_acc did not improve from 0.86765\n",
"Epoch 803/3000\n",
" - 39s - loss: 0.2699 - acc: 0.9689 - val_loss: 0.8947 - val_acc: 0.8610\n",
"\n",
"Epoch 00803: val_acc did not improve from 0.86765\n",
"Epoch 804/3000\n",
" - 39s - loss: 0.2867 - acc: 0.9639 - val_loss: 0.9387 - val_acc: 0.8509\n",
"\n",
"Epoch 00804: val_acc did not improve from 0.86765\n",
"Epoch 805/3000\n",
" - 40s - loss: 0.2954 - acc: 0.9621 - val_loss: 0.9670 - val_acc: 0.8521\n",
"\n",
"Epoch 00805: val_acc did not improve from 0.86765\n",
"Epoch 806/3000\n",
" - 39s - loss: 0.2764 - acc: 0.9651 - val_loss: 0.9275 - val_acc: 0.8548\n",
"\n",
"Epoch 00806: val_acc did not improve from 0.86765\n",
"Epoch 807/3000\n",
" - 39s - loss: 0.2878 - acc: 0.9629 - val_loss: 0.9517 - val_acc: 0.8529\n",
"\n",
"Epoch 00807: val_acc did not improve from 0.86765\n",
"Epoch 808/3000\n",
" - 39s - loss: 0.2891 - acc: 0.9626 - val_loss: 0.8864 - val_acc: 0.8591\n",
"\n",
"Epoch 00808: val_acc did not improve from 0.86765\n",
"Epoch 809/3000\n",
" - 39s - loss: 0.2790 - acc: 0.9653 - val_loss: 0.8962 - val_acc: 0.8591\n",
"\n",
"Epoch 00809: val_acc did not improve from 0.86765\n",
"Epoch 810/3000\n",
" - 40s - loss: 0.2791 - acc: 0.9672 - val_loss: 0.9036 - val_acc: 0.8575\n",
"\n",
"Epoch 00810: val_acc did not improve from 0.86765\n",
"Epoch 811/3000\n",
" - 39s - loss: 0.2815 - acc: 0.9621 - val_loss: 0.8976 - val_acc: 0.8622\n",
"\n",
"Epoch 00811: val_acc did not improve from 0.86765\n",
"Epoch 812/3000\n",
" - 39s - loss: 0.2793 - acc: 0.9675 - val_loss: 0.8967 - val_acc: 0.8583\n",
"\n",
"Epoch 00812: val_acc did not improve from 0.86765\n",
"Epoch 813/3000\n",
" - 39s - loss: 0.2730 - acc: 0.9700 - val_loss: 0.9122 - val_acc: 0.8571\n",
"\n",
"Epoch 00813: val_acc did not improve from 0.86765\n",
"Epoch 814/3000\n",
" - 39s - loss: 0.2846 - acc: 0.9645 - val_loss: 0.9050 - val_acc: 0.8587\n",
"\n",
"Epoch 00814: val_acc did not improve from 0.86765\n",
"Epoch 815/3000\n",
" - 39s - loss: 0.2723 - acc: 0.9675 - val_loss: 0.9011 - val_acc: 0.8560\n",
"\n",
"Epoch 00815: val_acc did not improve from 0.86765\n",
"Epoch 816/3000\n",
" - 39s - loss: 0.2765 - acc: 0.9662 - val_loss: 0.9365 - val_acc: 0.8482\n",
"\n",
"Epoch 00816: val_acc did not improve from 0.86765\n",
"Epoch 817/3000\n",
" - 39s - loss: 0.2826 - acc: 0.9639 - val_loss: 0.9098 - val_acc: 0.8575\n",
"\n",
"Epoch 00817: val_acc did not improve from 0.86765\n",
"Epoch 818/3000\n",
" - 39s - loss: 0.2846 - acc: 0.9641 - val_loss: 0.9574 - val_acc: 0.8525\n",
"\n",
"Epoch 00818: val_acc did not improve from 0.86765\n",
"Epoch 819/3000\n",
" - 40s - loss: 0.2812 - acc: 0.9659 - val_loss: 0.9651 - val_acc: 0.8482\n",
"\n",
"Epoch 00819: val_acc did not improve from 0.86765\n",
"Epoch 820/3000\n",
" - 39s - loss: 0.2794 - acc: 0.9659 - val_loss: 0.9131 - val_acc: 0.8536\n",
"\n",
"Epoch 00820: val_acc did not improve from 0.86765\n",
"Epoch 821/3000\n",
" - 39s - loss: 0.2878 - acc: 0.9639 - val_loss: 1.0026 - val_acc: 0.8424\n",
"\n",
"Epoch 00821: val_acc did not improve from 0.86765\n",
"Epoch 822/3000\n",
" - 39s - loss: 0.2864 - acc: 0.9629 - val_loss: 0.9296 - val_acc: 0.8513\n",
"\n",
"Epoch 00822: val_acc did not improve from 0.86765\n",
"Epoch 823/3000\n",
" - 39s - loss: 0.2793 - acc: 0.9657 - val_loss: 0.9175 - val_acc: 0.8575\n",
"\n",
"Epoch 00823: val_acc did not improve from 0.86765\n",
"Epoch 824/3000\n",
" - 39s - loss: 0.2714 - acc: 0.9689 - val_loss: 0.9200 - val_acc: 0.8529\n",
"\n",
"Epoch 00824: val_acc did not improve from 0.86765\n",
"\n",
"Epoch 00824: ReduceLROnPlateau reducing learning rate to 2.7738960125134326e-05.\n",
"Epoch 825/3000\n",
" - 39s - loss: 0.2668 - acc: 0.9684 - val_loss: 0.9377 - val_acc: 0.8466\n",
"\n",
"Epoch 00825: val_acc did not improve from 0.86765\n",
"Epoch 826/3000\n",
" - 39s - loss: 0.2777 - acc: 0.9659 - val_loss: 0.9424 - val_acc: 0.8525\n",
"\n",
"Epoch 00826: val_acc did not improve from 0.86765\n",
"Epoch 827/3000\n",
" - 39s - loss: 0.2730 - acc: 0.9681 - val_loss: 0.9109 - val_acc: 0.8548\n",
"\n",
"Epoch 00827: val_acc did not improve from 0.86765\n",
"Epoch 828/3000\n",
" - 39s - loss: 0.2793 - acc: 0.9672 - val_loss: 0.8981 - val_acc: 0.8564\n",
"\n",
"Epoch 00828: val_acc did not improve from 0.86765\n",
"Epoch 829/3000\n",
" - 39s - loss: 0.2794 - acc: 0.9674 - val_loss: 0.8856 - val_acc: 0.8630\n",
"\n",
"Epoch 00829: val_acc did not improve from 0.86765\n",
"Epoch 830/3000\n",
" - 39s - loss: 0.2857 - acc: 0.9632 - val_loss: 0.8871 - val_acc: 0.8571\n",
"\n",
"Epoch 00830: val_acc did not improve from 0.86765\n",
"Epoch 831/3000\n",
" - 39s - loss: 0.2796 - acc: 0.9641 - val_loss: 0.9087 - val_acc: 0.8560\n",
"\n",
"Epoch 00831: val_acc did not improve from 0.86765\n",
"Epoch 832/3000\n",
" - 39s - loss: 0.2803 - acc: 0.9665 - val_loss: 0.9592 - val_acc: 0.8490\n",
"\n",
"Epoch 00832: val_acc did not improve from 0.86765\n",
"Epoch 833/3000\n",
" - 39s - loss: 0.2778 - acc: 0.9672 - val_loss: 0.9842 - val_acc: 0.8490\n",
"\n",
"Epoch 00833: val_acc did not improve from 0.86765\n",
"Epoch 834/3000\n",
" - 40s - loss: 0.2756 - acc: 0.9684 - val_loss: 0.9498 - val_acc: 0.8548\n",
"\n",
"Epoch 00834: val_acc did not improve from 0.86765\n",
"Epoch 835/3000\n",
" - 39s - loss: 0.2723 - acc: 0.9659 - val_loss: 0.8817 - val_acc: 0.8661\n",
"\n",
"Epoch 00835: val_acc did not improve from 0.86765\n",
"Epoch 836/3000\n",
" - 40s - loss: 0.2778 - acc: 0.9665 - val_loss: 0.9431 - val_acc: 0.8552\n",
"\n",
"Epoch 00836: val_acc did not improve from 0.86765\n",
"Epoch 837/3000\n",
" - 39s - loss: 0.2793 - acc: 0.9642 - val_loss: 0.9221 - val_acc: 0.8513\n",
"\n",
"Epoch 00837: val_acc did not improve from 0.86765\n",
"Epoch 838/3000\n",
" - 39s - loss: 0.2845 - acc: 0.9672 - val_loss: 0.9088 - val_acc: 0.8548\n",
"\n",
"Epoch 00838: val_acc did not improve from 0.86765\n",
"Epoch 839/3000\n",
" - 40s - loss: 0.2788 - acc: 0.9665 - val_loss: 0.9014 - val_acc: 0.8579\n",
"\n",
"Epoch 00839: val_acc did not improve from 0.86765\n",
"Epoch 840/3000\n",
" - 39s - loss: 0.2772 - acc: 0.9654 - val_loss: 0.9364 - val_acc: 0.8556\n",
"\n",
"Epoch 00840: val_acc did not improve from 0.86765\n",
"Epoch 841/3000\n",
" - 40s - loss: 0.2839 - acc: 0.9638 - val_loss: 0.9346 - val_acc: 0.8529\n",
"\n",
"Epoch 00841: val_acc did not improve from 0.86765\n",
"Epoch 842/3000\n",
" - 39s - loss: 0.2756 - acc: 0.9668 - val_loss: 0.9324 - val_acc: 0.8552\n",
"\n",
"Epoch 00842: val_acc did not improve from 0.86765\n",
"Epoch 843/3000\n",
" - 39s - loss: 0.2699 - acc: 0.9660 - val_loss: 0.9185 - val_acc: 0.8533\n",
"\n",
"Epoch 00843: val_acc did not improve from 0.86765\n",
"Epoch 844/3000\n",
" - 39s - loss: 0.2659 - acc: 0.9692 - val_loss: 0.9411 - val_acc: 0.8556\n",
"\n",
"Epoch 00844: val_acc did not improve from 0.86765\n",
"Epoch 845/3000\n",
" - 38s - loss: 0.2721 - acc: 0.9681 - val_loss: 0.9333 - val_acc: 0.8556\n",
"\n",
"Epoch 00845: val_acc did not improve from 0.86765\n",
"Epoch 846/3000\n",
" - 39s - loss: 0.2703 - acc: 0.9695 - val_loss: 0.9727 - val_acc: 0.8540\n",
"\n",
"Epoch 00846: val_acc did not improve from 0.86765\n",
"Epoch 847/3000\n",
" - 39s - loss: 0.2790 - acc: 0.9672 - val_loss: 0.9842 - val_acc: 0.8486\n",
"\n",
"Epoch 00847: val_acc did not improve from 0.86765\n",
"Epoch 848/3000\n",
" - 39s - loss: 0.2789 - acc: 0.9666 - val_loss: 0.8609 - val_acc: 0.8579\n",
"\n",
"Epoch 00848: val_acc did not improve from 0.86765\n",
"Epoch 849/3000\n",
" - 39s - loss: 0.2717 - acc: 0.9654 - val_loss: 0.9184 - val_acc: 0.8568\n",
"\n",
"Epoch 00849: val_acc did not improve from 0.86765\n",
"Epoch 850/3000\n",
" - 40s - loss: 0.2760 - acc: 0.9689 - val_loss: 0.8725 - val_acc: 0.8564\n",
"\n",
"Epoch 00850: val_acc did not improve from 0.86765\n",
"Epoch 851/3000\n",
" - 39s - loss: 0.2848 - acc: 0.9660 - val_loss: 0.9761 - val_acc: 0.8509\n",
"\n",
"Epoch 00851: val_acc did not improve from 0.86765\n",
"Epoch 852/3000\n",
" - 39s - loss: 0.2893 - acc: 0.9639 - val_loss: 0.9037 - val_acc: 0.8540\n",
"\n",
"Epoch 00852: val_acc did not improve from 0.86765\n",
"Epoch 853/3000\n",
" - 39s - loss: 0.2875 - acc: 0.9598 - val_loss: 0.9386 - val_acc: 0.8533\n",
"\n",
"Epoch 00853: val_acc did not improve from 0.86765\n",
"Epoch 854/3000\n",
" - 39s - loss: 0.2745 - acc: 0.9665 - val_loss: 0.9106 - val_acc: 0.8486\n",
"\n",
"Epoch 00854: val_acc did not improve from 0.86765\n",
"\n",
"Epoch 00854: ReduceLROnPlateau reducing learning rate to 2.6352012810093584e-05.\n",
"Epoch 855/3000\n",
" - 39s - loss: 0.2748 - acc: 0.9660 - val_loss: 0.8811 - val_acc: 0.8548\n",
"\n",
"Epoch 00855: val_acc did not improve from 0.86765\n",
"Epoch 856/3000\n",
" - 39s - loss: 0.2754 - acc: 0.9684 - val_loss: 0.9337 - val_acc: 0.8517\n",
"\n",
"Epoch 00856: val_acc did not improve from 0.86765\n",
"Epoch 857/3000\n",
" - 39s - loss: 0.2750 - acc: 0.9666 - val_loss: 0.9643 - val_acc: 0.8494\n",
"\n",
"Epoch 00857: val_acc did not improve from 0.86765\n",
"Epoch 858/3000\n",
" - 39s - loss: 0.2956 - acc: 0.9629 - val_loss: 0.9885 - val_acc: 0.8533\n",
"\n",
"Epoch 00858: val_acc did not improve from 0.86765\n",
"Epoch 859/3000\n",
" - 39s - loss: 0.2851 - acc: 0.9656 - val_loss: 0.9386 - val_acc: 0.8564\n",
"\n",
"Epoch 00859: val_acc did not improve from 0.86765\n",
"Epoch 860/3000\n",
" - 39s - loss: 0.2880 - acc: 0.9630 - val_loss: 0.8919 - val_acc: 0.8540\n",
"\n",
"Epoch 00860: val_acc did not improve from 0.86765\n",
"Epoch 861/3000\n",
" - 39s - loss: 0.2840 - acc: 0.9659 - val_loss: 0.9075 - val_acc: 0.8490\n",
"\n",
"Epoch 00861: val_acc did not improve from 0.86765\n",
"Epoch 862/3000\n",
" - 39s - loss: 0.2773 - acc: 0.9675 - val_loss: 0.9273 - val_acc: 0.8560\n",
"\n",
"Epoch 00862: val_acc did not improve from 0.86765\n",
"Epoch 863/3000\n",
" - 39s - loss: 0.2611 - acc: 0.9688 - val_loss: 0.9148 - val_acc: 0.8606\n",
"\n",
"Epoch 00863: val_acc did not improve from 0.86765\n",
"Epoch 864/3000\n",
" - 39s - loss: 0.2638 - acc: 0.9691 - val_loss: 0.9168 - val_acc: 0.8587\n",
"\n",
"Epoch 00864: val_acc did not improve from 0.86765\n",
"Epoch 865/3000\n",
" - 39s - loss: 0.2778 - acc: 0.9642 - val_loss: 0.9238 - val_acc: 0.8540\n",
"\n",
"Epoch 00865: val_acc did not improve from 0.86765\n",
"Epoch 866/3000\n",
" - 39s - loss: 0.2923 - acc: 0.9630 - val_loss: 0.9227 - val_acc: 0.8548\n",
"\n",
"Epoch 00866: val_acc did not improve from 0.86765\n",
"Epoch 867/3000\n",
" - 39s - loss: 0.2839 - acc: 0.9650 - val_loss: 0.9232 - val_acc: 0.8587\n",
"\n",
"Epoch 00867: val_acc did not improve from 0.86765\n",
"Epoch 868/3000\n",
" - 39s - loss: 0.2839 - acc: 0.9644 - val_loss: 0.8876 - val_acc: 0.8568\n",
"\n",
"Epoch 00868: val_acc did not improve from 0.86765\n",
"Epoch 869/3000\n",
" - 39s - loss: 0.2805 - acc: 0.9660 - val_loss: 0.8901 - val_acc: 0.8568\n",
"\n",
"Epoch 00869: val_acc did not improve from 0.86765\n",
"Epoch 870/3000\n",
" - 39s - loss: 0.2745 - acc: 0.9672 - val_loss: 0.9161 - val_acc: 0.8556\n",
"\n",
"Epoch 00870: val_acc did not improve from 0.86765\n",
"Epoch 871/3000\n",
" - 39s - loss: 0.2767 - acc: 0.9656 - val_loss: 0.9435 - val_acc: 0.8571\n",
"\n",
"Epoch 00871: val_acc did not improve from 0.86765\n",
"Epoch 872/3000\n",
" - 39s - loss: 0.2778 - acc: 0.9665 - val_loss: 0.9252 - val_acc: 0.8540\n",
"\n",
"Epoch 00872: val_acc did not improve from 0.86765\n",
"Epoch 873/3000\n",
" - 38s - loss: 0.2818 - acc: 0.9642 - val_loss: 0.9597 - val_acc: 0.8548\n",
"\n",
"Epoch 00873: val_acc did not improve from 0.86765\n",
"Epoch 874/3000\n",
" - 39s - loss: 0.2685 - acc: 0.9681 - val_loss: 0.9170 - val_acc: 0.8583\n",
"\n",
"Epoch 00874: val_acc did not improve from 0.86765\n",
"Epoch 875/3000\n",
" - 40s - loss: 0.2723 - acc: 0.9666 - val_loss: 0.9440 - val_acc: 0.8517\n",
"\n",
"Epoch 00875: val_acc did not improve from 0.86765\n",
"Epoch 876/3000\n",
" - 39s - loss: 0.2628 - acc: 0.9681 - val_loss: 0.9244 - val_acc: 0.8540\n",
"\n",
"Epoch 00876: val_acc did not improve from 0.86765\n",
"Epoch 877/3000\n",
" - 39s - loss: 0.2675 - acc: 0.9688 - val_loss: 0.9393 - val_acc: 0.8540\n",
"\n",
"Epoch 00877: val_acc did not improve from 0.86765\n",
"Epoch 878/3000\n",
" - 39s - loss: 0.2735 - acc: 0.9681 - val_loss: 0.9207 - val_acc: 0.8614\n",
"\n",
"Epoch 00878: val_acc did not improve from 0.86765\n",
"Epoch 879/3000\n",
" - 39s - loss: 0.2819 - acc: 0.9671 - val_loss: 0.9909 - val_acc: 0.8486\n",
"\n",
"Epoch 00879: val_acc did not improve from 0.86765\n",
"Epoch 880/3000\n",
" - 39s - loss: 0.2822 - acc: 0.9663 - val_loss: 0.8859 - val_acc: 0.8622\n",
" - 39s - loss: 0.2437 - acc: 0.9728 - val_loss: 0.9414 - val_acc: 0.8540\n",
"\n",
"Epoch 01453: val_acc did not improve from 0.86765\n",
"Epoch 1454/3000\n",
" - 39s - loss: 0.2456 - acc: 0.9689 - val_loss: 0.9396 - val_acc: 0.8560\n",
"\n",
"Epoch 01454: val_acc did not improve from 0.86765\n",
"Epoch 1455/3000\n",
" - 39s - loss: 0.2501 - acc: 0.9704 - val_loss: 0.9704 - val_acc: 0.8540\n",
"\n",
"Epoch 01455: val_acc did not improve from 0.86765\n",
"Epoch 1456/3000\n",
" - 39s - loss: 0.2410 - acc: 0.9730 - val_loss: 0.9438 - val_acc: 0.8606\n",
"\n",
"Epoch 01456: val_acc did not improve from 0.86765\n",
"Epoch 1457/3000\n",
" - 39s - loss: 0.2434 - acc: 0.9704 - val_loss: 0.9349 - val_acc: 0.8591\n",
"\n",
"Epoch 01457: val_acc did not improve from 0.86765\n",
"Epoch 1458/3000\n",
" - 39s - loss: 0.2376 - acc: 0.9721 - val_loss: 0.9272 - val_acc: 0.8603\n",
"\n",
"Epoch 01458: val_acc did not improve from 0.86765\n",
"Epoch 1459/3000\n",
" - 39s - loss: 0.2536 - acc: 0.9677 - val_loss: 0.9463 - val_acc: 0.8540\n",
"\n",
"Epoch 01459: val_acc did not improve from 0.86765\n",
"Epoch 1460/3000\n",
" - 39s - loss: 0.2437 - acc: 0.9716 - val_loss: 0.9497 - val_acc: 0.8529\n",
"\n",
"Epoch 01460: val_acc did not improve from 0.86765\n",
"Epoch 1461/3000\n",
" - 39s - loss: 0.2477 - acc: 0.9725 - val_loss: 0.9798 - val_acc: 0.8533\n",
"\n",
"Epoch 01461: val_acc did not improve from 0.86765\n",
"Epoch 1462/3000\n",
" - 39s - loss: 0.2415 - acc: 0.9713 - val_loss: 0.9064 - val_acc: 0.8622\n",
"\n",
"Epoch 01462: val_acc did not improve from 0.86765\n",
"Epoch 1463/3000\n",
" - 39s - loss: 0.2492 - acc: 0.9691 - val_loss: 0.9249 - val_acc: 0.8599\n",
"\n",
"Epoch 01463: val_acc did not improve from 0.86765\n",
"Epoch 1464/3000\n",
" - 39s - loss: 0.2496 - acc: 0.9680 - val_loss: 0.9249 - val_acc: 0.8583\n",
"\n",
"Epoch 01464: val_acc did not improve from 0.86765\n",
"Epoch 1465/3000\n",
" - 39s - loss: 0.2397 - acc: 0.9719 - val_loss: 0.9813 - val_acc: 0.8552\n",
"\n",
"Epoch 01465: val_acc did not improve from 0.86765\n",
"Epoch 1466/3000\n",
" - 39s - loss: 0.2389 - acc: 0.9728 - val_loss: 0.9445 - val_acc: 0.8587\n",
"\n",
"Epoch 01466: val_acc did not improve from 0.86765\n",
"Epoch 1467/3000\n",
" - 39s - loss: 0.2439 - acc: 0.9706 - val_loss: 0.9341 - val_acc: 0.8571\n",
"\n",
"Epoch 01467: val_acc did not improve from 0.86765\n",
"Epoch 1468/3000\n",
" - 39s - loss: 0.2416 - acc: 0.9703 - val_loss: 0.9718 - val_acc: 0.8552\n",
"\n",
"Epoch 01468: val_acc did not improve from 0.86765\n",
"Epoch 1469/3000\n",
" - 39s - loss: 0.2418 - acc: 0.9707 - val_loss: 0.9249 - val_acc: 0.8591\n",
"\n",
"Epoch 01469: val_acc did not improve from 0.86765\n",
"Epoch 1470/3000\n",
" - 39s - loss: 0.2429 - acc: 0.9713 - val_loss: 0.9052 - val_acc: 0.8610\n",
"\n",
"Epoch 01470: val_acc did not improve from 0.86765\n",
"Epoch 1471/3000\n",
" - 39s - loss: 0.2460 - acc: 0.9727 - val_loss: 0.9187 - val_acc: 0.8579\n",
"\n",
"Epoch 01471: val_acc did not improve from 0.86765\n",
"Epoch 1472/3000\n",
" - 39s - loss: 0.2476 - acc: 0.9688 - val_loss: 0.9076 - val_acc: 0.8603\n",
"\n",
"Epoch 01472: val_acc did not improve from 0.86765\n",
"Epoch 1473/3000\n",
" - 39s - loss: 0.2420 - acc: 0.9722 - val_loss: 0.9097 - val_acc: 0.8595\n",
"\n",
"Epoch 01473: val_acc did not improve from 0.86765\n",
"Epoch 1474/3000\n",
" - 39s - loss: 0.2489 - acc: 0.9707 - val_loss: 0.9347 - val_acc: 0.8579\n",
"\n",
"Epoch 01474: val_acc did not improve from 0.86765\n",
"Epoch 1475/3000\n",
" - 39s - loss: 0.2462 - acc: 0.9706 - val_loss: 0.9509 - val_acc: 0.8556\n",
"\n",
"Epoch 01475: val_acc did not improve from 0.86765\n",
"Epoch 1476/3000\n",
" - 39s - loss: 0.2499 - acc: 0.9728 - val_loss: 0.9493 - val_acc: 0.8536\n",
"\n",
"Epoch 01476: val_acc did not improve from 0.86765\n",
"Epoch 1477/3000\n",
" - 39s - loss: 0.2415 - acc: 0.9716 - val_loss: 0.9297 - val_acc: 0.8579\n",
"\n",
"Epoch 01477: val_acc did not improve from 0.86765\n",
"Epoch 1478/3000\n",
" - 39s - loss: 0.2439 - acc: 0.9710 - val_loss: 0.9168 - val_acc: 0.8595\n",
"\n",
"Epoch 01481: val_acc did not improve from 0.86765\n",
"Epoch 1482/3000\n",
" - 39s - loss: 0.2395 - acc: 0.9718 - val_loss: 0.9125 - val_acc: 0.8587\n",
"\n",
"Epoch 01482: val_acc did not improve from 0.86765\n",
"Epoch 1483/3000\n",
" - 39s - loss: 0.2418 - acc: 0.9722 - val_loss: 0.9493 - val_acc: 0.8575\n",
"\n",
"Epoch 01483: val_acc did not improve from 0.86765\n",
"Epoch 1484/3000\n",
" - 39s - loss: 0.2469 - acc: 0.9697 - val_loss: 0.9119 - val_acc: 0.8603\n",
"\n",
"Epoch 01484: val_acc did not improve from 0.86765\n",
"Epoch 1485/3000\n",
" - 39s - loss: 0.2441 - acc: 0.9697 - val_loss: 0.9422 - val_acc: 0.8544\n",
"\n",
"Epoch 01485: val_acc did not improve from 0.86765\n",
"Epoch 1486/3000\n",
" - 39s - loss: 0.2403 - acc: 0.9724 - val_loss: 0.8943 - val_acc: 0.8630\n",
"\n",
"Epoch 01486: val_acc did not improve from 0.86765\n",
"Epoch 1487/3000\n",
" - 39s - loss: 0.2391 - acc: 0.9733 - val_loss: 0.9159 - val_acc: 0.8587\n",
"\n",
"Epoch 01487: val_acc did not improve from 0.86765\n",
"Epoch 1488/3000\n",
" - 39s - loss: 0.2531 - acc: 0.9701 - val_loss: 0.9343 - val_acc: 0.8548\n",
"\n",
"Epoch 01488: val_acc did not improve from 0.86765\n",
"Epoch 1489/3000\n",
" - 39s - loss: 0.2523 - acc: 0.9681 - val_loss: 0.9137 - val_acc: 0.8560\n",
"\n",
"Epoch 01489: val_acc did not improve from 0.86765\n",
"Epoch 1490/3000\n",
" - 39s - loss: 0.2489 - acc: 0.9692 - val_loss: 0.9188 - val_acc: 0.8614\n",
"\n",
"Epoch 01490: val_acc did not improve from 0.86765\n",
"Epoch 1491/3000\n",
" - 39s - loss: 0.2460 - acc: 0.9721 - val_loss: 0.9391 - val_acc: 0.8529\n",
"\n",
"Epoch 01491: val_acc did not improve from 0.86765\n",
"Epoch 1492/3000\n",
" - 39s - loss: 0.2413 - acc: 0.9731 - val_loss: 0.9119 - val_acc: 0.8579\n",
"\n",
"Epoch 01492: val_acc did not improve from 0.86765\n",
"Epoch 1493/3000\n",
" - 39s - loss: 0.2432 - acc: 0.9728 - val_loss: 0.9226 - val_acc: 0.8599\n",
"\n",
"Epoch 01493: val_acc did not improve from 0.86765\n",
"Epoch 1494/3000\n",
" - 39s - loss: 0.2494 - acc: 0.9700 - val_loss: 0.9348 - val_acc: 0.8595\n",
"\n",
"Epoch 01494: val_acc did not improve from 0.86765\n",
"Epoch 1495/3000\n",
" - 39s - loss: 0.2516 - acc: 0.9691 - val_loss: 0.9478 - val_acc: 0.8529\n",
"\n",
"Epoch 01495: val_acc did not improve from 0.86765\n",
"Epoch 1496/3000\n",
" - 39s - loss: 0.2404 - acc: 0.9730 - val_loss: 0.9172 - val_acc: 0.8575\n",
"\n",
"Epoch 01496: val_acc did not improve from 0.86765\n",
"Epoch 1497/3000\n",
" - 39s - loss: 0.2487 - acc: 0.9700 - val_loss: 0.9282 - val_acc: 0.8595\n",
"\n",
"Epoch 01497: val_acc did not improve from 0.86765\n",
"Epoch 1498/3000\n",
" - 39s - loss: 0.2415 - acc: 0.9703 - val_loss: 0.9244 - val_acc: 0.8583\n",
"\n",
"Epoch 01498: val_acc did not improve from 0.86765\n",
"Epoch 1499/3000\n",
" - 39s - loss: 0.2360 - acc: 0.9731 - val_loss: 0.9419 - val_acc: 0.8533\n",
"\n",
"Epoch 01499: val_acc did not improve from 0.86765\n",
"Epoch 1500/3000\n",
" - 40s - loss: 0.2498 - acc: 0.9707 - val_loss: 0.9439 - val_acc: 0.8571\n",
"\n",
"Epoch 01500: val_acc did not improve from 0.86765\n",
"Epoch 1501/3000\n",
" - 39s - loss: 0.2382 - acc: 0.9739 - val_loss: 0.9383 - val_acc: 0.8603\n",
"\n",
"Epoch 01501: val_acc did not improve from 0.86765\n",
"Epoch 1502/3000\n",
" - 39s - loss: 0.2510 - acc: 0.9691 - val_loss: 0.9453 - val_acc: 0.8540\n",
"\n",
"Epoch 01502: val_acc did not improve from 0.86765\n",
"Epoch 1503/3000\n",
" - 39s - loss: 0.2407 - acc: 0.9715 - val_loss: 0.9334 - val_acc: 0.8556\n",
"\n",
"Epoch 01503: val_acc did not improve from 0.86765\n",
"Epoch 1504/3000\n",
" - 39s - loss: 0.2410 - acc: 0.9719 - val_loss: 0.9173 - val_acc: 0.8591\n",
"\n",
"Epoch 01504: val_acc did not improve from 0.86765\n",
"Epoch 1505/3000\n",
" - 39s - loss: 0.2400 - acc: 0.9727 - val_loss: 0.9308 - val_acc: 0.8575\n",
"\n",
"Epoch 01505: val_acc did not improve from 0.86765\n",
"Epoch 1506/3000\n",
" - 39s - loss: 0.2380 - acc: 0.9716 - val_loss: 0.9309 - val_acc: 0.8583\n",
"\n",
"Epoch 01506: val_acc did not improve from 0.86765\n",
"Epoch 1507/3000\n",
" - 39s - loss: 0.2403 - acc: 0.9727 - val_loss: 0.9427 - val_acc: 0.8583\n",
"\n",
"Epoch 01507: val_acc did not improve from 0.86765\n",
"Epoch 1508/3000\n",
" - 39s - loss: 0.2587 - acc: 0.9684 - val_loss: 0.9172 - val_acc: 0.8571\n",
"\n",
"Epoch 01508: val_acc did not improve from 0.86765\n",
"Epoch 1509/3000\n",
" - 39s - loss: 0.2459 - acc: 0.9706 - val_loss: 0.9064 - val_acc: 0.8583\n",
"\n",
"Epoch 01509: val_acc did not improve from 0.86765\n",
"Epoch 1510/3000\n",
" - 39s - loss: 0.2416 - acc: 0.9718 - val_loss: 0.9717 - val_acc: 0.8517\n",
"\n",
"Epoch 01510: val_acc did not improve from 0.86765\n",
"Epoch 1511/3000\n",
" - 39s - loss: 0.2383 - acc: 0.9748 - val_loss: 0.9140 - val_acc: 0.8610\n",
"\n",
"Epoch 01511: val_acc did not improve from 0.86765\n",
"Epoch 1512/3000\n",
" - 40s - loss: 0.2466 - acc: 0.9703 - val_loss: 0.9460 - val_acc: 0.8571\n",
"\n",
"Epoch 01512: val_acc did not improve from 0.86765\n",
"Epoch 1513/3000\n",
" - 39s - loss: 0.2477 - acc: 0.9706 - val_loss: 0.9360 - val_acc: 0.8560\n",
"\n",
"Epoch 01513: val_acc did not improve from 0.86765\n",
"Epoch 1514/3000\n",
" - 39s - loss: 0.2462 - acc: 0.9703 - val_loss: 0.9291 - val_acc: 0.8560\n",
"\n",
"Epoch 01514: val_acc did not improve from 0.86765\n",
"Epoch 1515/3000\n",
" - 39s - loss: 0.2393 - acc: 0.9725 - val_loss: 0.9156 - val_acc: 0.8575\n",
"\n",
"Epoch 01515: val_acc did not improve from 0.86765\n",
"Epoch 1516/3000\n",
" - 39s - loss: 0.2503 - acc: 0.9700 - val_loss: 0.9251 - val_acc: 0.8548\n",
"\n",
"Epoch 01516: val_acc did not improve from 0.86765\n",
"Epoch 1517/3000\n",
" - 39s - loss: 0.2373 - acc: 0.9706 - val_loss: 0.9421 - val_acc: 0.8575\n",
"\n",
"Epoch 01517: val_acc did not improve from 0.86765\n",
"Epoch 1518/3000\n",
" - 39s - loss: 0.2344 - acc: 0.9736 - val_loss: 0.9315 - val_acc: 0.8560\n",
"\n",
"Epoch 01518: val_acc did not improve from 0.86765\n",
"Epoch 1519/3000\n",
" - 39s - loss: 0.2454 - acc: 0.9692 - val_loss: 0.9511 - val_acc: 0.8544\n",
"\n",
"Epoch 01519: val_acc did not improve from 0.86765\n",
"Epoch 1520/3000\n",
" - 39s - loss: 0.2413 - acc: 0.9713 - val_loss: 0.9209 - val_acc: 0.8583\n",
"\n",
"Epoch 01520: val_acc did not improve from 0.86765\n",
"Epoch 1521/3000\n",
" - 39s - loss: 0.2384 - acc: 0.9725 - val_loss: 0.9393 - val_acc: 0.8583\n",
"\n",
"Epoch 01521: val_acc did not improve from 0.86765\n",
"Epoch 1522/3000\n",
" - 39s - loss: 0.2409 - acc: 0.9718 - val_loss: 0.9752 - val_acc: 0.8536\n",
"\n",
"Epoch 01522: val_acc did not improve from 0.86765\n",
"Epoch 1523/3000\n",
" - 39s - loss: 0.2481 - acc: 0.9712 - val_loss: 0.9664 - val_acc: 0.8552\n",
"\n",
"Epoch 01523: val_acc did not improve from 0.86765\n",
"Epoch 1524/3000\n",
" - 39s - loss: 0.2389 - acc: 0.9718 - val_loss: 0.9558 - val_acc: 0.8606\n",
"\n",
"Epoch 01524: val_acc did not improve from 0.86765\n",
"Epoch 1525/3000\n",
" - 40s - loss: 0.2428 - acc: 0.9707 - val_loss: 0.9175 - val_acc: 0.8579\n",
"\n",
"Epoch 01525: val_acc did not improve from 0.86765\n",
"Epoch 1526/3000\n",
" - 39s - loss: 0.2370 - acc: 0.9710 - val_loss: 0.9364 - val_acc: 0.8568\n",
"\n",
"Epoch 01526: val_acc did not improve from 0.86765\n",
"Epoch 1527/3000\n",
" - 39s - loss: 0.2385 - acc: 0.9724 - val_loss: 0.9245 - val_acc: 0.8591\n",
"\n",
"Epoch 01527: val_acc did not improve from 0.86765\n",
"Epoch 1528/3000\n",
" - 39s - loss: 0.2444 - acc: 0.9713 - val_loss: 0.9479 - val_acc: 0.8552\n",
"\n",
"Epoch 01528: val_acc did not improve from 0.86765\n",
"Epoch 1529/3000\n",
" - 39s - loss: 0.2440 - acc: 0.9712 - val_loss: 0.9174 - val_acc: 0.8587\n",
"\n",
"Epoch 01529: val_acc did not improve from 0.86765\n",
"Epoch 1530/3000\n",
" - 39s - loss: 0.2414 - acc: 0.9730 - val_loss: 0.9120 - val_acc: 0.8610\n",
"\n",
"Epoch 01530: val_acc did not improve from 0.86765\n",
"Epoch 1531/3000\n",
" - 39s - loss: 0.2462 - acc: 0.9700 - val_loss: 0.9599 - val_acc: 0.8533\n",
"\n",
"Epoch 01531: val_acc did not improve from 0.86765\n",
"Epoch 1532/3000\n",
" - 39s - loss: 0.2468 - acc: 0.9703 - val_loss: 0.9481 - val_acc: 0.8575\n",
"\n",
"Epoch 01532: val_acc did not improve from 0.86765\n",
"Epoch 1533/3000\n",
" - 39s - loss: 0.2527 - acc: 0.9697 - val_loss: 0.9536 - val_acc: 0.8552\n",
"\n",
"Epoch 01533: val_acc did not improve from 0.86765\n",
"Epoch 1534/3000\n",
" - 39s - loss: 0.2402 - acc: 0.9725 - val_loss: 0.9329 - val_acc: 0.8579\n",
"\n",
"Epoch 01534: val_acc did not improve from 0.86765\n",
"Epoch 1535/3000\n",
" - 39s - loss: 0.2466 - acc: 0.9716 - val_loss: 0.9295 - val_acc: 0.8595\n",
"\n",
"Epoch 01535: val_acc did not improve from 0.86765\n",
"Epoch 1536/3000\n",
" - 39s - loss: 0.2510 - acc: 0.9689 - val_loss: 0.9513 - val_acc: 0.8575\n",
"\n",
"Epoch 01536: val_acc did not improve from 0.86765\n",
"Epoch 1537/3000\n",
" - 39s - loss: 0.2406 - acc: 0.9734 - val_loss: 0.9543 - val_acc: 0.8556\n",
"\n",
"Epoch 01537: val_acc did not improve from 0.86765\n",
"Epoch 1538/3000\n",
" - 39s - loss: 0.2365 - acc: 0.9742 - val_loss: 0.9489 - val_acc: 0.8568\n",
"\n",
"Epoch 01538: val_acc did not improve from 0.86765\n",
"Epoch 1539/3000\n",
" - 39s - loss: 0.2477 - acc: 0.9716 - val_loss: 0.9163 - val_acc: 0.8618\n",
"\n",
"Epoch 01539: val_acc did not improve from 0.86765\n",
"Epoch 1540/3000\n",
" - 39s - loss: 0.2436 - acc: 0.9710 - val_loss: 0.9541 - val_acc: 0.8579\n",
"\n",
"Epoch 01540: val_acc did not improve from 0.86765\n",
"Epoch 1541/3000\n",
" - 39s - loss: 0.2495 - acc: 0.9704 - val_loss: 0.9125 - val_acc: 0.8603\n",
"\n",
"Epoch 01541: val_acc did not improve from 0.86765\n",
"Epoch 1542/3000\n",
" - 39s - loss: 0.2509 - acc: 0.9691 - val_loss: 0.9325 - val_acc: 0.8579\n",
"\n",
"Epoch 01542: val_acc did not improve from 0.86765\n",
"Epoch 1543/3000\n",
" - 39s - loss: 0.2377 - acc: 0.9725 - val_loss: 0.9237 - val_acc: 0.8622\n",
"\n",
"Epoch 01543: val_acc did not improve from 0.86765\n",
"Epoch 1544/3000\n",
" - 39s - loss: 0.2403 - acc: 0.9700 - val_loss: 0.9509 - val_acc: 0.8560\n",
"\n",
"Epoch 01544: val_acc did not improve from 0.86765\n",
"Epoch 1545/3000\n",
" - 40s - loss: 0.2449 - acc: 0.9697 - val_loss: 0.9444 - val_acc: 0.8618\n",
"\n",
"Epoch 01545: val_acc did not improve from 0.86765\n",
"Epoch 1546/3000\n",
" - 39s - loss: 0.2476 - acc: 0.9712 - val_loss: 0.9380 - val_acc: 0.8595\n",
"\n",
"Epoch 01546: val_acc did not improve from 0.86765\n",
"Epoch 1547/3000\n",
" - 39s - loss: 0.2425 - acc: 0.9712 - val_loss: 0.9158 - val_acc: 0.8575\n",
"\n",
"Epoch 01547: val_acc did not improve from 0.86765\n",
"Epoch 1548/3000\n",
" - 39s - loss: 0.2398 - acc: 0.9718 - val_loss: 0.9345 - val_acc: 0.8583\n",
"\n",
"Epoch 01548: val_acc did not improve from 0.86765\n",
"Epoch 1549/3000\n",
" - 39s - loss: 0.2443 - acc: 0.9707 - val_loss: 0.9413 - val_acc: 0.8529\n",
"\n",
"Epoch 01549: val_acc did not improve from 0.86765\n",
"Epoch 1550/3000\n",
" - 39s - loss: 0.2405 - acc: 0.9718 - val_loss: 0.9338 - val_acc: 0.8548\n",
"\n",
"Epoch 01550: val_acc did not improve from 0.86765\n",
"Epoch 1551/3000\n",
" - 39s - loss: 0.2521 - acc: 0.9701 - val_loss: 0.9322 - val_acc: 0.8587\n",
"\n",
"Epoch 01551: val_acc did not improve from 0.86765\n",
"Epoch 1552/3000\n",
" - 40s - loss: 0.2453 - acc: 0.9728 - val_loss: 0.9305 - val_acc: 0.8568\n",
"\n",
"Epoch 01552: val_acc did not improve from 0.86765\n",
"Epoch 1553/3000\n",
" - 39s - loss: 0.2423 - acc: 0.9719 - val_loss: 0.9418 - val_acc: 0.8579\n",
"\n",
"Epoch 01553: val_acc did not improve from 0.86765\n",
"Epoch 1554/3000\n",
" - 39s - loss: 0.2467 - acc: 0.9710 - val_loss: 0.9277 - val_acc: 0.8587\n",
"\n",
"Epoch 01554: val_acc did not improve from 0.86765\n",
"Epoch 1555/3000\n",
" - 39s - loss: 0.2439 - acc: 0.9706 - val_loss: 0.9063 - val_acc: 0.8591\n",
"\n",
"Epoch 01555: val_acc did not improve from 0.86765\n",
"Epoch 1556/3000\n",
" - 39s - loss: 0.2427 - acc: 0.9721 - val_loss: 0.9451 - val_acc: 0.8540\n",
"\n",
"Epoch 01556: val_acc did not improve from 0.86765\n",
"Epoch 1557/3000\n",
" - 39s - loss: 0.2500 - acc: 0.9707 - val_loss: 0.9079 - val_acc: 0.8599\n",
"\n",
"Epoch 01557: val_acc did not improve from 0.86765\n",
"Epoch 1558/3000\n",
" - 39s - loss: 0.2409 - acc: 0.9709 - val_loss: 0.9328 - val_acc: 0.8552\n",
"\n",
"Epoch 01558: val_acc did not improve from 0.86765\n",
"Epoch 1559/3000\n",
" - 39s - loss: 0.2429 - acc: 0.9725 - val_loss: 0.9575 - val_acc: 0.8564\n",
"\n",
"Epoch 01559: val_acc did not improve from 0.86765\n",
"Epoch 1560/3000\n",
" - 40s - loss: 0.2338 - acc: 0.9734 - val_loss: 0.9515 - val_acc: 0.8568\n",
"\n",
"Epoch 01560: val_acc did not improve from 0.86765\n",
"Epoch 1561/3000\n",
" - 40s - loss: 0.2473 - acc: 0.9706 - val_loss: 0.9409 - val_acc: 0.8591\n",
"\n",
"Epoch 01561: val_acc did not improve from 0.86765\n",
"Epoch 1562/3000\n",
" - 39s - loss: 0.2456 - acc: 0.9712 - val_loss: 0.9535 - val_acc: 0.8560\n",
"\n",
"Epoch 01562: val_acc did not improve from 0.86765\n",
"Epoch 1563/3000\n",
" - 39s - loss: 0.2408 - acc: 0.9716 - val_loss: 0.9553 - val_acc: 0.8540\n",
"\n",
"Epoch 01563: val_acc did not improve from 0.86765\n",
"Epoch 1564/3000\n",
" - 39s - loss: 0.2381 - acc: 0.9746 - val_loss: 0.9454 - val_acc: 0.8579\n",
"\n",
"Epoch 01564: val_acc did not improve from 0.86765\n",
"Epoch 1565/3000\n",
" - 39s - loss: 0.2401 - acc: 0.9736 - val_loss: 0.9554 - val_acc: 0.8556\n",
"\n",
"Epoch 01565: val_acc did not improve from 0.86765\n",
"Epoch 1566/3000\n",
" - 39s - loss: 0.2458 - acc: 0.9695 - val_loss: 0.9125 - val_acc: 0.8591\n",
"\n",
"Epoch 01566: val_acc did not improve from 0.86765\n",
"Epoch 1567/3000\n",
" - 40s - loss: 0.2468 - acc: 0.9704 - val_loss: 0.9234 - val_acc: 0.8595\n",
"\n",
"Epoch 01567: val_acc did not improve from 0.86765\n",
"Epoch 1568/3000\n",
" - 39s - loss: 0.2473 - acc: 0.9686 - val_loss: 0.9318 - val_acc: 0.8556\n",
"\n",
"Epoch 01568: val_acc did not improve from 0.86765\n",
"Epoch 1569/3000\n",
" - 39s - loss: 0.2450 - acc: 0.9701 - val_loss: 0.9276 - val_acc: 0.8591\n",
"\n",
"Epoch 01569: val_acc did not improve from 0.86765\n",
"Epoch 1570/3000\n",
" - 39s - loss: 0.2446 - acc: 0.9698 - val_loss: 0.9424 - val_acc: 0.8548\n",
"\n",
"Epoch 01570: val_acc did not improve from 0.86765\n",
"Epoch 1571/3000\n",
" - 39s - loss: 0.2531 - acc: 0.9695 - val_loss: 0.9584 - val_acc: 0.8548\n",
"\n",
"Epoch 01571: val_acc did not improve from 0.86765\n",
"Epoch 1572/3000\n",
" - 39s - loss: 0.2353 - acc: 0.9734 - val_loss: 0.9187 - val_acc: 0.8618\n",
"\n",
"Epoch 01572: val_acc did not improve from 0.86765\n",
"Epoch 1573/3000\n",
" - 39s - loss: 0.2381 - acc: 0.9721 - val_loss: 0.9378 - val_acc: 0.8548\n",
"\n",
"Epoch 01573: val_acc did not improve from 0.86765\n",
"Epoch 1574/3000\n",
" - 39s - loss: 0.2368 - acc: 0.9734 - val_loss: 0.9305 - val_acc: 0.8560\n",
"\n",
"Epoch 01574: val_acc did not improve from 0.86765\n",
"Epoch 1575/3000\n",
" - 39s - loss: 0.2427 - acc: 0.9706 - val_loss: 0.9283 - val_acc: 0.8560\n",
"\n",
"Epoch 01575: val_acc did not improve from 0.86765\n",
"Epoch 1576/3000\n",
" - 39s - loss: 0.2452 - acc: 0.9713 - val_loss: 0.9388 - val_acc: 0.8568\n",
"\n",
"Epoch 01576: val_acc did not improve from 0.86765\n",
"Epoch 1577/3000\n",
" - 39s - loss: 0.2353 - acc: 0.9731 - val_loss: 0.9383 - val_acc: 0.8583\n",
"\n",
"Epoch 01577: val_acc did not improve from 0.86765\n",
"Epoch 1578/3000\n",
" - 39s - loss: 0.2318 - acc: 0.9737 - val_loss: 0.9308 - val_acc: 0.8556\n",
"\n",
"Epoch 01578: val_acc did not improve from 0.86765\n",
"Epoch 1579/3000\n",
" - 39s - loss: 0.2462 - acc: 0.9737 - val_loss: 0.9294 - val_acc: 0.8540\n",
"\n",
"Epoch 01579: val_acc did not improve from 0.86765\n",
"Epoch 1580/3000\n",
" - 39s - loss: 0.2419 - acc: 0.9722 - val_loss: 0.9507 - val_acc: 0.8525\n",
"\n",
"Epoch 01580: val_acc did not improve from 0.86765\n",
"Epoch 1581/3000\n",
" - 39s - loss: 0.2438 - acc: 0.9727 - val_loss: 0.9298 - val_acc: 0.8568\n",
"\n",
"Epoch 01581: val_acc did not improve from 0.86765\n",
"Epoch 1582/3000\n",
" - 39s - loss: 0.2423 - acc: 0.9727 - val_loss: 0.9501 - val_acc: 0.8556\n",
"\n",
"Epoch 01582: val_acc did not improve from 0.86765\n",
"Epoch 1583/3000\n",
" - 39s - loss: 0.2363 - acc: 0.9727 - val_loss: 0.9519 - val_acc: 0.8552\n",
"\n",
"Epoch 01583: val_acc did not improve from 0.86765\n",
"Epoch 1584/3000\n",
" - 39s - loss: 0.2477 - acc: 0.9686 - val_loss: 0.9393 - val_acc: 0.8571\n",
"\n",
"Epoch 01584: val_acc did not improve from 0.86765\n",
"Epoch 1585/3000\n",
" - 39s - loss: 0.2445 - acc: 0.9706 - val_loss: 0.9423 - val_acc: 0.8536\n",
"\n",
"Epoch 01585: val_acc did not improve from 0.86765\n",
"Epoch 1586/3000\n",
" - 39s - loss: 0.2391 - acc: 0.9722 - val_loss: 0.9277 - val_acc: 0.8564\n",
"\n",
"Epoch 01586: val_acc did not improve from 0.86765\n",
"Epoch 1587/3000\n",
" - 39s - loss: 0.2368 - acc: 0.9736 - val_loss: 0.9419 - val_acc: 0.8568\n",
"\n",
"Epoch 01587: val_acc did not improve from 0.86765\n",
"Epoch 1588/3000\n",
" - 39s - loss: 0.2463 - acc: 0.9703 - val_loss: 0.9391 - val_acc: 0.8579\n",
"\n",
"Epoch 01588: val_acc did not improve from 0.86765\n",
"Epoch 1589/3000\n",
" - 39s - loss: 0.2509 - acc: 0.9715 - val_loss: 0.8978 - val_acc: 0.8599\n",
"\n",
"Epoch 01589: val_acc did not improve from 0.86765\n",
"Epoch 1590/3000\n",
" - 39s - loss: 0.2401 - acc: 0.9736 - val_loss: 0.9471 - val_acc: 0.8552\n",
"\n",
"Epoch 01590: val_acc did not improve from 0.86765\n",
"Epoch 1591/3000\n",
" - 39s - loss: 0.2393 - acc: 0.9745 - val_loss: 0.9482 - val_acc: 0.8560\n",
"\n",
"Epoch 01591: val_acc did not improve from 0.86765\n",
"Epoch 1592/3000\n",
" - 39s - loss: 0.2456 - acc: 0.9706 - val_loss: 0.9385 - val_acc: 0.8583\n",
"\n",
"Epoch 01592: val_acc did not improve from 0.86765\n",
"Epoch 1593/3000\n",
" - 39s - loss: 0.2496 - acc: 0.9698 - val_loss: 0.9354 - val_acc: 0.8568\n",
"\n",
"Epoch 01593: val_acc did not improve from 0.86765\n",
"Epoch 1594/3000\n",
" - 39s - loss: 0.2427 - acc: 0.9722 - val_loss: 0.9388 - val_acc: 0.8556\n",
"\n",
"Epoch 01594: val_acc did not improve from 0.86765\n",
"Epoch 1595/3000\n",
" - 39s - loss: 0.2441 - acc: 0.9722 - val_loss: 0.9616 - val_acc: 0.8556\n",
"\n",
"Epoch 01595: val_acc did not improve from 0.86765\n",
"Epoch 1596/3000\n",
" - 39s - loss: 0.2371 - acc: 0.9736 - val_loss: 0.9283 - val_acc: 0.8610\n",
"\n",
"Epoch 01596: val_acc did not improve from 0.86765\n",
"Epoch 1597/3000\n",
" - 39s - loss: 0.2392 - acc: 0.9718 - val_loss: 0.9350 - val_acc: 0.8568\n",
"\n",
"Epoch 01597: val_acc did not improve from 0.86765\n",
"Epoch 1598/3000\n",
" - 39s - loss: 0.2397 - acc: 0.9730 - val_loss: 0.9188 - val_acc: 0.8587\n",
"\n",
"Epoch 01598: val_acc did not improve from 0.86765\n",
"Epoch 1599/3000\n",
" - 39s - loss: 0.2416 - acc: 0.9728 - val_loss: 0.9390 - val_acc: 0.8568\n",
"\n",
"Epoch 01599: val_acc did not improve from 0.86765\n",
"Epoch 1600/3000\n",
" - 39s - loss: 0.2415 - acc: 0.9719 - val_loss: 0.9558 - val_acc: 0.8529\n",
"\n",
"Epoch 01600: val_acc did not improve from 0.86765\n",
"Epoch 1601/3000\n",
" - 40s - loss: 0.2359 - acc: 0.9739 - val_loss: 0.9396 - val_acc: 0.8552\n",
"\n",
"Epoch 01601: val_acc did not improve from 0.86765\n",
"Epoch 1602/3000\n",
" - 39s - loss: 0.2448 - acc: 0.9709 - val_loss: 0.9179 - val_acc: 0.8595\n",
"\n",
"Epoch 01602: val_acc did not improve from 0.86765\n",
"Epoch 1603/3000\n",
" - 39s - loss: 0.2373 - acc: 0.9725 - val_loss: 0.9440 - val_acc: 0.8575\n",
"\n",
"Epoch 01603: val_acc did not improve from 0.86765\n",
"Epoch 1604/3000\n",
" - 39s - loss: 0.2330 - acc: 0.9755 - val_loss: 0.9561 - val_acc: 0.8579\n",
"\n",
"Epoch 01604: val_acc did not improve from 0.86765\n",
"Epoch 1605/3000\n",
" - 39s - loss: 0.2442 - acc: 0.9709 - val_loss: 0.9349 - val_acc: 0.8568\n",
"\n",
"Epoch 01605: val_acc did not improve from 0.86765\n",
"Epoch 1606/3000\n",
" - 39s - loss: 0.2401 - acc: 0.9718 - val_loss: 0.9517 - val_acc: 0.8552\n",
"\n",
"Epoch 01606: val_acc did not improve from 0.86765\n",
"Epoch 1607/3000\n",
" - 39s - loss: 0.2407 - acc: 0.9700 - val_loss: 0.9624 - val_acc: 0.8533\n",
"\n",
"Epoch 01607: val_acc did not improve from 0.86765\n",
"Epoch 1608/3000\n",
" - 39s - loss: 0.2463 - acc: 0.9722 - val_loss: 0.9382 - val_acc: 0.8575\n",
"\n",
"Epoch 01608: val_acc did not improve from 0.86765\n",
"Epoch 1609/3000\n",
" - 40s - loss: 0.2430 - acc: 0.9721 - val_loss: 0.9283 - val_acc: 0.8579\n",
"\n",
"Epoch 01609: val_acc did not improve from 0.86765\n",
"Epoch 1610/3000\n",
" - 39s - loss: 0.2407 - acc: 0.9716 - val_loss: 0.9335 - val_acc: 0.8552\n",
"\n",
"Epoch 01610: val_acc did not improve from 0.86765\n",
"Epoch 1611/3000\n",
" - 39s - loss: 0.2451 - acc: 0.9707 - val_loss: 0.9679 - val_acc: 0.8529\n",
"\n",
"Epoch 01611: val_acc did not improve from 0.86765\n",
"Epoch 1612/3000\n",
" - 39s - loss: 0.2400 - acc: 0.9719 - val_loss: 0.9498 - val_acc: 0.8540\n",
"\n",
"Epoch 01612: val_acc did not improve from 0.86765\n",
"Epoch 1613/3000\n",
" - 39s - loss: 0.2493 - acc: 0.9698 - val_loss: 0.9202 - val_acc: 0.8556\n",
"\n",
"Epoch 01613: val_acc did not improve from 0.86765\n",
"Epoch 1614/3000\n",
" - 39s - loss: 0.2435 - acc: 0.9719 - val_loss: 0.9701 - val_acc: 0.8513\n",
"\n",
"Epoch 01614: val_acc did not improve from 0.86765\n",
"Epoch 1615/3000\n",
" - 39s - loss: 0.2496 - acc: 0.9704 - val_loss: 0.9406 - val_acc: 0.8494\n",
"\n",
"Epoch 01615: val_acc did not improve from 0.86765\n",
"Epoch 1616/3000\n",
" - 39s - loss: 0.2410 - acc: 0.9727 - val_loss: 0.9507 - val_acc: 0.8544\n",
"\n",
"Epoch 01616: val_acc did not improve from 0.86765\n",
"Epoch 1617/3000\n",
" - 39s - loss: 0.2406 - acc: 0.9725 - val_loss: 0.9349 - val_acc: 0.8599\n",
"\n",
"Epoch 01617: val_acc did not improve from 0.86765\n",
"Epoch 1618/3000\n",
" - 39s - loss: 0.2343 - acc: 0.9740 - val_loss: 0.9635 - val_acc: 0.8556\n",
"\n",
"Epoch 01618: val_acc did not improve from 0.86765\n",
"Epoch 1619/3000\n",
" - 39s - loss: 0.2390 - acc: 0.9727 - val_loss: 0.9410 - val_acc: 0.8583\n",
"\n",
"Epoch 01619: val_acc did not improve from 0.86765\n",
"Epoch 1620/3000\n",
" - 39s - loss: 0.2365 - acc: 0.9740 - val_loss: 0.9448 - val_acc: 0.8571\n",
"\n",
"Epoch 01620: val_acc did not improve from 0.86765\n",
"Epoch 1621/3000\n",
" - 39s - loss: 0.2352 - acc: 0.9736 - val_loss: 0.9504 - val_acc: 0.8564\n",
"\n",
"Epoch 01621: val_acc did not improve from 0.86765\n",
"Epoch 1622/3000\n",
" - 39s - loss: 0.2404 - acc: 0.9706 - val_loss: 0.9619 - val_acc: 0.8591\n",
"\n",
"Epoch 01622: val_acc did not improve from 0.86765\n",
"Epoch 1623/3000\n",
" - 39s - loss: 0.2453 - acc: 0.9725 - val_loss: 0.9266 - val_acc: 0.8591\n",
"\n",
"Epoch 01623: val_acc did not improve from 0.86765\n",
"Epoch 1624/3000\n",
" - 39s - loss: 0.2396 - acc: 0.9716 - val_loss: 0.9552 - val_acc: 0.8571\n",
"\n",
"Epoch 01624: val_acc did not improve from 0.86765\n",
"Epoch 1625/3000\n",
" - 39s - loss: 0.2455 - acc: 0.9707 - val_loss: 0.9090 - val_acc: 0.8595\n",
"\n",
"Epoch 01625: val_acc did not improve from 0.86765\n",
"Epoch 1626/3000\n",
" - 39s - loss: 0.2400 - acc: 0.9727 - val_loss: 0.9586 - val_acc: 0.8575\n",
"\n",
"Epoch 01626: val_acc did not improve from 0.86765\n",
"Epoch 1627/3000\n",
" - 39s - loss: 0.2413 - acc: 0.9722 - val_loss: 0.9194 - val_acc: 0.8630\n",
"\n",
"Epoch 01627: val_acc did not improve from 0.86765\n",
"Epoch 1628/3000\n",
" - 39s - loss: 0.2453 - acc: 0.9701 - val_loss: 0.9331 - val_acc: 0.8564\n",
"\n",
"Epoch 01628: val_acc did not improve from 0.86765\n",
"Epoch 1629/3000\n",
" - 39s - loss: 0.2332 - acc: 0.9725 - val_loss: 0.9728 - val_acc: 0.8556\n",
"\n",
"Epoch 01629: val_acc did not improve from 0.86765\n",
"Epoch 1630/3000\n",
" - 39s - loss: 0.2458 - acc: 0.9718 - val_loss: 0.9460 - val_acc: 0.8575\n",
"\n",
"Epoch 01630: val_acc did not improve from 0.86765\n",
"Epoch 1631/3000\n",
" - 39s - loss: 0.2544 - acc: 0.9669 - val_loss: 0.9337 - val_acc: 0.8556\n",
"\n",
"Epoch 01631: val_acc did not improve from 0.86765\n",
"Epoch 1632/3000\n",
" - 39s - loss: 0.2498 - acc: 0.9698 - val_loss: 0.9509 - val_acc: 0.8552\n",
"\n",
"Epoch 01632: val_acc did not improve from 0.86765\n",
"Epoch 1633/3000\n",
" - 39s - loss: 0.2383 - acc: 0.9719 - val_loss: 0.9533 - val_acc: 0.8564\n",
"\n",
"Epoch 01633: val_acc did not improve from 0.86765\n",
"Epoch 1634/3000\n",
" - 39s - loss: 0.2489 - acc: 0.9704 - val_loss: 0.9220 - val_acc: 0.8595\n",
"\n",
"Epoch 01634: val_acc did not improve from 0.86765\n",
"Epoch 1635/3000\n",
" - 39s - loss: 0.2393 - acc: 0.9716 - val_loss: 0.9476 - val_acc: 0.8579\n",
"\n",
"Epoch 01635: val_acc did not improve from 0.86765\n",
"Epoch 1636/3000\n",
" - 39s - loss: 0.2425 - acc: 0.9716 - val_loss: 0.9419 - val_acc: 0.8560\n",
"\n",
"Epoch 01636: val_acc did not improve from 0.86765\n",
"Epoch 1637/3000\n",
" - 39s - loss: 0.2475 - acc: 0.9680 - val_loss: 0.9260 - val_acc: 0.8575\n",
"\n",
"Epoch 01637: val_acc did not improve from 0.86765\n",
"Epoch 1638/3000\n",
" - 39s - loss: 0.2448 - acc: 0.9707 - val_loss: 0.9294 - val_acc: 0.8579\n",
"\n",
"Epoch 01638: val_acc did not improve from 0.86765\n",
"Epoch 1639/3000\n",
" - 39s - loss: 0.2376 - acc: 0.9710 - val_loss: 0.9369 - val_acc: 0.8568\n",
"\n",
"Epoch 01639: val_acc did not improve from 0.86765\n",
"Epoch 1640/3000\n",
" - 39s - loss: 0.2451 - acc: 0.9688 - val_loss: 0.9558 - val_acc: 0.8571\n",
"\n",
"Epoch 01640: val_acc did not improve from 0.86765\n",
"Epoch 1641/3000\n",
" - 39s - loss: 0.2404 - acc: 0.9709 - val_loss: 0.9431 - val_acc: 0.8587\n",
"\n",
"Epoch 01641: val_acc did not improve from 0.86765\n",
"Epoch 1642/3000\n",
" - 40s - loss: 0.2439 - acc: 0.9683 - val_loss: 0.9447 - val_acc: 0.8552\n",
"\n",
"Epoch 01642: val_acc did not improve from 0.86765\n",
"Epoch 1643/3000\n",
" - 39s - loss: 0.2433 - acc: 0.9707 - val_loss: 0.9525 - val_acc: 0.8536\n",
"\n",
"Epoch 01643: val_acc did not improve from 0.86765\n",
"Epoch 1644/3000\n",
" - 39s - loss: 0.2329 - acc: 0.9757 - val_loss: 0.9513 - val_acc: 0.8595\n",
"\n",
"Epoch 01644: val_acc did not improve from 0.86765\n",
"Epoch 1645/3000\n",
" - 39s - loss: 0.2347 - acc: 0.9713 - val_loss: 0.9371 - val_acc: 0.8606\n",
"\n",
"Epoch 01645: val_acc did not improve from 0.86765\n",
"Epoch 1646/3000\n",
" - 39s - loss: 0.2341 - acc: 0.9757 - val_loss: 0.9519 - val_acc: 0.8521\n",
"\n",
"Epoch 01646: val_acc did not improve from 0.86765\n",
"Epoch 1647/3000\n",
" - 39s - loss: 0.2425 - acc: 0.9692 - val_loss: 0.9479 - val_acc: 0.8560\n",
"\n",
"Epoch 01647: val_acc did not improve from 0.86765\n",
"Epoch 1648/3000\n",
" - 39s - loss: 0.2429 - acc: 0.9718 - val_loss: 0.9273 - val_acc: 0.8599\n",
"\n",
"Epoch 01648: val_acc did not improve from 0.86765\n",
"Epoch 1649/3000\n",
" - 39s - loss: 0.2422 - acc: 0.9710 - val_loss: 0.9667 - val_acc: 0.8560\n",
"\n",
"Epoch 01649: val_acc did not improve from 0.86765\n",
"Epoch 1650/3000\n",
" - 39s - loss: 0.2579 - acc: 0.9684 - val_loss: 0.9098 - val_acc: 0.8599\n",
"\n",
"Epoch 01650: val_acc did not improve from 0.86765\n",
"Epoch 1651/3000\n",
" - 39s - loss: 0.2449 - acc: 0.9703 - val_loss: 0.9447 - val_acc: 0.8556\n",
"\n",
"Epoch 01651: val_acc did not improve from 0.86765\n",
"Epoch 1652/3000\n",
" - 39s - loss: 0.2403 - acc: 0.9725 - val_loss: 0.9293 - val_acc: 0.8575\n",
"\n",
"Epoch 01652: val_acc did not improve from 0.86765\n",
"Epoch 1653/3000\n",
" - 39s - loss: 0.2319 - acc: 0.9751 - val_loss: 0.9602 - val_acc: 0.8540\n",
"\n",
"Epoch 01653: val_acc did not improve from 0.86765\n",
"Epoch 1654/3000\n",
" - 39s - loss: 0.2309 - acc: 0.9748 - val_loss: 0.9506 - val_acc: 0.8564\n",
"\n",
"Epoch 01654: val_acc did not improve from 0.86765\n",
"Epoch 1655/3000\n",
" - 39s - loss: 0.2422 - acc: 0.9724 - val_loss: 0.9509 - val_acc: 0.8564\n",
"\n",
"Epoch 01655: val_acc did not improve from 0.86765\n",
"Epoch 1656/3000\n",
" - 39s - loss: 0.2348 - acc: 0.9736 - val_loss: 0.9294 - val_acc: 0.8571\n",
"\n",
"Epoch 01656: val_acc did not improve from 0.86765\n",
"Epoch 1657/3000\n",
" - 40s - loss: 0.2360 - acc: 0.9748 - val_loss: 0.9647 - val_acc: 0.8568\n",
"\n",
"Epoch 01657: val_acc did not improve from 0.86765\n",
"Epoch 1658/3000\n",
" - 39s - loss: 0.2443 - acc: 0.9733 - val_loss: 0.9579 - val_acc: 0.8556\n",
"\n",
"Epoch 01658: val_acc did not improve from 0.86765\n",
"Epoch 1659/3000\n",
" - 40s - loss: 0.2359 - acc: 0.9734 - val_loss: 0.9637 - val_acc: 0.8529\n",
"\n",
"Epoch 01659: val_acc did not improve from 0.86765\n",
"Epoch 1660/3000\n",
" - 39s - loss: 0.2457 - acc: 0.9719 - val_loss: 0.9487 - val_acc: 0.8517\n",
"\n",
"Epoch 01660: val_acc did not improve from 0.86765\n",
"Epoch 1661/3000\n",
" - 39s - loss: 0.2475 - acc: 0.9684 - val_loss: 0.9477 - val_acc: 0.8552\n",
"\n",
"Epoch 01661: val_acc did not improve from 0.86765\n",
"Epoch 1662/3000\n",
" - 39s - loss: 0.2420 - acc: 0.9700 - val_loss: 0.9458 - val_acc: 0.8568\n",
"\n",
"Epoch 01662: val_acc did not improve from 0.86765\n",
"Epoch 1663/3000\n",
" - 39s - loss: 0.2477 - acc: 0.9697 - val_loss: 0.9442 - val_acc: 0.8571\n",
"\n",
"Epoch 01663: val_acc did not improve from 0.86765\n",
"Epoch 1664/3000\n",
" - 39s - loss: 0.2430 - acc: 0.9700 - val_loss: 0.9368 - val_acc: 0.8606\n",
"\n",
"Epoch 01664: val_acc did not improve from 0.86765\n",
"Epoch 1665/3000\n",
" - 39s - loss: 0.2441 - acc: 0.9712 - val_loss: 0.9506 - val_acc: 0.8583\n",
"\n",
"Epoch 01665: val_acc did not improve from 0.86765\n",
"Epoch 1666/3000\n",
" - 39s - loss: 0.2415 - acc: 0.9716 - val_loss: 0.9450 - val_acc: 0.8587\n",
"\n",
"Epoch 01666: val_acc did not improve from 0.86765\n",
"Epoch 1667/3000\n",
" - 39s - loss: 0.2457 - acc: 0.9700 - val_loss: 0.9005 - val_acc: 0.8618\n",
"\n",
"Epoch 01667: val_acc did not improve from 0.86765\n",
"Epoch 1668/3000\n",
" - 39s - loss: 0.2392 - acc: 0.9724 - val_loss: 0.9510 - val_acc: 0.8560\n",
"\n",
"Epoch 01668: val_acc did not improve from 0.86765\n",
"Epoch 1669/3000\n",
" - 39s - loss: 0.2405 - acc: 0.9718 - val_loss: 0.9585 - val_acc: 0.8536\n",
"\n",
"Epoch 01669: val_acc did not improve from 0.86765\n",
"Epoch 1670/3000\n",
" - 39s - loss: 0.2377 - acc: 0.9724 - val_loss: 0.9441 - val_acc: 0.8564\n",
"\n",
"Epoch 01670: val_acc did not improve from 0.86765\n",
"Epoch 1671/3000\n",
" - 40s - loss: 0.2461 - acc: 0.9713 - val_loss: 0.9328 - val_acc: 0.8564\n",
"\n",
"Epoch 01671: val_acc did not improve from 0.86765\n",
"Epoch 1672/3000\n",
" - 39s - loss: 0.2339 - acc: 0.9743 - val_loss: 0.9569 - val_acc: 0.8564\n",
"\n",
"Epoch 01672: val_acc did not improve from 0.86765\n",
"Epoch 1673/3000\n",
" - 39s - loss: 0.2422 - acc: 0.9730 - val_loss: 0.9678 - val_acc: 0.8536\n",
"\n",
"Epoch 01673: val_acc did not improve from 0.86765\n",
"Epoch 1674/3000\n",
" - 39s - loss: 0.2379 - acc: 0.9736 - val_loss: 0.9578 - val_acc: 0.8548\n",
"\n",
"Epoch 01674: val_acc did not improve from 0.86765\n",
"Epoch 1675/3000\n",
" - 39s - loss: 0.2431 - acc: 0.9719 - val_loss: 0.9396 - val_acc: 0.8540\n",
"\n",
"Epoch 01675: val_acc did not improve from 0.86765\n",
"Epoch 1676/3000\n",
" - 39s - loss: 0.2344 - acc: 0.9739 - val_loss: 0.9647 - val_acc: 0.8540\n",
"\n",
"Epoch 01676: val_acc did not improve from 0.86765\n",
"Epoch 1677/3000\n",
" - 39s - loss: 0.2400 - acc: 0.9716 - val_loss: 0.9397 - val_acc: 0.8552\n",
"\n",
"Epoch 01677: val_acc did not improve from 0.86765\n",
"Epoch 1678/3000\n",
" - 39s - loss: 0.2354 - acc: 0.9725 - val_loss: 0.9399 - val_acc: 0.8614\n",
"\n",
"Epoch 01678: val_acc did not improve from 0.86765\n",
"Epoch 1679/3000\n",
" - 40s - loss: 0.2425 - acc: 0.9730 - val_loss: 0.9595 - val_acc: 0.8552\n",
"\n",
"Epoch 01679: val_acc did not improve from 0.86765\n",
"Epoch 1680/3000\n",
" - 39s - loss: 0.2448 - acc: 0.9701 - val_loss: 0.9767 - val_acc: 0.8575\n",
"\n",
"Epoch 01680: val_acc did not improve from 0.86765\n",
"Epoch 1681/3000\n",
" - 39s - loss: 0.2462 - acc: 0.9716 - val_loss: 0.9558 - val_acc: 0.8571\n",
"\n",
"Epoch 01681: val_acc did not improve from 0.86765\n",
"Epoch 1682/3000\n",
" - 39s - loss: 0.2443 - acc: 0.9710 - val_loss: 0.9277 - val_acc: 0.8591\n",
"\n",
"Epoch 01682: val_acc did not improve from 0.86765\n",
"Epoch 1683/3000\n",
" - 39s - loss: 0.2431 - acc: 0.9712 - val_loss: 0.9203 - val_acc: 0.8579\n",
"\n",
"Epoch 01683: val_acc did not improve from 0.86765\n",
"Epoch 1684/3000\n",
" - 39s - loss: 0.2371 - acc: 0.9736 - val_loss: 0.9380 - val_acc: 0.8552\n",
"\n",
"Epoch 01684: val_acc did not improve from 0.86765\n",
"Epoch 1685/3000\n",
" - 39s - loss: 0.2406 - acc: 0.9725 - val_loss: 0.9291 - val_acc: 0.8583\n",
"\n",
"Epoch 01685: val_acc did not improve from 0.86765\n",
"Epoch 1686/3000\n",
" - 40s - loss: 0.2362 - acc: 0.9725 - val_loss: 0.9439 - val_acc: 0.8540\n",
"\n",
"Epoch 01686: val_acc did not improve from 0.86765\n",
"Epoch 1687/3000\n",
" - 39s - loss: 0.2435 - acc: 0.9686 - val_loss: 0.9464 - val_acc: 0.8575\n",
"\n",
"Epoch 01687: val_acc did not improve from 0.86765\n",
"Epoch 1688/3000\n",
" - 39s - loss: 0.2474 - acc: 0.9710 - val_loss: 0.9844 - val_acc: 0.8533\n",
"\n",
"Epoch 01688: val_acc did not improve from 0.86765\n",
"Epoch 1689/3000\n",
" - 39s - loss: 0.2352 - acc: 0.9730 - val_loss: 0.9340 - val_acc: 0.8560\n",
"\n",
"Epoch 01689: val_acc did not improve from 0.86765\n",
"Epoch 1690/3000\n",
" - 39s - loss: 0.2444 - acc: 0.9706 - val_loss: 0.9369 - val_acc: 0.8579\n",
"\n",
"Epoch 01690: val_acc did not improve from 0.86765\n",
"Epoch 1691/3000\n",
" - 39s - loss: 0.2453 - acc: 0.9724 - val_loss: 0.9335 - val_acc: 0.8568\n",
"\n",
"Epoch 01691: val_acc did not improve from 0.86765\n",
"Epoch 1692/3000\n",
" - 39s - loss: 0.2431 - acc: 0.9724 - val_loss: 0.9536 - val_acc: 0.8540\n",
"\n",
"Epoch 01692: val_acc did not improve from 0.86765\n",
"Epoch 1693/3000\n",
" - 39s - loss: 0.2437 - acc: 0.9737 - val_loss: 0.9467 - val_acc: 0.8571\n",
"\n",
"Epoch 01693: val_acc did not improve from 0.86765\n",
"Epoch 1694/3000\n",
" - 40s - loss: 0.2387 - acc: 0.9743 - val_loss: 0.9340 - val_acc: 0.8591\n",
"\n",
"Epoch 01694: val_acc did not improve from 0.86765\n",
"Epoch 1695/3000\n",
" - 39s - loss: 0.2499 - acc: 0.9695 - val_loss: 0.9462 - val_acc: 0.8571\n",
"\n",
"Epoch 01695: val_acc did not improve from 0.86765\n",
"Epoch 1696/3000\n",
" - 39s - loss: 0.2421 - acc: 0.9710 - val_loss: 0.9613 - val_acc: 0.8575\n",
"\n",
"Epoch 01696: val_acc did not improve from 0.86765\n",
"Epoch 1697/3000\n",
" - 40s - loss: 0.2481 - acc: 0.9722 - val_loss: 0.9473 - val_acc: 0.8583\n",
"\n",
"Epoch 01697: val_acc did not improve from 0.86765\n",
"Epoch 1698/3000\n",
" - 40s - loss: 0.2469 - acc: 0.9713 - val_loss: 0.9375 - val_acc: 0.8560\n",
"\n",
"Epoch 01698: val_acc did not improve from 0.86765\n",
"Epoch 1699/3000\n",
" - 39s - loss: 0.2516 - acc: 0.9692 - val_loss: 0.9508 - val_acc: 0.8548\n",
"\n",
"Epoch 01699: val_acc did not improve from 0.86765\n",
"Epoch 1700/3000\n",
" - 39s - loss: 0.2453 - acc: 0.9715 - val_loss: 0.9098 - val_acc: 0.8610\n",
"\n",
"Epoch 01700: val_acc did not improve from 0.86765\n",
"Epoch 1701/3000\n",
" - 39s - loss: 0.2466 - acc: 0.9713 - val_loss: 0.9457 - val_acc: 0.8568\n",
"\n",
"Epoch 01701: val_acc did not improve from 0.86765\n",
"Epoch 1702/3000\n",
" - 39s - loss: 0.2518 - acc: 0.9712 - val_loss: 0.9541 - val_acc: 0.8540\n",
"\n",
"Epoch 01702: val_acc did not improve from 0.86765\n",
"Epoch 1703/3000\n",
" - 39s - loss: 0.2430 - acc: 0.9716 - val_loss: 0.9383 - val_acc: 0.8571\n",
"\n",
"Epoch 01703: val_acc did not improve from 0.86765\n",
"Epoch 1704/3000\n",
" - 39s - loss: 0.2395 - acc: 0.9721 - val_loss: 0.9496 - val_acc: 0.8571\n",
"\n",
"Epoch 01704: val_acc did not improve from 0.86765\n",
"Epoch 1705/3000\n",
" - 39s - loss: 0.2404 - acc: 0.9721 - val_loss: 0.9436 - val_acc: 0.8579\n",
"\n",
"Epoch 01705: val_acc did not improve from 0.86765\n",
"Epoch 1706/3000\n",
" - 39s - loss: 0.2380 - acc: 0.9701 - val_loss: 0.9312 - val_acc: 0.8599\n",
"\n",
"Epoch 01706: val_acc did not improve from 0.86765\n",
"Epoch 1707/3000\n",
" - 39s - loss: 0.2296 - acc: 0.9731 - val_loss: 0.9177 - val_acc: 0.8591\n",
"\n",
"Epoch 01707: val_acc did not improve from 0.86765\n",
"Epoch 1708/3000\n",
" - 39s - loss: 0.2388 - acc: 0.9697 - val_loss: 0.9459 - val_acc: 0.8505\n",
"\n",
"Epoch 01708: val_acc did not improve from 0.86765\n",
"Epoch 1709/3000\n",
" - 39s - loss: 0.2434 - acc: 0.9724 - val_loss: 0.9242 - val_acc: 0.8599\n",
"\n",
"Epoch 01709: val_acc did not improve from 0.86765\n",
"Epoch 1710/3000\n",
" - 39s - loss: 0.2486 - acc: 0.9691 - val_loss: 0.9343 - val_acc: 0.8552\n",
"\n",
"Epoch 01710: val_acc did not improve from 0.86765\n",
"Epoch 1711/3000\n",
" - 39s - loss: 0.2421 - acc: 0.9698 - val_loss: 0.9341 - val_acc: 0.8579\n",
"\n",
"Epoch 01711: val_acc did not improve from 0.86765\n",
"Epoch 1712/3000\n",
" - 39s - loss: 0.2464 - acc: 0.9733 - val_loss: 0.9185 - val_acc: 0.8606\n",
"\n",
"Epoch 01712: val_acc did not improve from 0.86765\n",
"Epoch 1713/3000\n",
" - 39s - loss: 0.2354 - acc: 0.9749 - val_loss: 0.9516 - val_acc: 0.8552\n",
"\n",
"Epoch 01713: val_acc did not improve from 0.86765\n",
"Epoch 1714/3000\n",
" - 39s - loss: 0.2401 - acc: 0.9725 - val_loss: 0.9506 - val_acc: 0.8571\n",
"\n",
"Epoch 01714: val_acc did not improve from 0.86765\n",
"Epoch 1715/3000\n",
" - 39s - loss: 0.2400 - acc: 0.9706 - val_loss: 0.9420 - val_acc: 0.8544\n",
"\n",
"Epoch 01715: val_acc did not improve from 0.86765\n",
"Epoch 1716/3000\n",
" - 39s - loss: 0.2410 - acc: 0.9722 - val_loss: 0.9052 - val_acc: 0.8626\n",
"\n",
"Epoch 01716: val_acc did not improve from 0.86765\n",
"Epoch 1717/3000\n",
" - 39s - loss: 0.2403 - acc: 0.9710 - val_loss: 0.9490 - val_acc: 0.8548\n",
"\n",
"Epoch 01717: val_acc did not improve from 0.86765\n",
"Epoch 1718/3000\n",
" - 39s - loss: 0.2377 - acc: 0.9716 - val_loss: 0.9385 - val_acc: 0.8564\n",
"\n",
"Epoch 01718: val_acc did not improve from 0.86765\n",
"Epoch 1719/3000\n",
" - 39s - loss: 0.2404 - acc: 0.9737 - val_loss: 0.9540 - val_acc: 0.8568\n",
"\n",
"Epoch 01719: val_acc did not improve from 0.86765\n",
"Epoch 1720/3000\n",
" - 39s - loss: 0.2437 - acc: 0.9692 - val_loss: 0.9487 - val_acc: 0.8564\n",
"\n",
"Epoch 01720: val_acc did not improve from 0.86765\n",
"Epoch 1721/3000\n",
" - 40s - loss: 0.2416 - acc: 0.9700 - val_loss: 0.9389 - val_acc: 0.8540\n",
"\n",
"Epoch 01721: val_acc did not improve from 0.86765\n",
"Epoch 1722/3000\n",
" - 39s - loss: 0.2386 - acc: 0.9713 - val_loss: 0.9300 - val_acc: 0.8587\n",
"\n",
"Epoch 01722: val_acc did not improve from 0.86765\n",
"Epoch 1723/3000\n",
" - 39s - loss: 0.2559 - acc: 0.9683 - val_loss: 0.9424 - val_acc: 0.8571\n",
"\n",
"Epoch 01723: val_acc did not improve from 0.86765\n",
"Epoch 1724/3000\n",
" - 39s - loss: 0.2483 - acc: 0.9695 - val_loss: 0.9352 - val_acc: 0.8571\n",
"\n",
"Epoch 01724: val_acc did not improve from 0.86765\n",
"Epoch 1725/3000\n",
" - 39s - loss: 0.2453 - acc: 0.9728 - val_loss: 0.9717 - val_acc: 0.8509\n",
"\n",
"Epoch 01725: val_acc did not improve from 0.86765\n",
"Epoch 1726/3000\n",
" - 39s - loss: 0.2370 - acc: 0.9743 - val_loss: 0.9454 - val_acc: 0.8548\n",
"\n",
"Epoch 01726: val_acc did not improve from 0.86765\n",
"Epoch 1727/3000\n",
" - 39s - loss: 0.2430 - acc: 0.9718 - val_loss: 0.9518 - val_acc: 0.8533\n",
"\n",
"Epoch 01727: val_acc did not improve from 0.86765\n",
"Epoch 1728/3000\n",
" - 39s - loss: 0.2425 - acc: 0.9718 - val_loss: 0.9127 - val_acc: 0.8610\n",
"\n",
"Epoch 01728: val_acc did not improve from 0.86765\n",
"Epoch 1729/3000\n",
" - 40s - loss: 0.2421 - acc: 0.9713 - val_loss: 0.9084 - val_acc: 0.8638\n",
"\n",
"Epoch 01729: val_acc did not improve from 0.86765\n",
"Epoch 1730/3000\n",
" - 39s - loss: 0.2365 - acc: 0.9730 - val_loss: 0.9388 - val_acc: 0.8587\n",
"\n",
"Epoch 01730: val_acc did not improve from 0.86765\n",
"Epoch 1731/3000\n",
" - 39s - loss: 0.2448 - acc: 0.9709 - val_loss: 0.9196 - val_acc: 0.8591\n",
"\n",
"Epoch 01731: val_acc did not improve from 0.86765\n",
"Epoch 1732/3000\n",
" - 39s - loss: 0.2353 - acc: 0.9725 - val_loss: 0.9516 - val_acc: 0.8544\n",
"\n",
"Epoch 01732: val_acc did not improve from 0.86765\n",
"Epoch 1733/3000\n",
" - 39s - loss: 0.2402 - acc: 0.9712 - val_loss: 0.9143 - val_acc: 0.8595\n",
"\n",
"Epoch 01733: val_acc did not improve from 0.86765\n",
"Epoch 1734/3000\n",
" - 39s - loss: 0.2426 - acc: 0.9733 - val_loss: 0.9374 - val_acc: 0.8564\n",
"\n",
"Epoch 01734: val_acc did not improve from 0.86765\n",
"Epoch 1735/3000\n",
" - 39s - loss: 0.2436 - acc: 0.9710 - val_loss: 0.9604 - val_acc: 0.8548\n",
"\n",
"Epoch 01735: val_acc did not improve from 0.86765\n",
"Epoch 1736/3000\n",
" - 39s - loss: 0.2415 - acc: 0.9721 - val_loss: 0.9298 - val_acc: 0.8556\n",
"\n",
"Epoch 01736: val_acc did not improve from 0.86765\n",
"Epoch 1737/3000\n",
" - 39s - loss: 0.2399 - acc: 0.9725 - val_loss: 0.9215 - val_acc: 0.8583\n",
"\n",
"Epoch 01737: val_acc did not improve from 0.86765\n",
"Epoch 1738/3000\n",
" - 39s - loss: 0.2310 - acc: 0.9736 - val_loss: 0.9474 - val_acc: 0.8591\n",
"\n",
"Epoch 01738: val_acc did not improve from 0.86765\n",
"Epoch 1739/3000\n",
" - 39s - loss: 0.2395 - acc: 0.9712 - val_loss: 0.9407 - val_acc: 0.8587\n",
"\n",
"Epoch 01739: val_acc did not improve from 0.86765\n",
"Epoch 1740/3000\n",
" - 39s - loss: 0.2460 - acc: 0.9703 - val_loss: 0.9558 - val_acc: 0.8548\n",
"\n",
"Epoch 01740: val_acc did not improve from 0.86765\n",
"Epoch 1741/3000\n",
" - 39s - loss: 0.2414 - acc: 0.9694 - val_loss: 0.9251 - val_acc: 0.8579\n",
"\n",
"Epoch 01741: val_acc did not improve from 0.86765\n",
"Epoch 1742/3000\n",
" - 39s - loss: 0.2349 - acc: 0.9715 - val_loss: 0.9195 - val_acc: 0.8595\n",
"\n",
"Epoch 01742: val_acc did not improve from 0.86765\n",
"Epoch 1743/3000\n",
" - 39s - loss: 0.2456 - acc: 0.9689 - val_loss: 0.9571 - val_acc: 0.8540\n",
"\n",
"Epoch 01743: val_acc did not improve from 0.86765\n",
"Epoch 1744/3000\n",
" - 39s - loss: 0.2492 - acc: 0.9719 - val_loss: 0.9217 - val_acc: 0.8606\n",
"\n",
"Epoch 01744: val_acc did not improve from 0.86765\n",
"Epoch 1745/3000\n",
" - 39s - loss: 0.2412 - acc: 0.9683 - val_loss: 0.9430 - val_acc: 0.8560\n",
"\n",
"Epoch 01745: val_acc did not improve from 0.86765\n",
"Epoch 1746/3000\n",
" - 39s - loss: 0.2401 - acc: 0.9710 - val_loss: 0.9504 - val_acc: 0.8564\n",
"\n",
"Epoch 01746: val_acc did not improve from 0.86765\n",
"Epoch 1747/3000\n",
" - 39s - loss: 0.2436 - acc: 0.9704 - val_loss: 0.9241 - val_acc: 0.8575\n",
"\n",
"Epoch 01747: val_acc did not improve from 0.86765\n",
"Epoch 1748/3000\n",
" - 39s - loss: 0.2376 - acc: 0.9722 - val_loss: 0.9375 - val_acc: 0.8552\n",
"\n",
"Epoch 01748: val_acc did not improve from 0.86765\n",
"Epoch 1749/3000\n",
" - 39s - loss: 0.2347 - acc: 0.9751 - val_loss: 0.9259 - val_acc: 0.8583\n",
"\n",
"Epoch 01749: val_acc did not improve from 0.86765\n",
"Epoch 1750/3000\n",
" - 40s - loss: 0.2415 - acc: 0.9743 - val_loss: 0.9516 - val_acc: 0.8575\n",
"\n",
"Epoch 01750: val_acc did not improve from 0.86765\n",
"Epoch 1751/3000\n",
" - 40s - loss: 0.2434 - acc: 0.9724 - val_loss: 0.9135 - val_acc: 0.8568\n",
"\n",
"Epoch 01751: val_acc did not improve from 0.86765\n",
"Epoch 1752/3000\n",
" - 39s - loss: 0.2428 - acc: 0.9703 - val_loss: 0.9301 - val_acc: 0.8575\n",
"\n",
"Epoch 01752: val_acc did not improve from 0.86765\n",
"Epoch 1753/3000\n",
" - 39s - loss: 0.2417 - acc: 0.9727 - val_loss: 0.9306 - val_acc: 0.8568\n",
"\n",
"Epoch 01753: val_acc did not improve from 0.86765\n",
"Epoch 1754/3000\n",
" - 39s - loss: 0.2461 - acc: 0.9712 - val_loss: 0.9300 - val_acc: 0.8564\n",
"\n",
"Epoch 01754: val_acc did not improve from 0.86765\n",
"Epoch 1755/3000\n",
" - 39s - loss: 0.2412 - acc: 0.9727 - val_loss: 0.9236 - val_acc: 0.8610\n",
"\n",
"Epoch 01755: val_acc did not improve from 0.86765\n",
"Epoch 1756/3000\n",
" - 39s - loss: 0.2459 - acc: 0.9709 - val_loss: 0.9483 - val_acc: 0.8587\n",
"\n",
"Epoch 01756: val_acc did not improve from 0.86765\n",
"Epoch 1757/3000\n",
" - 39s - loss: 0.2420 - acc: 0.9710 - val_loss: 0.9172 - val_acc: 0.8603\n",
"\n",
"Epoch 01757: val_acc did not improve from 0.86765\n",
"Epoch 1758/3000\n",
" - 39s - loss: 0.2355 - acc: 0.9763 - val_loss: 0.9278 - val_acc: 0.8595\n",
"\n",
"Epoch 01758: val_acc did not improve from 0.86765\n",
"Epoch 1759/3000\n",
" - 39s - loss: 0.2438 - acc: 0.9692 - val_loss: 0.9344 - val_acc: 0.8564\n",
"\n",
"Epoch 01759: val_acc did not improve from 0.86765\n",
"Epoch 1760/3000\n",
" - 39s - loss: 0.2326 - acc: 0.9737 - val_loss: 0.9386 - val_acc: 0.8599\n",
"\n",
"Epoch 01760: val_acc did not improve from 0.86765\n",
"Epoch 1761/3000\n",
" - 39s - loss: 0.2436 - acc: 0.9718 - val_loss: 0.9084 - val_acc: 0.8614\n",
"\n",
"Epoch 01761: val_acc did not improve from 0.86765\n",
"Epoch 1762/3000\n",
" - 39s - loss: 0.2398 - acc: 0.9722 - val_loss: 0.9349 - val_acc: 0.8595\n",
"\n",
"Epoch 01762: val_acc did not improve from 0.86765\n",
"Epoch 1763/3000\n",
" - 39s - loss: 0.2419 - acc: 0.9722 - val_loss: 0.9150 - val_acc: 0.8587\n",
"\n",
"Epoch 01763: val_acc did not improve from 0.86765\n",
"Epoch 1764/3000\n",
" - 39s - loss: 0.2461 - acc: 0.9697 - val_loss: 0.9402 - val_acc: 0.8583\n",
"\n",
"Epoch 01764: val_acc did not improve from 0.86765\n",
"Epoch 1765/3000\n",
" - 39s - loss: 0.2457 - acc: 0.9681 - val_loss: 0.9335 - val_acc: 0.8568\n",
"\n",
"Epoch 01765: val_acc did not improve from 0.86765\n",
"Epoch 1766/3000\n",
" - 39s - loss: 0.2419 - acc: 0.9713 - val_loss: 0.9420 - val_acc: 0.8536\n",
"\n",
"Epoch 01766: val_acc did not improve from 0.86765\n",
"Epoch 1767/3000\n",
" - 39s - loss: 0.2509 - acc: 0.9691 - val_loss: 0.9329 - val_acc: 0.8591\n",
"\n",
"Epoch 01767: val_acc did not improve from 0.86765\n",
"Epoch 1768/3000\n",
" - 39s - loss: 0.2456 - acc: 0.9730 - val_loss: 0.9043 - val_acc: 0.8618\n",
"\n",
"Epoch 01768: val_acc did not improve from 0.86765\n",
"Epoch 1769/3000\n",
" - 39s - loss: 0.2478 - acc: 0.9694 - val_loss: 0.9261 - val_acc: 0.8599\n",
"\n",
"Epoch 01769: val_acc did not improve from 0.86765\n",
"Epoch 1770/3000\n",
" - 39s - loss: 0.2429 - acc: 0.9725 - val_loss: 0.9364 - val_acc: 0.8544\n",
"\n",
"Epoch 01770: val_acc did not improve from 0.86765\n",
"Epoch 1771/3000\n",
" - 39s - loss: 0.2381 - acc: 0.9733 - val_loss: 0.9368 - val_acc: 0.8544\n",
"\n",
"Epoch 01771: val_acc did not improve from 0.86765\n",
"Epoch 1772/3000\n",
" - 39s - loss: 0.2486 - acc: 0.9692 - val_loss: 0.9250 - val_acc: 0.8560\n",
"\n",
"Epoch 01772: val_acc did not improve from 0.86765\n",
"Epoch 1773/3000\n",
" - 39s - loss: 0.2443 - acc: 0.9687 - val_loss: 0.9404 - val_acc: 0.8568\n",
"\n",
"Epoch 01773: val_acc did not improve from 0.86765\n",
"Epoch 1774/3000\n",
" - 39s - loss: 0.2495 - acc: 0.9692 - val_loss: 0.9378 - val_acc: 0.8564\n",
"\n",
"Epoch 01774: val_acc did not improve from 0.86765\n",
"Epoch 1775/3000\n",
" - 39s - loss: 0.2415 - acc: 0.9719 - val_loss: 0.9238 - val_acc: 0.8610\n",
"\n",
"Epoch 01775: val_acc did not improve from 0.86765\n",
"Epoch 1776/3000\n",
" - 39s - loss: 0.2394 - acc: 0.9716 - val_loss: 0.9160 - val_acc: 0.8614\n",
"\n",
"Epoch 01776: val_acc did not improve from 0.86765\n",
"Epoch 1777/3000\n",
" - 39s - loss: 0.2438 - acc: 0.9715 - val_loss: 0.9439 - val_acc: 0.8575\n",
"\n",
"Epoch 01777: val_acc did not improve from 0.86765\n",
"Epoch 1778/3000\n",
" - 39s - loss: 0.2338 - acc: 0.9737 - val_loss: 0.9241 - val_acc: 0.8599\n",
"\n",
"Epoch 01778: val_acc did not improve from 0.86765\n",
"Epoch 1779/3000\n",
" - 39s - loss: 0.2466 - acc: 0.9706 - val_loss: 0.9761 - val_acc: 0.8509\n",
"\n",
"Epoch 01779: val_acc did not improve from 0.86765\n",
"Epoch 1780/3000\n",
" - 39s - loss: 0.2373 - acc: 0.9724 - val_loss: 0.9495 - val_acc: 0.8571\n",
"\n",
"Epoch 01780: val_acc did not improve from 0.86765\n",
"Epoch 1781/3000\n",
" - 39s - loss: 0.2429 - acc: 0.9677 - val_loss: 0.9758 - val_acc: 0.8564\n",
"\n",
"Epoch 01781: val_acc did not improve from 0.86765\n",
"Epoch 1782/3000\n",
" - 39s - loss: 0.2457 - acc: 0.9700 - val_loss: 0.9374 - val_acc: 0.8552\n",
"\n",
"Epoch 01782: val_acc did not improve from 0.86765\n",
"Epoch 1783/3000\n",
" - 39s - loss: 0.2374 - acc: 0.9721 - val_loss: 0.9752 - val_acc: 0.8540\n",
"\n",
"Epoch 01783: val_acc did not improve from 0.86765\n",
"Epoch 1784/3000\n",
" - 39s - loss: 0.2423 - acc: 0.9719 - val_loss: 0.9427 - val_acc: 0.8544\n",
"\n",
"Epoch 01784: val_acc did not improve from 0.86765\n",
"Epoch 1785/3000\n",
" - 39s - loss: 0.2375 - acc: 0.9700 - val_loss: 0.9716 - val_acc: 0.8513\n",
"\n",
"Epoch 01785: val_acc did not improve from 0.86765\n",
"Epoch 1786/3000\n",
" - 39s - loss: 0.2413 - acc: 0.9719 - val_loss: 0.9522 - val_acc: 0.8548\n",
"\n",
"Epoch 01786: val_acc did not improve from 0.86765\n",
"Epoch 1787/3000\n",
" - 39s - loss: 0.2407 - acc: 0.9704 - val_loss: 0.9640 - val_acc: 0.8529\n",
"\n",
"Epoch 01787: val_acc did not improve from 0.86765\n",
"Epoch 1788/3000\n",
" - 39s - loss: 0.2317 - acc: 0.9748 - val_loss: 0.9631 - val_acc: 0.8552\n",
"\n",
"Epoch 01788: val_acc did not improve from 0.86765\n",
"Epoch 1789/3000\n",
" - 39s - loss: 0.2437 - acc: 0.9710 - val_loss: 0.9231 - val_acc: 0.8587\n",
"\n",
"Epoch 01789: val_acc did not improve from 0.86765\n",
"Epoch 1790/3000\n",
" - 39s - loss: 0.2380 - acc: 0.9733 - val_loss: 0.9308 - val_acc: 0.8548\n",
"\n",
"Epoch 01790: val_acc did not improve from 0.86765\n",
"Epoch 1791/3000\n",
" - 39s - loss: 0.2389 - acc: 0.9721 - val_loss: 0.9357 - val_acc: 0.8564\n",
"\n",
"Epoch 01791: val_acc did not improve from 0.86765\n",
"Epoch 1792/3000\n",
" - 39s - loss: 0.2424 - acc: 0.9737 - val_loss: 0.9492 - val_acc: 0.8564\n",
"\n",
"Epoch 01792: val_acc did not improve from 0.86765\n",
"Epoch 1793/3000\n",
" - 39s - loss: 0.2438 - acc: 0.9713 - val_loss: 0.9325 - val_acc: 0.8622\n",
"\n",
"Epoch 01793: val_acc did not improve from 0.86765\n",
"Epoch 1794/3000\n",
" - 39s - loss: 0.2479 - acc: 0.9675 - val_loss: 0.9442 - val_acc: 0.8603\n",
"\n",
"Epoch 01794: val_acc did not improve from 0.86765\n",
"Epoch 1795/3000\n",
" - 39s - loss: 0.2486 - acc: 0.9712 - val_loss: 0.9469 - val_acc: 0.8595\n",
"\n",
"Epoch 01795: val_acc did not improve from 0.86765\n",
"Epoch 1796/3000\n",
" - 39s - loss: 0.2459 - acc: 0.9704 - val_loss: 0.9402 - val_acc: 0.8571\n",
"\n",
"Epoch 01796: val_acc did not improve from 0.86765\n",
"Epoch 1797/3000\n",
" - 39s - loss: 0.2418 - acc: 0.9700 - val_loss: 0.9580 - val_acc: 0.8556\n",
"\n",
"Epoch 01797: val_acc did not improve from 0.86765\n",
"Epoch 1798/3000\n",
" - 39s - loss: 0.2442 - acc: 0.9700 - val_loss: 0.9435 - val_acc: 0.8564\n",
"\n",
"Epoch 01798: val_acc did not improve from 0.86765\n",
"Epoch 1799/3000\n",
" - 39s - loss: 0.2406 - acc: 0.9724 - val_loss: 0.9391 - val_acc: 0.8583\n",
"\n",
"Epoch 01799: val_acc did not improve from 0.86765\n",
"Epoch 1800/3000\n",
" - 39s - loss: 0.2368 - acc: 0.9736 - val_loss: 0.9343 - val_acc: 0.8568\n",
"\n",
"Epoch 01800: val_acc did not improve from 0.86765\n",
"Epoch 1801/3000\n",
" - 39s - loss: 0.2366 - acc: 0.9734 - val_loss: 0.9168 - val_acc: 0.8606\n",
"\n",
"Epoch 01801: val_acc did not improve from 0.86765\n",
"Epoch 1802/3000\n",
" - 39s - loss: 0.2345 - acc: 0.9730 - val_loss: 0.9597 - val_acc: 0.8571\n",
"\n",
"Epoch 01802: val_acc did not improve from 0.86765\n",
"Epoch 1803/3000\n",
" - 39s - loss: 0.2450 - acc: 0.9700 - val_loss: 0.9367 - val_acc: 0.8595\n",
"\n",
"Epoch 01803: val_acc did not improve from 0.86765\n",
"Epoch 1804/3000\n",
" - 39s - loss: 0.2403 - acc: 0.9730 - val_loss: 0.9376 - val_acc: 0.8591\n",
"\n",
"Epoch 01804: val_acc did not improve from 0.86765\n",
"Epoch 1805/3000\n",
" - 39s - loss: 0.2388 - acc: 0.9712 - val_loss: 0.9378 - val_acc: 0.8564\n",
"\n",
"Epoch 01805: val_acc did not improve from 0.86765\n",
"Epoch 1806/3000\n",
" - 39s - loss: 0.2422 - acc: 0.9718 - val_loss: 0.9483 - val_acc: 0.8564\n",
"\n",
"Epoch 01806: val_acc did not improve from 0.86765\n",
"Epoch 1807/3000\n",
" - 39s - loss: 0.2417 - acc: 0.9701 - val_loss: 0.9790 - val_acc: 0.8521\n",
"\n",
"Epoch 01807: val_acc did not improve from 0.86765\n",
"Epoch 1808/3000\n",
" - 39s - loss: 0.2371 - acc: 0.9734 - val_loss: 0.9583 - val_acc: 0.8560\n",
"\n",
"Epoch 01808: val_acc did not improve from 0.86765\n",
"Epoch 1809/3000\n",
" - 39s - loss: 0.2282 - acc: 0.9749 - val_loss: 0.9693 - val_acc: 0.8533\n",
"\n",
"Epoch 01809: val_acc did not improve from 0.86765\n",
"Epoch 1810/3000\n",
" - 39s - loss: 0.2411 - acc: 0.9695 - val_loss: 0.9553 - val_acc: 0.8552\n",
"\n",
"Epoch 01810: val_acc did not improve from 0.86765\n",
"Epoch 1811/3000\n",
" - 39s - loss: 0.2520 - acc: 0.9709 - val_loss: 0.9570 - val_acc: 0.8548\n",
"\n",
"Epoch 01811: val_acc did not improve from 0.86765\n",
"Epoch 1812/3000\n",
" - 39s - loss: 0.2471 - acc: 0.9698 - val_loss: 0.9281 - val_acc: 0.8599\n",
"\n",
"Epoch 01812: val_acc did not improve from 0.86765\n",
"Epoch 1813/3000\n",
" - 39s - loss: 0.2417 - acc: 0.9721 - val_loss: 0.9480 - val_acc: 0.8591\n",
"\n",
"Epoch 01813: val_acc did not improve from 0.86765\n",
"Epoch 1814/3000\n",
" - 39s - loss: 0.2373 - acc: 0.9746 - val_loss: 0.9497 - val_acc: 0.8583\n",
"\n",
"Epoch 01814: val_acc did not improve from 0.86765\n",
"Epoch 1815/3000\n",
" - 39s - loss: 0.2368 - acc: 0.9736 - val_loss: 0.9438 - val_acc: 0.8575\n",
"\n",
"Epoch 01815: val_acc did not improve from 0.86765\n",
"Epoch 1816/3000\n",
" - 39s - loss: 0.2418 - acc: 0.9712 - val_loss: 0.9399 - val_acc: 0.8595\n",
"\n",
"Epoch 01816: val_acc did not improve from 0.86765\n",
"Epoch 1817/3000\n",
" - 39s - loss: 0.2405 - acc: 0.9710 - val_loss: 0.9412 - val_acc: 0.8575\n",
"\n",
"Epoch 01817: val_acc did not improve from 0.86765\n",
"Epoch 1818/3000\n",
" - 39s - loss: 0.2468 - acc: 0.9700 - val_loss: 0.9614 - val_acc: 0.8544\n",
"\n",
"Epoch 01818: val_acc did not improve from 0.86765\n",
"Epoch 1819/3000\n",
" - 39s - loss: 0.2416 - acc: 0.9695 - val_loss: 0.9304 - val_acc: 0.8610\n",
"\n",
"Epoch 01819: val_acc did not improve from 0.86765\n",
"Epoch 1820/3000\n",
" - 39s - loss: 0.2365 - acc: 0.9727 - val_loss: 0.9495 - val_acc: 0.8564\n",
"\n",
"Epoch 01820: val_acc did not improve from 0.86765\n",
"Epoch 1821/3000\n",
" - 39s - loss: 0.2381 - acc: 0.9721 - val_loss: 0.9580 - val_acc: 0.8606\n",
"\n",
"Epoch 01821: val_acc did not improve from 0.86765\n",
"Epoch 1822/3000\n",
" - 39s - loss: 0.2413 - acc: 0.9721 - val_loss: 0.9527 - val_acc: 0.8595\n",
"\n",
"Epoch 01822: val_acc did not improve from 0.86765\n",
"Epoch 1823/3000\n",
" - 39s - loss: 0.2466 - acc: 0.9731 - val_loss: 0.9803 - val_acc: 0.8544\n",
"\n",
"Epoch 01823: val_acc did not improve from 0.86765\n",
"Epoch 1824/3000\n",
" - 39s - loss: 0.2419 - acc: 0.9688 - val_loss: 0.9607 - val_acc: 0.8571\n",
"\n",
"Epoch 01824: val_acc did not improve from 0.86765\n",
"Epoch 1825/3000\n",
" - 39s - loss: 0.2390 - acc: 0.9734 - val_loss: 0.9452 - val_acc: 0.8583\n",
"\n",
"Epoch 01825: val_acc did not improve from 0.86765\n",
"Epoch 1826/3000\n",
" - 40s - loss: 0.2399 - acc: 0.9710 - val_loss: 0.9513 - val_acc: 0.8591\n",
"\n",
"Epoch 01826: val_acc did not improve from 0.86765\n",
"Epoch 1827/3000\n",
" - 39s - loss: 0.2378 - acc: 0.9713 - val_loss: 0.9697 - val_acc: 0.8583\n",
"\n",
"Epoch 01827: val_acc did not improve from 0.86765\n",
"Epoch 1828/3000\n",
" - 39s - loss: 0.2388 - acc: 0.9713 - val_loss: 0.9881 - val_acc: 0.8517\n",
"\n",
"Epoch 01828: val_acc did not improve from 0.86765\n",
"Epoch 1829/3000\n",
" - 39s - loss: 0.2408 - acc: 0.9704 - val_loss: 0.9500 - val_acc: 0.8544\n",
"\n",
"Epoch 01829: val_acc did not improve from 0.86765\n",
"Epoch 1830/3000\n",
" - 39s - loss: 0.2329 - acc: 0.9728 - val_loss: 0.9716 - val_acc: 0.8544\n",
"\n",
"Epoch 01830: val_acc did not improve from 0.86765\n",
"Epoch 1831/3000\n",
" - 39s - loss: 0.2355 - acc: 0.9758 - val_loss: 0.9836 - val_acc: 0.8525\n",
"\n",
"Epoch 01831: val_acc did not improve from 0.86765\n",
"Epoch 1832/3000\n",
" - 39s - loss: 0.2432 - acc: 0.9707 - val_loss: 0.9901 - val_acc: 0.8509\n",
"\n",
"Epoch 01832: val_acc did not improve from 0.86765\n",
"Epoch 1833/3000\n",
" - 40s - loss: 0.2499 - acc: 0.9686 - val_loss: 0.9511 - val_acc: 0.8575\n",
"\n",
"Epoch 01833: val_acc did not improve from 0.86765\n",
"Epoch 1834/3000\n",
" - 39s - loss: 0.2431 - acc: 0.9681 - val_loss: 0.9292 - val_acc: 0.8587\n",
"\n",
"Epoch 01834: val_acc did not improve from 0.86765\n",
"Epoch 1835/3000\n",
" - 39s - loss: 0.2391 - acc: 0.9707 - val_loss: 0.9398 - val_acc: 0.8626\n",
"\n",
"Epoch 01835: val_acc did not improve from 0.86765\n",
"Epoch 1836/3000\n",
" - 39s - loss: 0.2392 - acc: 0.9698 - val_loss: 0.9527 - val_acc: 0.8571\n",
"\n",
"Epoch 01836: val_acc did not improve from 0.86765\n",
"Epoch 1837/3000\n",
" - 39s - loss: 0.2405 - acc: 0.9707 - val_loss: 0.9451 - val_acc: 0.8587\n",
"\n",
"Epoch 01837: val_acc did not improve from 0.86765\n",
"Epoch 1838/3000\n",
" - 39s - loss: 0.2516 - acc: 0.9695 - val_loss: 0.9610 - val_acc: 0.8544\n",
"\n",
"Epoch 01838: val_acc did not improve from 0.86765\n",
"Epoch 1839/3000\n",
" - 39s - loss: 0.2381 - acc: 0.9749 - val_loss: 0.9451 - val_acc: 0.8571\n",
"\n",
"Epoch 01839: val_acc did not improve from 0.86765\n",
"Epoch 1840/3000\n",
" - 39s - loss: 0.2408 - acc: 0.9697 - val_loss: 0.9727 - val_acc: 0.8529\n",
"\n",
"Epoch 01840: val_acc did not improve from 0.86765\n",
"Epoch 1841/3000\n",
" - 39s - loss: 0.2402 - acc: 0.9712 - val_loss: 0.9786 - val_acc: 0.8548\n",
"\n",
"Epoch 01841: val_acc did not improve from 0.86765\n",
"Epoch 1842/3000\n",
" - 39s - loss: 0.2398 - acc: 0.9721 - val_loss: 0.9552 - val_acc: 0.8575\n",
"\n",
"Epoch 01842: val_acc did not improve from 0.86765\n",
"Epoch 1843/3000\n",
" - 39s - loss: 0.2389 - acc: 0.9713 - val_loss: 0.9592 - val_acc: 0.8548\n",
"\n",
"Epoch 01843: val_acc did not improve from 0.86765\n",
"Epoch 1844/3000\n",
" - 39s - loss: 0.2357 - acc: 0.9743 - val_loss: 0.9337 - val_acc: 0.8571\n",
"\n",
"Epoch 01844: val_acc did not improve from 0.86765\n",
"Epoch 1845/3000\n",
" - 39s - loss: 0.2399 - acc: 0.9704 - val_loss: 0.9657 - val_acc: 0.8521\n",
"\n",
"Epoch 01845: val_acc did not improve from 0.86765\n",
"Epoch 1846/3000\n",
" - 39s - loss: 0.2329 - acc: 0.9742 - val_loss: 0.9439 - val_acc: 0.8575\n",
"\n",
"Epoch 01846: val_acc did not improve from 0.86765\n",
"Epoch 1847/3000\n",
" - 39s - loss: 0.2435 - acc: 0.9704 - val_loss: 0.9366 - val_acc: 0.8603\n",
"\n",
"Epoch 01847: val_acc did not improve from 0.86765\n",
"Epoch 1848/3000\n",
" - 39s - loss: 0.2470 - acc: 0.9694 - val_loss: 0.9374 - val_acc: 0.8575\n",
"\n",
"Epoch 01848: val_acc did not improve from 0.86765\n",
"Epoch 1849/3000\n",
" - 39s - loss: 0.2396 - acc: 0.9718 - val_loss: 0.9438 - val_acc: 0.8568\n",
"\n",
"Epoch 01849: val_acc did not improve from 0.86765\n",
"Epoch 1850/3000\n",
" - 39s - loss: 0.2414 - acc: 0.9719 - val_loss: 0.9080 - val_acc: 0.8603\n",
"\n",
"Epoch 01850: val_acc did not improve from 0.86765\n",
"Epoch 1851/3000\n",
" - 39s - loss: 0.2434 - acc: 0.9697 - val_loss: 0.9104 - val_acc: 0.8599\n",
"\n",
"Epoch 01851: val_acc did not improve from 0.86765\n",
"Epoch 1852/3000\n",
" - 39s - loss: 0.2470 - acc: 0.9713 - val_loss: 0.9503 - val_acc: 0.8603\n",
"\n",
"Epoch 01852: val_acc did not improve from 0.86765\n",
"Epoch 1853/3000\n",
" - 39s - loss: 0.2453 - acc: 0.9700 - val_loss: 0.9346 - val_acc: 0.8595\n",
"\n",
"Epoch 01853: val_acc did not improve from 0.86765\n",
"Epoch 1854/3000\n",
" - 39s - loss: 0.2381 - acc: 0.9739 - val_loss: 0.9314 - val_acc: 0.8614\n",
"\n",
"Epoch 01854: val_acc did not improve from 0.86765\n",
"Epoch 1855/3000\n",
" - 39s - loss: 0.2400 - acc: 0.9724 - val_loss: 0.9511 - val_acc: 0.8583\n",
"\n",
"Epoch 01855: val_acc did not improve from 0.86765\n",
"Epoch 1856/3000\n",
" - 39s - loss: 0.2430 - acc: 0.9703 - val_loss: 0.9523 - val_acc: 0.8556\n",
"\n",
"Epoch 01856: val_acc did not improve from 0.86765\n",
"Epoch 1857/3000\n",
" - 39s - loss: 0.2361 - acc: 0.9721 - val_loss: 0.9428 - val_acc: 0.8575\n",
"\n",
"Epoch 01857: val_acc did not improve from 0.86765\n",
"Epoch 1858/3000\n",
" - 39s - loss: 0.2398 - acc: 0.9715 - val_loss: 0.9580 - val_acc: 0.8575\n",
"\n",
"Epoch 01858: val_acc did not improve from 0.86765\n",
"Epoch 1859/3000\n",
" - 39s - loss: 0.2347 - acc: 0.9722 - val_loss: 0.9519 - val_acc: 0.8587\n",
"\n",
"Epoch 01859: val_acc did not improve from 0.86765\n",
"Epoch 1860/3000\n",
" - 39s - loss: 0.2367 - acc: 0.9724 - val_loss: 0.9812 - val_acc: 0.8552\n",
"\n",
"Epoch 01860: val_acc did not improve from 0.86765\n",
"Epoch 1861/3000\n",
" - 39s - loss: 0.2304 - acc: 0.9746 - val_loss: 0.9847 - val_acc: 0.8595\n",
"\n",
"Epoch 01861: val_acc did not improve from 0.86765\n",
"Epoch 1862/3000\n",
" - 39s - loss: 0.2331 - acc: 0.9755 - val_loss: 0.9559 - val_acc: 0.8599\n",
"\n",
"Epoch 01862: val_acc did not improve from 0.86765\n",
"Epoch 1863/3000\n",
" - 39s - loss: 0.2399 - acc: 0.9687 - val_loss: 0.9609 - val_acc: 0.8587\n",
"\n",
"Epoch 01863: val_acc did not improve from 0.86765\n",
"Epoch 1864/3000\n",
" - 39s - loss: 0.2407 - acc: 0.9706 - val_loss: 0.9607 - val_acc: 0.8544\n",
"\n",
"Epoch 01864: val_acc did not improve from 0.86765\n",
"Epoch 1865/3000\n",
" - 39s - loss: 0.2422 - acc: 0.9727 - val_loss: 0.9649 - val_acc: 0.8579\n",
"\n",
"Epoch 01865: val_acc did not improve from 0.86765\n",
"Epoch 1866/3000\n",
" - 39s - loss: 0.2418 - acc: 0.9719 - val_loss: 0.9655 - val_acc: 0.8564\n",
"\n",
"Epoch 01866: val_acc did not improve from 0.86765\n",
"Epoch 1867/3000\n",
" - 39s - loss: 0.2550 - acc: 0.9694 - val_loss: 0.9776 - val_acc: 0.8525\n",
"\n",
"Epoch 01867: val_acc did not improve from 0.86765\n",
"Epoch 1868/3000\n",
" - 39s - loss: 0.2452 - acc: 0.9709 - val_loss: 0.9613 - val_acc: 0.8536\n",
"\n",
"Epoch 01868: val_acc did not improve from 0.86765\n",
"Epoch 1869/3000\n",
" - 39s - loss: 0.2403 - acc: 0.9728 - val_loss: 0.9542 - val_acc: 0.8552\n",
"\n",
"Epoch 01869: val_acc did not improve from 0.86765\n",
"Epoch 1870/3000\n",
" - 39s - loss: 0.2402 - acc: 0.9721 - val_loss: 0.9660 - val_acc: 0.8579\n",
"\n",
"Epoch 01870: val_acc did not improve from 0.86765\n",
"Epoch 1871/3000\n",
" - 39s - loss: 0.2373 - acc: 0.9727 - val_loss: 0.9606 - val_acc: 0.8575\n",
"\n",
"Epoch 01871: val_acc did not improve from 0.86765\n",
"Epoch 1872/3000\n",
" - 39s - loss: 0.2404 - acc: 0.9718 - val_loss: 0.9379 - val_acc: 0.8591\n",
"\n",
"Epoch 01872: val_acc did not improve from 0.86765\n",
"Epoch 1873/3000\n",
" - 39s - loss: 0.2452 - acc: 0.9716 - val_loss: 0.9386 - val_acc: 0.8571\n",
"\n",
"Epoch 01873: val_acc did not improve from 0.86765\n",
"Epoch 1874/3000\n",
" - 39s - loss: 0.2387 - acc: 0.9721 - val_loss: 0.9413 - val_acc: 0.8556\n",
"\n",
"Epoch 01874: val_acc did not improve from 0.86765\n",
"Epoch 1875/3000\n",
" - 39s - loss: 0.2489 - acc: 0.9695 - val_loss: 0.9421 - val_acc: 0.8587\n",
"\n",
"Epoch 01875: val_acc did not improve from 0.86765\n",
"Epoch 1876/3000\n",
" - 39s - loss: 0.2346 - acc: 0.9718 - val_loss: 0.9266 - val_acc: 0.8603\n",
"\n",
"Epoch 01876: val_acc did not improve from 0.86765\n",
"Epoch 1877/3000\n",
" - 39s - loss: 0.2360 - acc: 0.9713 - val_loss: 0.9745 - val_acc: 0.8564\n",
"\n",
"Epoch 01877: val_acc did not improve from 0.86765\n",
"Epoch 1878/3000\n",
" - 39s - loss: 0.2408 - acc: 0.9727 - val_loss: 0.9634 - val_acc: 0.8587\n",
"\n",
"Epoch 01878: val_acc did not improve from 0.86765\n",
"Epoch 1879/3000\n",
" - 39s - loss: 0.2326 - acc: 0.9733 - val_loss: 0.9706 - val_acc: 0.8556\n",
"\n",
"Epoch 01879: val_acc did not improve from 0.86765\n",
"Epoch 1880/3000\n",
" - 39s - loss: 0.2360 - acc: 0.9746 - val_loss: 0.9577 - val_acc: 0.8571\n",
"\n",
"Epoch 01880: val_acc did not improve from 0.86765\n",
"Epoch 1881/3000\n",
" - 39s - loss: 0.2414 - acc: 0.9713 - val_loss: 0.9587 - val_acc: 0.8560\n",
"\n",
"Epoch 01881: val_acc did not improve from 0.86765\n",
"Epoch 1882/3000\n",
" - 39s - loss: 0.2357 - acc: 0.9713 - val_loss: 0.9674 - val_acc: 0.8544\n",
"\n",
"Epoch 01882: val_acc did not improve from 0.86765\n",
"Epoch 1883/3000\n",
" - 39s - loss: 0.2440 - acc: 0.9727 - val_loss: 0.9707 - val_acc: 0.8560\n",
"\n",
"Epoch 01883: val_acc did not improve from 0.86765\n",
"Epoch 1884/3000\n",
" - 39s - loss: 0.2407 - acc: 0.9709 - val_loss: 0.9582 - val_acc: 0.8540\n",
"\n",
"Epoch 01884: val_acc did not improve from 0.86765\n",
"Epoch 1885/3000\n",
" - 39s - loss: 0.2495 - acc: 0.9689 - val_loss: 0.9689 - val_acc: 0.8540\n",
"\n",
"Epoch 01885: val_acc did not improve from 0.86765\n",
"Epoch 1886/3000\n",
" - 39s - loss: 0.2439 - acc: 0.9713 - val_loss: 0.9941 - val_acc: 0.8490\n",
"\n",
"Epoch 01886: val_acc did not improve from 0.86765\n",
"Epoch 1887/3000\n",
" - 39s - loss: 0.2344 - acc: 0.9727 - val_loss: 0.9802 - val_acc: 0.8540\n",
"\n",
"Epoch 01887: val_acc did not improve from 0.86765\n",
"Epoch 1888/3000\n",
" - 39s - loss: 0.2310 - acc: 0.9722 - val_loss: 0.9709 - val_acc: 0.8529\n",
"\n",
"Epoch 01888: val_acc did not improve from 0.86765\n",
"Epoch 1889/3000\n",
" - 39s - loss: 0.2405 - acc: 0.9721 - val_loss: 0.9663 - val_acc: 0.8548\n",
"\n",
"Epoch 01889: val_acc did not improve from 0.86765\n",
"Epoch 1890/3000\n",
" - 39s - loss: 0.2370 - acc: 0.9728 - val_loss: 0.9883 - val_acc: 0.8525\n",
"\n",
"Epoch 01890: val_acc did not improve from 0.86765\n",
"Epoch 1891/3000\n",
" - 39s - loss: 0.2374 - acc: 0.9734 - val_loss: 0.9596 - val_acc: 0.8587\n",
"\n",
"Epoch 01891: val_acc did not improve from 0.86765\n",
"Epoch 1892/3000\n",
" - 39s - loss: 0.2392 - acc: 0.9733 - val_loss: 0.9252 - val_acc: 0.8614\n",
"\n",
"Epoch 01892: val_acc did not improve from 0.86765\n",
"Epoch 1893/3000\n",
" - 39s - loss: 0.2387 - acc: 0.9739 - val_loss: 0.9429 - val_acc: 0.8606\n",
"\n",
"Epoch 01893: val_acc did not improve from 0.86765\n",
"Epoch 1894/3000\n",
" - 39s - loss: 0.2370 - acc: 0.9718 - val_loss: 0.9621 - val_acc: 0.8568\n",
"\n",
"Epoch 01894: val_acc did not improve from 0.86765\n",
"Epoch 1895/3000\n",
" - 39s - loss: 0.2471 - acc: 0.9706 - val_loss: 0.9455 - val_acc: 0.8571\n",
"\n",
"Epoch 01895: val_acc did not improve from 0.86765\n",
"Epoch 1896/3000\n",
" - 39s - loss: 0.2334 - acc: 0.9724 - val_loss: 0.9653 - val_acc: 0.8560\n",
"\n",
"Epoch 01896: val_acc did not improve from 0.86765\n",
"Epoch 1897/3000\n",
" - 39s - loss: 0.2416 - acc: 0.9710 - val_loss: 0.9728 - val_acc: 0.8560\n",
"\n",
"Epoch 01897: val_acc did not improve from 0.86765\n",
"Epoch 1898/3000\n",
" - 39s - loss: 0.2371 - acc: 0.9737 - val_loss: 0.9766 - val_acc: 0.8556\n",
"\n",
"Epoch 01898: val_acc did not improve from 0.86765\n",
"Epoch 1899/3000\n",
" - 39s - loss: 0.2401 - acc: 0.9706 - val_loss: 0.9570 - val_acc: 0.8564\n",
"\n",
"Epoch 01899: val_acc did not improve from 0.86765\n",
"Epoch 1900/3000\n",
" - 39s - loss: 0.2432 - acc: 0.9707 - val_loss: 0.9477 - val_acc: 0.8591\n",
"\n",
"Epoch 01900: val_acc did not improve from 0.86765\n",
"Epoch 1901/3000\n",
" - 39s - loss: 0.2355 - acc: 0.9739 - val_loss: 0.9596 - val_acc: 0.8552\n",
"\n",
"Epoch 01901: val_acc did not improve from 0.86765\n",
"Epoch 1902/3000\n",
" - 39s - loss: 0.2412 - acc: 0.9721 - val_loss: 0.9618 - val_acc: 0.8513\n",
"\n",
"Epoch 01902: val_acc did not improve from 0.86765\n",
"Epoch 1903/3000\n",
" - 39s - loss: 0.2387 - acc: 0.9719 - val_loss: 0.9374 - val_acc: 0.8606\n",
"\n",
"Epoch 01903: val_acc did not improve from 0.86765\n",
"Epoch 1904/3000\n",
" - 39s - loss: 0.2339 - acc: 0.9728 - val_loss: 0.9285 - val_acc: 0.8579\n",
"\n",
"Epoch 01904: val_acc did not improve from 0.86765\n",
"Epoch 1905/3000\n",
" - 39s - loss: 0.2372 - acc: 0.9746 - val_loss: 0.9552 - val_acc: 0.8571\n",
"\n",
"Epoch 01905: val_acc did not improve from 0.86765\n",
"Epoch 1906/3000\n",
" - 39s - loss: 0.2459 - acc: 0.9709 - val_loss: 0.9340 - val_acc: 0.8599\n",
"\n",
"Epoch 01906: val_acc did not improve from 0.86765\n",
"Epoch 1907/3000\n",
" - 39s - loss: 0.2344 - acc: 0.9733 - val_loss: 0.9364 - val_acc: 0.8587\n",
"\n",
"Epoch 01907: val_acc did not improve from 0.86765\n",
"Epoch 1908/3000\n",
" - 39s - loss: 0.2379 - acc: 0.9725 - val_loss: 0.9703 - val_acc: 0.8587\n",
"\n",
"Epoch 01908: val_acc did not improve from 0.86765\n",
"Epoch 1909/3000\n",
" - 39s - loss: 0.2433 - acc: 0.9713 - val_loss: 0.9424 - val_acc: 0.8568\n",
"\n",
"Epoch 01909: val_acc did not improve from 0.86765\n",
"Epoch 1910/3000\n",
" - 39s - loss: 0.2371 - acc: 0.9742 - val_loss: 0.9427 - val_acc: 0.8579\n",
"\n",
"Epoch 01910: val_acc did not improve from 0.86765\n",
"Epoch 1911/3000\n",
" - 39s - loss: 0.2467 - acc: 0.9706 - val_loss: 0.9509 - val_acc: 0.8540\n",
"\n",
"Epoch 01911: val_acc did not improve from 0.86765\n",
"Epoch 1912/3000\n",
" - 39s - loss: 0.2344 - acc: 0.9724 - val_loss: 0.9493 - val_acc: 0.8556\n",
"\n",
"Epoch 01912: val_acc did not improve from 0.86765\n",
"Epoch 1913/3000\n",
" - 39s - loss: 0.2317 - acc: 0.9734 - val_loss: 0.9739 - val_acc: 0.8533\n",
"\n",
"Epoch 01913: val_acc did not improve from 0.86765\n",
"Epoch 1914/3000\n",
" - 39s - loss: 0.2358 - acc: 0.9725 - val_loss: 0.9722 - val_acc: 0.8548\n",
"\n",
"Epoch 01914: val_acc did not improve from 0.86765\n",
"Epoch 1915/3000\n",
" - 39s - loss: 0.2417 - acc: 0.9719 - val_loss: 0.9384 - val_acc: 0.8583\n",
"\n",
"Epoch 01915: val_acc did not improve from 0.86765\n",
"Epoch 1916/3000\n",
" - 39s - loss: 0.2421 - acc: 0.9722 - val_loss: 0.9634 - val_acc: 0.8536\n",
"\n",
"Epoch 01916: val_acc did not improve from 0.86765\n",
"Epoch 1917/3000\n",
" - 39s - loss: 0.2352 - acc: 0.9722 - val_loss: 0.9507 - val_acc: 0.8552\n",
"\n",
"Epoch 01917: val_acc did not improve from 0.86765\n",
"Epoch 1918/3000\n",
" - 40s - loss: 0.2385 - acc: 0.9709 - val_loss: 0.9990 - val_acc: 0.8497\n",
"\n",
"Epoch 01918: val_acc did not improve from 0.86765\n",
"Epoch 1919/3000\n",
" - 39s - loss: 0.2386 - acc: 0.9739 - val_loss: 0.9816 - val_acc: 0.8521\n",
"\n",
"Epoch 01919: val_acc did not improve from 0.86765\n",
"Epoch 1920/3000\n",
" - 39s - loss: 0.2395 - acc: 0.9722 - val_loss: 0.9367 - val_acc: 0.8591\n",
"\n",
"Epoch 01920: val_acc did not improve from 0.86765\n",
"Epoch 1921/3000\n",
" - 39s - loss: 0.2367 - acc: 0.9730 - val_loss: 0.9630 - val_acc: 0.8591\n",
"\n",
"Epoch 01921: val_acc did not improve from 0.86765\n",
"Epoch 1922/3000\n",
" - 39s - loss: 0.2380 - acc: 0.9716 - val_loss: 0.9766 - val_acc: 0.8556\n",
"\n",
"Epoch 01922: val_acc did not improve from 0.86765\n",
"Epoch 1923/3000\n",
" - 39s - loss: 0.2269 - acc: 0.9752 - val_loss: 0.9504 - val_acc: 0.8595\n",
"\n",
"Epoch 01923: val_acc did not improve from 0.86765\n",
"Epoch 1924/3000\n",
" - 39s - loss: 0.2434 - acc: 0.9704 - val_loss: 0.9474 - val_acc: 0.8560\n",
"\n",
"Epoch 01924: val_acc did not improve from 0.86765\n",
"Epoch 1925/3000\n",
" - 39s - loss: 0.2433 - acc: 0.9716 - val_loss: 0.9505 - val_acc: 0.8548\n",
"\n",
"Epoch 01925: val_acc did not improve from 0.86765\n",
"Epoch 1926/3000\n",
" - 39s - loss: 0.2325 - acc: 0.9721 - val_loss: 0.9665 - val_acc: 0.8525\n",
"\n",
"Epoch 01926: val_acc did not improve from 0.86765\n",
"Epoch 1927/3000\n",
" - 39s - loss: 0.2381 - acc: 0.9740 - val_loss: 0.9683 - val_acc: 0.8533\n",
"\n",
"Epoch 01927: val_acc did not improve from 0.86765\n",
"Epoch 1928/3000\n",
" - 39s - loss: 0.2500 - acc: 0.9698 - val_loss: 0.9197 - val_acc: 0.8587\n",
"\n",
"Epoch 01928: val_acc did not improve from 0.86765\n",
"Epoch 1929/3000\n",
" - 39s - loss: 0.2470 - acc: 0.9701 - val_loss: 0.9270 - val_acc: 0.8595\n",
"\n",
"Epoch 01929: val_acc did not improve from 0.86765\n",
"Epoch 1930/3000\n",
" - 40s - loss: 0.2369 - acc: 0.9709 - val_loss: 0.9403 - val_acc: 0.8591\n",
"\n",
"Epoch 01930: val_acc did not improve from 0.86765\n",
"Epoch 1931/3000\n",
" - 39s - loss: 0.2388 - acc: 0.9742 - val_loss: 0.9646 - val_acc: 0.8529\n",
"\n",
"Epoch 01931: val_acc did not improve from 0.86765\n",
"Epoch 1932/3000\n",
" - 39s - loss: 0.2443 - acc: 0.9715 - val_loss: 0.9565 - val_acc: 0.8536\n",
"\n",
"Epoch 01932: val_acc did not improve from 0.86765\n",
"Epoch 1933/3000\n",
" - 39s - loss: 0.2288 - acc: 0.9758 - val_loss: 0.9446 - val_acc: 0.8564\n",
"\n",
"Epoch 01933: val_acc did not improve from 0.86765\n",
"Epoch 1934/3000\n",
" - 39s - loss: 0.2394 - acc: 0.9715 - val_loss: 0.9450 - val_acc: 0.8564\n",
"\n",
"Epoch 01934: val_acc did not improve from 0.86765\n",
"Epoch 1935/3000\n",
" - 39s - loss: 0.2435 - acc: 0.9719 - val_loss: 0.9380 - val_acc: 0.8544\n",
"\n",
"Epoch 01935: val_acc did not improve from 0.86765\n",
"Epoch 1936/3000\n",
" - 39s - loss: 0.2404 - acc: 0.9713 - val_loss: 0.9436 - val_acc: 0.8552\n",
"\n",
"Epoch 01936: val_acc did not improve from 0.86765\n",
"Epoch 1937/3000\n",
" - 39s - loss: 0.2463 - acc: 0.9707 - val_loss: 0.9316 - val_acc: 0.8568\n",
"\n",
"Epoch 01937: val_acc did not improve from 0.86765\n",
"Epoch 1938/3000\n",
" - 39s - loss: 0.2431 - acc: 0.9710 - val_loss: 0.9430 - val_acc: 0.8548\n",
"\n",
"Epoch 01938: val_acc did not improve from 0.86765\n",
"Epoch 1939/3000\n",
" - 39s - loss: 0.2424 - acc: 0.9713 - val_loss: 0.9609 - val_acc: 0.8571\n",
"\n",
"Epoch 01939: val_acc did not improve from 0.86765\n",
"Epoch 1940/3000\n",
" - 39s - loss: 0.2343 - acc: 0.9727 - val_loss: 0.9469 - val_acc: 0.8552\n",
"\n",
"Epoch 01940: val_acc did not improve from 0.86765\n",
"Epoch 1941/3000\n",
" - 39s - loss: 0.2327 - acc: 0.9739 - val_loss: 0.9730 - val_acc: 0.8521\n",
"\n",
"Epoch 01941: val_acc did not improve from 0.86765\n",
"Epoch 1942/3000\n",
" - 39s - loss: 0.2485 - acc: 0.9701 - val_loss: 0.9613 - val_acc: 0.8533\n",
"\n",
"Epoch 01942: val_acc did not improve from 0.86765\n",
"Epoch 1943/3000\n",
" - 39s - loss: 0.2386 - acc: 0.9712 - val_loss: 0.9709 - val_acc: 0.8509\n",
"\n",
"Epoch 01943: val_acc did not improve from 0.86765\n",
"Epoch 1944/3000\n",
" - 39s - loss: 0.2444 - acc: 0.9719 - val_loss: 0.9265 - val_acc: 0.8587\n",
"\n",
"Epoch 01944: val_acc did not improve from 0.86765\n",
"Epoch 1945/3000\n",
" - 39s - loss: 0.2355 - acc: 0.9736 - val_loss: 0.9613 - val_acc: 0.8552\n",
"\n",
"Epoch 01945: val_acc did not improve from 0.86765\n",
"Epoch 1946/3000\n",
" - 39s - loss: 0.2347 - acc: 0.9731 - val_loss: 0.9583 - val_acc: 0.8521\n",
"\n",
"Epoch 01946: val_acc did not improve from 0.86765\n",
"Epoch 1947/3000\n",
" - 39s - loss: 0.2366 - acc: 0.9727 - val_loss: 0.9558 - val_acc: 0.8571\n",
"\n",
"Epoch 01947: val_acc did not improve from 0.86765\n",
"Epoch 1948/3000\n",
" - 39s - loss: 0.2376 - acc: 0.9719 - val_loss: 0.9360 - val_acc: 0.8583\n",
"\n",
"Epoch 01948: val_acc did not improve from 0.86765\n",
"Epoch 1949/3000\n",
" - 39s - loss: 0.2278 - acc: 0.9742 - val_loss: 0.9444 - val_acc: 0.8587\n",
"\n",
"Epoch 01949: val_acc did not improve from 0.86765\n",
"Epoch 1950/3000\n",
" - 39s - loss: 0.2490 - acc: 0.9703 - val_loss: 0.9326 - val_acc: 0.8599\n",
"\n",
"Epoch 01950: val_acc did not improve from 0.86765\n",
"Epoch 1951/3000\n",
" - 39s - loss: 0.2354 - acc: 0.9736 - val_loss: 0.9240 - val_acc: 0.8575\n",
"\n",
"Epoch 01951: val_acc did not improve from 0.86765\n",
"Epoch 1952/3000\n",
" - 39s - loss: 0.2372 - acc: 0.9724 - val_loss: 0.9178 - val_acc: 0.8587\n",
"\n",
"Epoch 01952: val_acc did not improve from 0.86765\n",
"Epoch 1953/3000\n",
" - 39s - loss: 0.2367 - acc: 0.9715 - val_loss: 0.9453 - val_acc: 0.8540\n",
"\n",
"Epoch 01953: val_acc did not improve from 0.86765\n",
"Epoch 1954/3000\n",
" - 39s - loss: 0.2439 - acc: 0.9698 - val_loss: 0.9189 - val_acc: 0.8579\n",
"\n",
"Epoch 01954: val_acc did not improve from 0.86765\n",
"Epoch 1955/3000\n",
" - 39s - loss: 0.2349 - acc: 0.9748 - val_loss: 0.9128 - val_acc: 0.8575\n",
"\n",
"Epoch 01955: val_acc did not improve from 0.86765\n",
"Epoch 1956/3000\n",
" - 39s - loss: 0.2483 - acc: 0.9721 - val_loss: 0.9379 - val_acc: 0.8571\n",
"\n",
"Epoch 01956: val_acc did not improve from 0.86765\n",
"Epoch 1957/3000\n",
" - 39s - loss: 0.2426 - acc: 0.9713 - val_loss: 0.9401 - val_acc: 0.8564\n",
"\n",
"Epoch 01957: val_acc did not improve from 0.86765\n",
"Epoch 1958/3000\n",
" - 39s - loss: 0.2320 - acc: 0.9725 - val_loss: 0.9494 - val_acc: 0.8548\n",
"\n",
"Epoch 01958: val_acc did not improve from 0.86765\n",
"Epoch 1959/3000\n",
" - 39s - loss: 0.2462 - acc: 0.9709 - val_loss: 0.9283 - val_acc: 0.8575\n",
"\n",
"Epoch 01959: val_acc did not improve from 0.86765\n",
"Epoch 1960/3000\n",
" - 39s - loss: 0.2322 - acc: 0.9731 - val_loss: 0.9307 - val_acc: 0.8595\n",
"\n",
"Epoch 01960: val_acc did not improve from 0.86765\n",
"Epoch 1961/3000\n",
" - 39s - loss: 0.2386 - acc: 0.9725 - val_loss: 0.9237 - val_acc: 0.8599\n",
"\n",
"Epoch 01961: val_acc did not improve from 0.86765\n",
"Epoch 1962/3000\n",
" - 39s - loss: 0.2421 - acc: 0.9718 - val_loss: 0.9575 - val_acc: 0.8579\n",
"\n",
"Epoch 01962: val_acc did not improve from 0.86765\n",
"Epoch 1963/3000\n",
" - 39s - loss: 0.2354 - acc: 0.9716 - val_loss: 0.9744 - val_acc: 0.8536\n",
"\n",
"Epoch 01963: val_acc did not improve from 0.86765\n",
"Epoch 1964/3000\n",
" - 39s - loss: 0.2351 - acc: 0.9707 - val_loss: 0.9565 - val_acc: 0.8560\n",
"\n",
"Epoch 01964: val_acc did not improve from 0.86765\n",
"Epoch 1965/3000\n",
" - 39s - loss: 0.2380 - acc: 0.9739 - val_loss: 0.9547 - val_acc: 0.8568\n",
"\n",
"Epoch 01965: val_acc did not improve from 0.86765\n",
"Epoch 1966/3000\n",
" - 39s - loss: 0.2353 - acc: 0.9737 - val_loss: 0.9342 - val_acc: 0.8587\n",
"\n",
"Epoch 01966: val_acc did not improve from 0.86765\n",
"Epoch 1967/3000\n",
" - 39s - loss: 0.2357 - acc: 0.9733 - val_loss: 0.9350 - val_acc: 0.8575\n",
"\n",
"Epoch 01967: val_acc did not improve from 0.86765\n",
"Epoch 1968/3000\n",
" - 39s - loss: 0.2315 - acc: 0.9734 - val_loss: 0.9570 - val_acc: 0.8568\n",
"\n",
"Epoch 01968: val_acc did not improve from 0.86765\n",
"Epoch 1969/3000\n",
" - 39s - loss: 0.2418 - acc: 0.9700 - val_loss: 0.9726 - val_acc: 0.8529\n",
"\n",
"Epoch 01969: val_acc did not improve from 0.86765\n",
"Epoch 1970/3000\n",
" - 39s - loss: 0.2409 - acc: 0.9716 - val_loss: 0.9846 - val_acc: 0.8529\n",
"\n",
"Epoch 01970: val_acc did not improve from 0.86765\n",
"Epoch 1971/3000\n",
" - 39s - loss: 0.2337 - acc: 0.9733 - val_loss: 0.9560 - val_acc: 0.8556\n",
"\n",
"Epoch 01971: val_acc did not improve from 0.86765\n",
"Epoch 1972/3000\n",
" - 39s - loss: 0.2431 - acc: 0.9716 - val_loss: 0.9473 - val_acc: 0.8583\n",
"\n",
"Epoch 01972: val_acc did not improve from 0.86765\n",
"Epoch 1973/3000\n",
" - 39s - loss: 0.2388 - acc: 0.9713 - val_loss: 0.9491 - val_acc: 0.8595\n",
"\n",
"Epoch 01973: val_acc did not improve from 0.86765\n",
"Epoch 1974/3000\n",
" - 39s - loss: 0.2320 - acc: 0.9742 - val_loss: 0.9748 - val_acc: 0.8536\n",
"\n",
"Epoch 01974: val_acc did not improve from 0.86765\n",
"Epoch 1975/3000\n",
" - 39s - loss: 0.2362 - acc: 0.9712 - val_loss: 0.9847 - val_acc: 0.8525\n",
"\n",
"Epoch 01975: val_acc did not improve from 0.86765\n",
"Epoch 1976/3000\n",
" - 39s - loss: 0.2345 - acc: 0.9737 - val_loss: 0.9525 - val_acc: 0.8548\n",
"\n",
"Epoch 01976: val_acc did not improve from 0.86765\n",
"Epoch 1977/3000\n",
" - 39s - loss: 0.2358 - acc: 0.9728 - val_loss: 0.9498 - val_acc: 0.8564\n",
"\n",
"Epoch 01977: val_acc did not improve from 0.86765\n",
"Epoch 1978/3000\n",
" - 39s - loss: 0.2410 - acc: 0.9719 - val_loss: 0.9489 - val_acc: 0.8587\n",
"\n",
"Epoch 01978: val_acc did not improve from 0.86765\n",
"Epoch 1979/3000\n",
" - 40s - loss: 0.2351 - acc: 0.9719 - val_loss: 0.9670 - val_acc: 0.8548\n",
"\n",
"Epoch 01979: val_acc did not improve from 0.86765\n",
"Epoch 1980/3000\n",
" - 39s - loss: 0.2305 - acc: 0.9755 - val_loss: 0.9234 - val_acc: 0.8591\n",
"\n",
"Epoch 01980: val_acc did not improve from 0.86765\n",
"Epoch 1981/3000\n",
" - 39s - loss: 0.2433 - acc: 0.9709 - val_loss: 0.9302 - val_acc: 0.8544\n",
"\n",
"Epoch 01981: val_acc did not improve from 0.86765\n",
"Epoch 1982/3000\n",
" - 39s - loss: 0.2398 - acc: 0.9710 - val_loss: 0.9491 - val_acc: 0.8599\n",
"\n",
"Epoch 01982: val_acc did not improve from 0.86765\n",
"Epoch 1983/3000\n",
" - 39s - loss: 0.2390 - acc: 0.9719 - val_loss: 0.9484 - val_acc: 0.8610\n",
"\n",
"Epoch 01983: val_acc did not improve from 0.86765\n",
"Epoch 1984/3000\n",
" - 39s - loss: 0.2371 - acc: 0.9710 - val_loss: 0.9230 - val_acc: 0.8575\n",
"\n",
"Epoch 01984: val_acc did not improve from 0.86765\n",
"Epoch 1985/3000\n",
" - 39s - loss: 0.2370 - acc: 0.9718 - val_loss: 0.9209 - val_acc: 0.8595\n",
"\n",
"Epoch 01985: val_acc did not improve from 0.86765\n",
"Epoch 1986/3000\n",
" - 39s - loss: 0.2372 - acc: 0.9731 - val_loss: 0.9489 - val_acc: 0.8564\n",
"\n",
"Epoch 01986: val_acc did not improve from 0.86765\n",
"Epoch 1987/3000\n",
" - 39s - loss: 0.2398 - acc: 0.9746 - val_loss: 0.9605 - val_acc: 0.8564\n",
"\n",
"Epoch 01987: val_acc did not improve from 0.86765\n",
"Epoch 1988/3000\n",
" - 40s - loss: 0.2354 - acc: 0.9724 - val_loss: 0.9813 - val_acc: 0.8533\n",
"\n",
"Epoch 01988: val_acc did not improve from 0.86765\n",
"Epoch 1989/3000\n",
" - 39s - loss: 0.2438 - acc: 0.9718 - val_loss: 0.9515 - val_acc: 0.8571\n",
"\n",
"Epoch 01989: val_acc did not improve from 0.86765\n",
"Epoch 1990/3000\n",
" - 39s - loss: 0.2381 - acc: 0.9724 - val_loss: 0.9762 - val_acc: 0.8544\n",
"\n",
"Epoch 01990: val_acc did not improve from 0.86765\n",
"Epoch 1991/3000\n",
" - 39s - loss: 0.2378 - acc: 0.9716 - val_loss: 0.9334 - val_acc: 0.8610\n",
"\n",
"Epoch 01991: val_acc did not improve from 0.86765\n",
"Epoch 1992/3000\n",
" - 40s - loss: 0.2349 - acc: 0.9710 - val_loss: 0.9288 - val_acc: 0.8587\n",
"\n",
"Epoch 01992: val_acc did not improve from 0.86765\n",
"Epoch 1993/3000\n",
" - 39s - loss: 0.2394 - acc: 0.9724 - val_loss: 0.9279 - val_acc: 0.8575\n",
"\n",
"Epoch 01993: val_acc did not improve from 0.86765\n",
"Epoch 1994/3000\n",
" - 39s - loss: 0.2526 - acc: 0.9697 - val_loss: 0.9624 - val_acc: 0.8552\n",
"\n",
"Epoch 01994: val_acc did not improve from 0.86765\n",
"Epoch 1995/3000\n",
" - 39s - loss: 0.2390 - acc: 0.9710 - val_loss: 0.9695 - val_acc: 0.8509\n",
"\n",
"Epoch 01995: val_acc did not improve from 0.86765\n",
"Epoch 1996/3000\n",
" - 39s - loss: 0.2423 - acc: 0.9725 - val_loss: 0.9476 - val_acc: 0.8575\n",
"\n",
"Epoch 01996: val_acc did not improve from 0.86765\n",
"Epoch 1997/3000\n",
" - 39s - loss: 0.2412 - acc: 0.9707 - val_loss: 0.9169 - val_acc: 0.8595\n",
"\n",
"Epoch 01997: val_acc did not improve from 0.86765\n",
"Epoch 1998/3000\n",
" - 39s - loss: 0.2388 - acc: 0.9731 - val_loss: 0.9313 - val_acc: 0.8568\n",
"\n",
"Epoch 01998: val_acc did not improve from 0.86765\n",
"Epoch 1999/3000\n",
" - 39s - loss: 0.2416 - acc: 0.9718 - val_loss: 0.9367 - val_acc: 0.8575\n",
"\n",
"Epoch 01999: val_acc did not improve from 0.86765\n",
"Epoch 2000/3000\n",
" - 39s - loss: 0.2459 - acc: 0.9697 - val_loss: 0.9481 - val_acc: 0.8540\n",
"\n",
"Epoch 02000: val_acc did not improve from 0.86765\n",
"Epoch 2001/3000\n",
" - 39s - loss: 0.2441 - acc: 0.9719 - val_loss: 0.9557 - val_acc: 0.8556\n",
"\n",
"Epoch 02001: val_acc did not improve from 0.86765\n",
"Epoch 2002/3000\n",
" - 39s - loss: 0.2404 - acc: 0.9709 - val_loss: 0.9537 - val_acc: 0.8533\n",
"\n",
"Epoch 02002: val_acc did not improve from 0.86765\n",
"Epoch 2003/3000\n",
" - 39s - loss: 0.2445 - acc: 0.9695 - val_loss: 0.9502 - val_acc: 0.8564\n",
"\n",
"Epoch 02003: val_acc did not improve from 0.86765\n",
"Epoch 2004/3000\n",
" - 39s - loss: 0.2377 - acc: 0.9728 - val_loss: 0.9505 - val_acc: 0.8556\n",
"\n",
"Epoch 02004: val_acc did not improve from 0.86765\n",
"Epoch 2005/3000\n",
" - 40s - loss: 0.2439 - acc: 0.9712 - val_loss: 0.9444 - val_acc: 0.8575\n",
"\n",
"Epoch 02005: val_acc did not improve from 0.86765\n",
"Epoch 2006/3000\n",
" - 39s - loss: 0.2411 - acc: 0.9698 - val_loss: 0.9675 - val_acc: 0.8552\n",
"\n",
"Epoch 02006: val_acc did not improve from 0.86765\n",
"Epoch 2007/3000\n",
" - 39s - loss: 0.2413 - acc: 0.9725 - val_loss: 0.9461 - val_acc: 0.8571\n",
"\n",
"Epoch 02007: val_acc did not improve from 0.86765\n",
"Epoch 2008/3000\n",
" - 39s - loss: 0.2400 - acc: 0.9716 - val_loss: 0.9730 - val_acc: 0.8474\n",
"\n",
"Epoch 02008: val_acc did not improve from 0.86765\n",
"Epoch 2009/3000\n",
" - 39s - loss: 0.2425 - acc: 0.9691 - val_loss: 0.9421 - val_acc: 0.8583\n",
"\n",
"Epoch 02009: val_acc did not improve from 0.86765\n",
"Epoch 2010/3000\n",
" - 39s - loss: 0.2375 - acc: 0.9740 - val_loss: 0.9678 - val_acc: 0.8568\n",
"\n",
"Epoch 02010: val_acc did not improve from 0.86765\n",
"Epoch 2011/3000\n",
" - 39s - loss: 0.2391 - acc: 0.9704 - val_loss: 0.9357 - val_acc: 0.8571\n",
"\n",
"Epoch 02011: val_acc did not improve from 0.86765\n",
"Epoch 2012/3000\n",
" - 39s - loss: 0.2442 - acc: 0.9715 - val_loss: 0.9617 - val_acc: 0.8556\n",
"\n",
"Epoch 02012: val_acc did not improve from 0.86765\n",
"Epoch 2013/3000\n",
" - 39s - loss: 0.2321 - acc: 0.9739 - val_loss: 0.9584 - val_acc: 0.8540\n",
"\n",
"Epoch 02013: val_acc did not improve from 0.86765\n",
"Epoch 2014/3000\n",
" - 39s - loss: 0.2488 - acc: 0.9706 - val_loss: 0.9299 - val_acc: 0.8579\n",
"\n",
"Epoch 02014: val_acc did not improve from 0.86765\n",
"Epoch 2015/3000\n",
" - 39s - loss: 0.2391 - acc: 0.9694 - val_loss: 0.9469 - val_acc: 0.8595\n",
"\n",
"Epoch 02015: val_acc did not improve from 0.86765\n",
"Epoch 2016/3000\n",
" - 39s - loss: 0.2331 - acc: 0.9734 - val_loss: 0.9361 - val_acc: 0.8595\n",
"\n",
"Epoch 02016: val_acc did not improve from 0.86765\n",
"Epoch 2017/3000\n",
" - 39s - loss: 0.2400 - acc: 0.9709 - val_loss: 0.9471 - val_acc: 0.8571\n",
"\n",
"Epoch 02017: val_acc did not improve from 0.86765\n",
"Epoch 2018/3000\n",
" - 39s - loss: 0.2352 - acc: 0.9737 - val_loss: 0.9547 - val_acc: 0.8548\n",
"\n",
"Epoch 02018: val_acc did not improve from 0.86765\n",
"Epoch 2019/3000\n",
" - 39s - loss: 0.2366 - acc: 0.9724 - val_loss: 0.9492 - val_acc: 0.8536\n",
"\n",
"Epoch 02019: val_acc did not improve from 0.86765\n",
"Epoch 2020/3000\n",
" - 39s - loss: 0.2387 - acc: 0.9713 - val_loss: 0.9556 - val_acc: 0.8536\n",
"\n",
"Epoch 02020: val_acc did not improve from 0.86765\n",
"Epoch 2021/3000\n",
" - 39s - loss: 0.2384 - acc: 0.9701 - val_loss: 0.9613 - val_acc: 0.8548\n",
"\n",
"Epoch 02021: val_acc did not improve from 0.86765\n",
"Epoch 2022/3000\n",
" - 39s - loss: 0.2335 - acc: 0.9704 - val_loss: 0.9613 - val_acc: 0.8583\n",
"\n",
"Epoch 02022: val_acc did not improve from 0.86765\n",
"Epoch 2023/3000\n",
" - 39s - loss: 0.2354 - acc: 0.9718 - val_loss: 0.9449 - val_acc: 0.8583\n",
"\n",
"Epoch 02023: val_acc did not improve from 0.86765\n",
"Epoch 2024/3000\n",
" - 39s - loss: 0.2371 - acc: 0.9742 - val_loss: 0.9633 - val_acc: 0.8540\n",
"\n",
"Epoch 02024: val_acc did not improve from 0.86765\n",
"Epoch 2025/3000\n",
" - 39s - loss: 0.2383 - acc: 0.9719 - val_loss: 0.9627 - val_acc: 0.8579\n",
"\n",
"Epoch 02025: val_acc did not improve from 0.86765\n",
"Epoch 2026/3000\n",
" - 39s - loss: 0.2476 - acc: 0.9712 - val_loss: 0.9534 - val_acc: 0.8533\n",
"\n",
"Epoch 02026: val_acc did not improve from 0.86765\n",
"Epoch 2027/3000\n",
" - 39s - loss: 0.2458 - acc: 0.9700 - val_loss: 0.9551 - val_acc: 0.8540\n",
"\n",
"Epoch 02027: val_acc did not improve from 0.86765\n",
"Epoch 2028/3000\n",
" - 39s - loss: 0.2380 - acc: 0.9716 - val_loss: 0.9125 - val_acc: 0.8579\n",
"\n",
"Epoch 02028: val_acc did not improve from 0.86765\n",
"Epoch 2029/3000\n",
" - 39s - loss: 0.2400 - acc: 0.9715 - val_loss: 0.9414 - val_acc: 0.8544\n",
"\n",
"Epoch 02029: val_acc did not improve from 0.86765\n",
"Epoch 2030/3000\n",
" - 39s - loss: 0.2453 - acc: 0.9718 - val_loss: 0.9240 - val_acc: 0.8540\n",
"\n",
"Epoch 02030: val_acc did not improve from 0.86765\n",
"Epoch 2031/3000\n",
" - 39s - loss: 0.2412 - acc: 0.9706 - val_loss: 0.9580 - val_acc: 0.8544\n",
"\n",
"Epoch 02031: val_acc did not improve from 0.86765\n",
"Epoch 2032/3000\n",
" - 39s - loss: 0.2414 - acc: 0.9692 - val_loss: 0.9617 - val_acc: 0.8544\n",
"\n",
"Epoch 02032: val_acc did not improve from 0.86765\n",
"Epoch 2033/3000\n",
" - 39s - loss: 0.2322 - acc: 0.9724 - val_loss: 0.9299 - val_acc: 0.8575\n",
"\n",
"Epoch 02033: val_acc did not improve from 0.86765\n",
"Epoch 2034/3000\n",
" - 39s - loss: 0.2407 - acc: 0.9722 - val_loss: 0.9504 - val_acc: 0.8540\n",
"\n",
"Epoch 02034: val_acc did not improve from 0.86765\n",
"Epoch 2035/3000\n",
" - 39s - loss: 0.2412 - acc: 0.9704 - val_loss: 0.9584 - val_acc: 0.8525\n",
"\n",
"Epoch 02035: val_acc did not improve from 0.86765\n",
"Epoch 2036/3000\n",
" - 39s - loss: 0.2291 - acc: 0.9752 - val_loss: 0.9353 - val_acc: 0.8575\n",
"\n",
"Epoch 02036: val_acc did not improve from 0.86765\n",
"Epoch 2037/3000\n",
" - 39s - loss: 0.2465 - acc: 0.9713 - val_loss: 0.9433 - val_acc: 0.8548\n",
"\n",
"Epoch 02037: val_acc did not improve from 0.86765\n",
"Epoch 2038/3000\n",
" - 39s - loss: 0.2313 - acc: 0.9755 - val_loss: 0.9622 - val_acc: 0.8564\n",
"\n",
"Epoch 02038: val_acc did not improve from 0.86765\n",
"Epoch 2039/3000\n",
" - 39s - loss: 0.2351 - acc: 0.9742 - val_loss: 0.9695 - val_acc: 0.8544\n",
"\n",
"Epoch 02039: val_acc did not improve from 0.86765\n",
"Epoch 2040/3000\n",
" - 39s - loss: 0.2315 - acc: 0.9748 - val_loss: 0.9526 - val_acc: 0.8552\n",
"\n",
"Epoch 02040: val_acc did not improve from 0.86765\n",
"Epoch 2041/3000\n",
" - 39s - loss: 0.2416 - acc: 0.9721 - val_loss: 0.9791 - val_acc: 0.8521\n",
"\n",
"Epoch 02041: val_acc did not improve from 0.86765\n",
"Epoch 2042/3000\n",
" - 39s - loss: 0.2393 - acc: 0.9700 - val_loss: 0.9864 - val_acc: 0.8494\n",
"\n",
"Epoch 02042: val_acc did not improve from 0.86765\n",
"Epoch 2043/3000\n",
" - 39s - loss: 0.2399 - acc: 0.9706 - val_loss: 0.9681 - val_acc: 0.8533\n",
"\n",
"Epoch 02043: val_acc did not improve from 0.86765\n",
"Epoch 2044/3000\n",
" - 39s - loss: 0.2462 - acc: 0.9707 - val_loss: 0.9471 - val_acc: 0.8568\n",
"\n",
"Epoch 02044: val_acc did not improve from 0.86765\n",
"Epoch 2045/3000\n",
" - 39s - loss: 0.2297 - acc: 0.9731 - val_loss: 0.9520 - val_acc: 0.8568\n",
"\n",
"Epoch 02045: val_acc did not improve from 0.86765\n",
"Epoch 2046/3000\n",
" - 39s - loss: 0.2403 - acc: 0.9728 - val_loss: 0.9474 - val_acc: 0.8571\n",
"\n",
"Epoch 02046: val_acc did not improve from 0.86765\n",
"Epoch 2047/3000\n",
" - 39s - loss: 0.2361 - acc: 0.9730 - val_loss: 0.9686 - val_acc: 0.8533\n",
"\n",
"Epoch 02047: val_acc did not improve from 0.86765\n",
"Epoch 2048/3000\n",
" - 39s - loss: 0.2373 - acc: 0.9740 - val_loss: 0.9266 - val_acc: 0.8595\n",
"\n",
"Epoch 02048: val_acc did not improve from 0.86765\n",
"Epoch 2049/3000\n",
" - 39s - loss: 0.2426 - acc: 0.9719 - val_loss: 0.9389 - val_acc: 0.8556\n",
"\n",
"Epoch 02049: val_acc did not improve from 0.86765\n",
"Epoch 2050/3000\n",
" - 39s - loss: 0.2379 - acc: 0.9725 - val_loss: 0.9371 - val_acc: 0.8583\n",
"\n",
"Epoch 02050: val_acc did not improve from 0.86765\n",
"Epoch 2051/3000\n",
" - 39s - loss: 0.2403 - acc: 0.9701 - val_loss: 0.9588 - val_acc: 0.8548\n",
"\n",
"Epoch 02051: val_acc did not improve from 0.86765\n",
"Epoch 2052/3000\n",
" - 39s - loss: 0.2376 - acc: 0.9722 - val_loss: 0.9510 - val_acc: 0.8564\n",
"\n",
"Epoch 02052: val_acc did not improve from 0.86765\n",
"Epoch 2053/3000\n",
" - 39s - loss: 0.2430 - acc: 0.9695 - val_loss: 0.9629 - val_acc: 0.8552\n",
"\n",
"Epoch 02053: val_acc did not improve from 0.86765\n",
"Epoch 2054/3000\n",
" - 39s - loss: 0.2442 - acc: 0.9691 - val_loss: 0.9616 - val_acc: 0.8552\n",
"\n",
"Epoch 02054: val_acc did not improve from 0.86765\n",
"Epoch 2055/3000\n",
" - 39s - loss: 0.2328 - acc: 0.9730 - val_loss: 0.9661 - val_acc: 0.8521\n",
"\n",
"Epoch 02055: val_acc did not improve from 0.86765\n",
"Epoch 2056/3000\n",
" - 39s - loss: 0.2396 - acc: 0.9728 - val_loss: 0.9707 - val_acc: 0.8540\n",
"\n",
"Epoch 02056: val_acc did not improve from 0.86765\n",
"Epoch 2057/3000\n",
" - 39s - loss: 0.2385 - acc: 0.9701 - val_loss: 0.9474 - val_acc: 0.8571\n",
"\n",
"Epoch 02057: val_acc did not improve from 0.86765\n",
"Epoch 2058/3000\n",
" - 39s - loss: 0.2517 - acc: 0.9683 - val_loss: 0.9756 - val_acc: 0.8533\n",
"\n",
"Epoch 02058: val_acc did not improve from 0.86765\n",
"Epoch 2059/3000\n",
" - 39s - loss: 0.2406 - acc: 0.9703 - val_loss: 0.9670 - val_acc: 0.8544\n",
"\n",
"Epoch 02059: val_acc did not improve from 0.86765\n",
"Epoch 2060/3000\n",
" - 39s - loss: 0.2333 - acc: 0.9727 - val_loss: 0.9609 - val_acc: 0.8540\n",
"\n",
"Epoch 02060: val_acc did not improve from 0.86765\n",
"Epoch 2061/3000\n",
" - 39s - loss: 0.2347 - acc: 0.9734 - val_loss: 0.9288 - val_acc: 0.8560\n",
"\n",
"Epoch 02061: val_acc did not improve from 0.86765\n",
"Epoch 2062/3000\n",
" - 39s - loss: 0.2367 - acc: 0.9709 - val_loss: 0.9407 - val_acc: 0.8548\n",
"\n",
"Epoch 02062: val_acc did not improve from 0.86765\n",
"Epoch 2063/3000\n",
" - 39s - loss: 0.2401 - acc: 0.9695 - val_loss: 0.9461 - val_acc: 0.8552\n",
"\n",
"Epoch 02063: val_acc did not improve from 0.86765\n",
"Epoch 2064/3000\n",
" - 39s - loss: 0.2355 - acc: 0.9721 - val_loss: 0.9359 - val_acc: 0.8575\n",
"\n",
"Epoch 02064: val_acc did not improve from 0.86765\n",
"Epoch 2065/3000\n",
" - 39s - loss: 0.2380 - acc: 0.9740 - val_loss: 0.9771 - val_acc: 0.8560\n",
"\n",
"Epoch 02065: val_acc did not improve from 0.86765\n",
"Epoch 2066/3000\n",
" - 39s - loss: 0.2330 - acc: 0.9709 - val_loss: 0.9569 - val_acc: 0.8571\n",
"\n",
"Epoch 02066: val_acc did not improve from 0.86765\n",
"Epoch 2067/3000\n",
" - 39s - loss: 0.2330 - acc: 0.9730 - val_loss: 0.9549 - val_acc: 0.8564\n",
"\n",
"Epoch 02067: val_acc did not improve from 0.86765\n",
"Epoch 2068/3000\n",
" - 39s - loss: 0.2374 - acc: 0.9695 - val_loss: 0.9608 - val_acc: 0.8548\n",
"\n",
"Epoch 02068: val_acc did not improve from 0.86765\n",
"Epoch 2069/3000\n",
" - 39s - loss: 0.2412 - acc: 0.9709 - val_loss: 0.9994 - val_acc: 0.8525\n",
"\n",
"Epoch 02069: val_acc did not improve from 0.86765\n",
"Epoch 2070/3000\n",
" - 39s - loss: 0.2426 - acc: 0.9721 - val_loss: 0.9681 - val_acc: 0.8529\n",
"\n",
"Epoch 02070: val_acc did not improve from 0.86765\n",
"Epoch 2071/3000\n",
" - 39s - loss: 0.2438 - acc: 0.9709 - val_loss: 0.9335 - val_acc: 0.8595\n",
"\n",
"Epoch 02071: val_acc did not improve from 0.86765\n",
"Epoch 2072/3000\n",
" - 39s - loss: 0.2370 - acc: 0.9706 - val_loss: 0.9464 - val_acc: 0.8583\n",
"\n",
"Epoch 02072: val_acc did not improve from 0.86765\n",
"Epoch 2073/3000\n",
" - 39s - loss: 0.2379 - acc: 0.9716 - val_loss: 0.9395 - val_acc: 0.8610\n",
"\n",
"Epoch 02073: val_acc did not improve from 0.86765\n",
"Epoch 2074/3000\n",
" - 39s - loss: 0.2335 - acc: 0.9725 - val_loss: 0.9276 - val_acc: 0.8579\n",
"\n",
"Epoch 02074: val_acc did not improve from 0.86765\n",
"Epoch 2075/3000\n",
" - 39s - loss: 0.2359 - acc: 0.9707 - val_loss: 0.9434 - val_acc: 0.8583\n",
"\n",
"Epoch 02075: val_acc did not improve from 0.86765\n",
"Epoch 2076/3000\n",
" - 39s - loss: 0.2351 - acc: 0.9725 - val_loss: 0.9271 - val_acc: 0.8614\n",
"\n",
"Epoch 02076: val_acc did not improve from 0.86765\n",
"Epoch 2077/3000\n",
" - 39s - loss: 0.2339 - acc: 0.9739 - val_loss: 0.9493 - val_acc: 0.8560\n",
"\n",
"Epoch 02077: val_acc did not improve from 0.86765\n",
"Epoch 2078/3000\n",
" - 39s - loss: 0.2408 - acc: 0.9713 - val_loss: 0.9289 - val_acc: 0.8595\n",
"\n",
"Epoch 02078: val_acc did not improve from 0.86765\n",
"Epoch 2079/3000\n",
" - 39s - loss: 0.2311 - acc: 0.9731 - val_loss: 0.9743 - val_acc: 0.8540\n",
"\n",
"Epoch 02079: val_acc did not improve from 0.86765\n",
"Epoch 2080/3000\n",
" - 39s - loss: 0.2391 - acc: 0.9712 - val_loss: 0.9787 - val_acc: 0.8552\n",
"\n",
"Epoch 02080: val_acc did not improve from 0.86765\n",
"Epoch 2081/3000\n",
" - 39s - loss: 0.2416 - acc: 0.9721 - val_loss: 0.9481 - val_acc: 0.8591\n",
"\n",
"Epoch 02081: val_acc did not improve from 0.86765\n",
"Epoch 2082/3000\n",
" - 39s - loss: 0.2349 - acc: 0.9746 - val_loss: 0.9670 - val_acc: 0.8556\n",
"\n",
"Epoch 02082: val_acc did not improve from 0.86765\n",
"Epoch 2083/3000\n",
" - 39s - loss: 0.2428 - acc: 0.9700 - val_loss: 0.9585 - val_acc: 0.8525\n",
"\n",
"Epoch 02083: val_acc did not improve from 0.86765\n",
"Epoch 2084/3000\n",
" - 39s - loss: 0.2419 - acc: 0.9707 - val_loss: 0.9589 - val_acc: 0.8533\n",
"\n",
"Epoch 02084: val_acc did not improve from 0.86765\n",
"Epoch 2085/3000\n",
" - 39s - loss: 0.2325 - acc: 0.9718 - val_loss: 0.9546 - val_acc: 0.8583\n",
"\n",
"Epoch 02085: val_acc did not improve from 0.86765\n",
"Epoch 2086/3000\n",
" - 39s - loss: 0.2456 - acc: 0.9695 - val_loss: 0.9727 - val_acc: 0.8533\n",
"\n",
"Epoch 02086: val_acc did not improve from 0.86765\n",
"Epoch 2087/3000\n",
" - 39s - loss: 0.2377 - acc: 0.9725 - val_loss: 0.9415 - val_acc: 0.8591\n",
"\n",
"Epoch 02087: val_acc did not improve from 0.86765\n",
"Epoch 2088/3000\n",
" - 39s - loss: 0.2340 - acc: 0.9730 - val_loss: 0.9524 - val_acc: 0.8587\n",
"\n",
"Epoch 02088: val_acc did not improve from 0.86765\n",
"Epoch 2089/3000\n",
" - 40s - loss: 0.2349 - acc: 0.9739 - val_loss: 0.9492 - val_acc: 0.8575\n",
"\n",
"Epoch 02089: val_acc did not improve from 0.86765\n",
"Epoch 2090/3000\n",
" - 39s - loss: 0.2492 - acc: 0.9692 - val_loss: 0.9358 - val_acc: 0.8606\n",
"\n",
"Epoch 02090: val_acc did not improve from 0.86765\n",
"Epoch 2091/3000\n",
" - 39s - loss: 0.2351 - acc: 0.9716 - val_loss: 0.9357 - val_acc: 0.8556\n",
"\n",
"Epoch 02091: val_acc did not improve from 0.86765\n",
"Epoch 2092/3000\n",
" - 39s - loss: 0.2311 - acc: 0.9728 - val_loss: 0.9582 - val_acc: 0.8568\n",
"\n",
"Epoch 02092: val_acc did not improve from 0.86765\n",
"Epoch 2093/3000\n",
" - 40s - loss: 0.2418 - acc: 0.9727 - val_loss: 0.9491 - val_acc: 0.8587\n",
"\n",
"Epoch 02093: val_acc did not improve from 0.86765\n",
"Epoch 2094/3000\n",
" - 39s - loss: 0.2311 - acc: 0.9751 - val_loss: 0.9501 - val_acc: 0.8579\n",
"\n",
"Epoch 02094: val_acc did not improve from 0.86765\n",
"Epoch 2095/3000\n",
" - 39s - loss: 0.2330 - acc: 0.9709 - val_loss: 0.9506 - val_acc: 0.8560\n",
"\n",
"Epoch 02095: val_acc did not improve from 0.86765\n",
"Epoch 2096/3000\n",
" - 39s - loss: 0.2446 - acc: 0.9706 - val_loss: 0.9476 - val_acc: 0.8583\n",
"\n",
"Epoch 02096: val_acc did not improve from 0.86765\n",
"Epoch 2097/3000\n",
" - 39s - loss: 0.2413 - acc: 0.9698 - val_loss: 0.9480 - val_acc: 0.8568\n",
"\n",
"Epoch 02097: val_acc did not improve from 0.86765\n",
"Epoch 2098/3000\n",
" - 39s - loss: 0.2435 - acc: 0.9706 - val_loss: 0.9637 - val_acc: 0.8575\n",
"\n",
"Epoch 02098: val_acc did not improve from 0.86765\n",
"Epoch 2099/3000\n",
" - 38s - loss: 0.2336 - acc: 0.9710 - val_loss: 0.9526 - val_acc: 0.8552\n",
"\n",
"Epoch 02099: val_acc did not improve from 0.86765\n",
"Epoch 2100/3000\n",
" - 39s - loss: 0.2388 - acc: 0.9710 - val_loss: 0.9775 - val_acc: 0.8548\n",
"\n",
"Epoch 02100: val_acc did not improve from 0.86765\n",
"Epoch 2101/3000\n",
" - 39s - loss: 0.2267 - acc: 0.9755 - val_loss: 0.9836 - val_acc: 0.8575\n",
"\n",
"Epoch 02101: val_acc did not improve from 0.86765\n",
"Epoch 2102/3000\n",
" - 39s - loss: 0.2341 - acc: 0.9737 - val_loss: 0.9518 - val_acc: 0.8591\n",
"\n",
"Epoch 02102: val_acc did not improve from 0.86765\n",
"Epoch 2103/3000\n",
" - 39s - loss: 0.2455 - acc: 0.9709 - val_loss: 0.9928 - val_acc: 0.8505\n",
"\n",
"Epoch 02103: val_acc did not improve from 0.86765\n",
"Epoch 2104/3000\n",
" - 39s - loss: 0.2408 - acc: 0.9713 - val_loss: 0.9514 - val_acc: 0.8556\n",
"\n",
"Epoch 02104: val_acc did not improve from 0.86765\n",
"Epoch 2105/3000\n",
" - 39s - loss: 0.2374 - acc: 0.9727 - val_loss: 0.9577 - val_acc: 0.8556\n",
"\n",
"Epoch 02105: val_acc did not improve from 0.86765\n",
"Epoch 2106/3000\n",
" - 39s - loss: 0.2509 - acc: 0.9694 - val_loss: 0.9963 - val_acc: 0.8544\n",
"\n",
"Epoch 02106: val_acc did not improve from 0.86765\n",
"Epoch 2107/3000\n",
" - 39s - loss: 0.2378 - acc: 0.9739 - val_loss: 0.9823 - val_acc: 0.8540\n",
"\n",
"Epoch 02107: val_acc did not improve from 0.86765\n",
"Epoch 2108/3000\n",
" - 39s - loss: 0.2395 - acc: 0.9722 - val_loss: 0.9448 - val_acc: 0.8564\n",
"\n",
"Epoch 02108: val_acc did not improve from 0.86765\n",
"Epoch 2109/3000\n",
" - 39s - loss: 0.2436 - acc: 0.9713 - val_loss: 0.9368 - val_acc: 0.8583\n",
"\n",
"Epoch 02109: val_acc did not improve from 0.86765\n",
"Epoch 2110/3000\n",
" - 39s - loss: 0.2356 - acc: 0.9727 - val_loss: 0.9681 - val_acc: 0.8571\n",
"\n",
"Epoch 02110: val_acc did not improve from 0.86765\n",
"Epoch 2111/3000\n",
" - 39s - loss: 0.2287 - acc: 0.9752 - val_loss: 0.9240 - val_acc: 0.8599\n",
"\n",
"Epoch 02111: val_acc did not improve from 0.86765\n",
"Epoch 2112/3000\n",
" - 39s - loss: 0.2469 - acc: 0.9695 - val_loss: 0.9354 - val_acc: 0.8591\n",
"\n",
"Epoch 02112: val_acc did not improve from 0.86765\n",
"Epoch 2113/3000\n",
" - 39s - loss: 0.2347 - acc: 0.9736 - val_loss: 0.9155 - val_acc: 0.8638\n",
"\n",
"Epoch 02113: val_acc did not improve from 0.86765\n",
"Epoch 2114/3000\n",
" - 39s - loss: 0.2462 - acc: 0.9680 - val_loss: 0.9165 - val_acc: 0.8579\n",
"\n",
"Epoch 02114: val_acc did not improve from 0.86765\n",
"Epoch 2115/3000\n",
" - 39s - loss: 0.2421 - acc: 0.9722 - val_loss: 0.9211 - val_acc: 0.8591\n",
"\n",
"Epoch 02115: val_acc did not improve from 0.86765\n",
"Epoch 2116/3000\n",
" - 39s - loss: 0.2292 - acc: 0.9728 - val_loss: 0.9327 - val_acc: 0.8575\n",
"\n",
"Epoch 02116: val_acc did not improve from 0.86765\n",
"Epoch 2117/3000\n",
" - 39s - loss: 0.2303 - acc: 0.9725 - val_loss: 0.9244 - val_acc: 0.8583\n",
"\n",
"Epoch 02117: val_acc did not improve from 0.86765\n",
"Epoch 2118/3000\n",
" - 39s - loss: 0.2364 - acc: 0.9727 - val_loss: 0.9234 - val_acc: 0.8634\n",
"\n",
"Epoch 02118: val_acc did not improve from 0.86765\n",
"Epoch 2119/3000\n",
" - 39s - loss: 0.2371 - acc: 0.9725 - val_loss: 0.9618 - val_acc: 0.8560\n",
"\n",
"Epoch 02119: val_acc did not improve from 0.86765\n",
"Epoch 2120/3000\n",
" - 39s - loss: 0.2456 - acc: 0.9689 - val_loss: 0.9649 - val_acc: 0.8529\n",
"\n",
"Epoch 02120: val_acc did not improve from 0.86765\n",
"Epoch 2121/3000\n",
" - 39s - loss: 0.2364 - acc: 0.9743 - val_loss: 0.9671 - val_acc: 0.8575\n",
"\n",
"Epoch 02121: val_acc did not improve from 0.86765\n",
"Epoch 2122/3000\n",
" - 39s - loss: 0.2430 - acc: 0.9695 - val_loss: 0.9523 - val_acc: 0.8568\n",
"\n",
"Epoch 02122: val_acc did not improve from 0.86765\n",
"Epoch 2123/3000\n",
" - 39s - loss: 0.2408 - acc: 0.9707 - val_loss: 0.9457 - val_acc: 0.8579\n",
"\n",
"Epoch 02123: val_acc did not improve from 0.86765\n",
"Epoch 2124/3000\n",
" - 39s - loss: 0.2469 - acc: 0.9706 - val_loss: 0.9600 - val_acc: 0.8583\n",
"\n",
"Epoch 02124: val_acc did not improve from 0.86765\n",
"Epoch 2125/3000\n",
" - 39s - loss: 0.2411 - acc: 0.9709 - val_loss: 0.9726 - val_acc: 0.8575\n",
"\n",
"Epoch 02125: val_acc did not improve from 0.86765\n",
"Epoch 2126/3000\n",
" - 39s - loss: 0.2395 - acc: 0.9721 - val_loss: 0.9435 - val_acc: 0.8575\n",
"\n",
"Epoch 02126: val_acc did not improve from 0.86765\n",
"Epoch 2127/3000\n",
" - 39s - loss: 0.2358 - acc: 0.9725 - val_loss: 0.9640 - val_acc: 0.8536\n",
"\n",
"Epoch 02127: val_acc did not improve from 0.86765\n",
"Epoch 2128/3000\n",
" - 39s - loss: 0.2387 - acc: 0.9718 - val_loss: 0.9418 - val_acc: 0.8587\n",
"\n",
"Epoch 02128: val_acc did not improve from 0.86765\n",
"Epoch 2129/3000\n",
" - 39s - loss: 0.2473 - acc: 0.9689 - val_loss: 0.9574 - val_acc: 0.8560\n",
"\n",
"Epoch 02129: val_acc did not improve from 0.86765\n",
"Epoch 2130/3000\n",
" - 39s - loss: 0.2325 - acc: 0.9707 - val_loss: 0.9537 - val_acc: 0.8591\n",
"\n",
"Epoch 02130: val_acc did not improve from 0.86765\n",
"Epoch 2131/3000\n",
" - 39s - loss: 0.2376 - acc: 0.9737 - val_loss: 0.9685 - val_acc: 0.8548\n",
"\n",
"Epoch 02131: val_acc did not improve from 0.86765\n",
"Epoch 2132/3000\n",
" - 39s - loss: 0.2299 - acc: 0.9730 - val_loss: 0.9493 - val_acc: 0.8575\n",
"\n",
"Epoch 02132: val_acc did not improve from 0.86765\n",
"Epoch 2133/3000\n",
" - 39s - loss: 0.2387 - acc: 0.9728 - val_loss: 0.9621 - val_acc: 0.8556\n",
"\n",
"Epoch 02133: val_acc did not improve from 0.86765\n",
"Epoch 2134/3000\n",
" - 39s - loss: 0.2414 - acc: 0.9719 - val_loss: 0.9472 - val_acc: 0.8564\n",