predictgold.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "# 使用PaddlePaddle完成黄金预测\n",
    "\n",
    "如果能预测黄金价格走势曲线，就可以在炒黄金时赚钱。\n",
    "\n",
    "国内黄金价格和国际黄金价格曲线基本一致。\n",
    "\n",
    "# 本模型效果图\n",
    "![](https://ai-studio-static-online.cdn.bcebos.com/fa98f86108af4edb928dd5fe67b0c359b55258ac05334b458bf3fab56209b69e)\n",
    "\n",
    "# 目标\n",
    "只需预测未来黄金价格曲线图，因为只需知道价格是上升还是下降，就可以决定现在是买入、卖出、还是持仓。\n",
    "不用将价格预测准确，只需要将趋势图预测准确就可以。\n",
    "\n",
    "# 贡献方法\n",
    "请大家把更新后的代码导出后，（点击右上角：文件-》导出为ipynb）\n",
    "pull request到我github的这个文件:\n",
    "[https://github.com/guojiahuiEmily/predictgoldprice/blob/main/predictgold.ipynb](https://github.com/guojiahuiEmily/predictgoldprice/blob/main/predictgold.ipynb)\n",
    "\n",
    "\n",
    "# 分享成绩\n",
    "并把自己的预测效果及截图分享到这里，看看谁的效果好：\n",
    "\n",
    "[https://github.com/guojiahuiEmily/predictgoldprice/blob/main/rank](https://github.com/guojiahuiEmily/predictgoldprice/blob/main/rank)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "# 所使用的的模型\n",
    "\n",
    "# 时间卷积网络（TCN，Temporal Convolutional Networks）\n",
    "时间序列是指按照时间先后顺序排列而成的序列，例如每日发电量、每小时营业额等组成的序列。通过分析时间序列中的发展过程、方向和趋势，我们可以预测下一段时间可能出现的情况。在本例中，我们使用时间卷积网络TCN进行建模，将学习到的特征接入全连接层完成预测。TCN的网络如下所示：\n",
    "\n",
    "<center><img src=\"https://ai-studio-static-online.cdn.bcebos.com/214011346cb24535ad57c50b936a2e4a207da6f7ce3b4e4babbe2571003f1aee\" width=\"500\" height=\"300\" ></center>\n",
    "<br><center>图1：TCN示意图</center></br>\n",
    "\n",
    "图中是一个filters number=3, dilated rate=1的时间卷积网络，它能够学习前T个时序的数据特征。关于TCN更详细的资料请参考论文：[An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling](https://arxiv.org/pdf/1803.01271.pdf)。\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "# 代码\n",
    "\n",
    "## 准备环境\n",
    "\n",
    "我们首先需要导入必要的包。\n",
    "\n",
    "这里我们使用`paddlenlp.seq2vec`中内置好的模型，关于`seq2vec`的详细介绍可参考这个项目：[seq2vec是什么? ](https://aistudio.baidu.com/aistudio/projectdetail/1283423)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "!pip install paddlenlp>=2.0.0b -i https://pypi.org/simple"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "如果上面运行失败可执行下面"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# !pip install --upgrade setuptools && python -m pip install --upgrade pip\n",
    "# !pip install paddlenlp==2.0.0rc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "import os\n",
    "import sys\n",
    "\n",
    "import paddle\n",
    "import paddle.nn as nn\n",
    "import numpy as np\n",
    "\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "from pylab import rcParams\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib import rc\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from pandas.plotting import register_matplotlib_converters\n",
    "\n",
    "sys.path.append(os.path.abspath(os.path.join(os.getcwd(), \"../..\")))\n",
    "from paddlenlp.seq2vec import TCNEncoder"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "为了更好地展示数据结果，我们在这里配置画图功能。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# config matplotlib\n",
    "%matplotlib inline\n",
    "%config InlineBackend.figure_format='retina'\n",
    "sns.set(style='whitegrid', palette='muted', font_scale=1.2)\n",
    "HAPPY_COLORS_PALETTE = [\"#01BEFE\", \"#FFDD00\", \"#FF7D00\", \"#FF006D\", \"#93D30C\", \"#8F00FF\"]\n",
    "sns.set_palette(sns.color_palette(HAPPY_COLORS_PALETTE))\n",
    "rcParams['figure.figsize'] = 14, 10\n",
    "register_matplotlib_converters()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "## 数据\n",
    "\n",
    "在csv文件夹内。您也可以点击左侧“数据集”，自己划分这3个文件。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "## 数据预览\n",
    "\n",
    "数据集中包含了国际每日黄金价格，单位：美元。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>days</th>\n",
       "      <th>price</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1978/12/29</td>\n",
       "      <td>226</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1979/1/1</td>\n",
       "      <td>226</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1979/1/2</td>\n",
       "      <td>227</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1979/1/3</td>\n",
       "      <td>219</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1979/1/4</td>\n",
       "      <td>223</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         days  price\n",
       "0  1978/12/29    226\n",
       "1    1979/1/1    226\n",
       "2    1979/1/2    227\n",
       "3    1979/1/3    219\n",
       "4    1979/1/4    223"
      ]
     },
     "execution_count": null,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_all = pd.read_csv('csv/all.csv')\n",
    "df_all.head()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "我们将对价格曲线走势进行预测，不用预测准确价格。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>price</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>days</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1978-12-29</th>\n",
       "      <td>226</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1979-01-01</th>\n",
       "      <td>226</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1979-01-02</th>\n",
       "      <td>227</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1979-01-03</th>\n",
       "      <td>219</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1979-01-04</th>\n",
       "      <td>223</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            price\n",
       "days             \n",
       "1978-12-29    226\n",
       "1979-01-01    226\n",
       "1979-01-02    227\n",
       "1979-01-03    219\n",
       "1979-01-04    223"
      ]
     },
     "execution_count": null,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "daily_cases = pd.read_csv('csv/all.csv',index_col=0,parse_dates=True, header=0,dtype= {'price': np.int32})\n",
    "daily_cases.head()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/matplotlib/cbook/__init__.py:2349: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working\n",
      "  if isinstance(obj, collections.Iterator):\n",
      "/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/matplotlib/cbook/__init__.py:2366: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working\n",
      "  return list(data) if isinstance(data, collections.MappingView) else data\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(6,6))\n",
    "plt.plot(daily_cases)\n",
    "plt.title(\"daily\");"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "## 数据预处理\n",
    "\n",
    "首先划分数据集为训练集与验证集，我们取最后30天的数据作为测试集，其余作为训练集。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The number of the samples in train set is : 10951\n"
     ]
    }
   ],
   "source": [
    "TEST_DATA_SIZE = 30\n",
    "\n",
    "train_data = daily_cases[:-TEST_DATA_SIZE].astype('float32')\n",
    "test_data = daily_cases[-TEST_DATA_SIZE:].astype('float32')\n",
    "\n",
    "# train_data = train_data.astype('float32')\n",
    "\n",
    "print(\"The number of the samples in train set is : %i\"%train_data.shape[0])\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "为了提升模型收敛速度与性能，我们使用scikit-learn进行数据归一化。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The number of the samples in train set is : 10951\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>price</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>days</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1978-12-29</th>\n",
       "      <td>226.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1979-01-01</th>\n",
       "      <td>226.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1979-01-02</th>\n",
       "      <td>227.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1979-01-03</th>\n",
       "      <td>219.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1979-01-04</th>\n",
       "      <td>223.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            price\n",
       "days             \n",
       "1978-12-29  226.0\n",
       "1979-01-01  226.0\n",
       "1979-01-02  227.0\n",
       "1979-01-03  219.0\n",
       "1979-01-04  223.0"
      ]
     },
     "execution_count": null,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "print(\"The number of the samples in train set is : %i\"%train_data.shape[0])\n",
    "train_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>price</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>days</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2020-12-21</th>\n",
       "      <td>1880.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-12-22</th>\n",
       "      <td>1877.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-12-23</th>\n",
       "      <td>1875.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-12-24</th>\n",
       "      <td>1875.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-12-25</th>\n",
       "      <td>1875.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             price\n",
       "days              \n",
       "2020-12-21  1880.0\n",
       "2020-12-22  1877.0\n",
       "2020-12-23  1875.0\n",
       "2020-12-24  1875.0\n",
       "2020-12-25  1875.0"
      ]
     },
     "execution_count": null,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "test_data.head()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "# print(f\"bef train_data:{train_data[0]}\")\n",
    "# print(f\"bef test_data:{test_data[0]}\")\n",
    "\n",
    "scaler = MinMaxScaler()\n",
    "train_data = scaler.fit_transform(train_data)\n",
    "test_data = scaler.fit_transform(test_data)\n",
    "# train_data[0]\n",
    "# test_data[0]\n",
    "\n",
    "# print(f\"after train_data:{train_data[0]}\")\n",
    "# print(f\"after test_data:{test_data[0]}\")\n",
    "\n",
    "# train_data = scaler.fit_transform(np.expand_dims(train_data, axis=1)).astype('float32')\n",
    "# test_data = scaler.transform(np.expand_dims(test_data, axis=1)).astype('float32')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "现在开始组建时间序列"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The shape of x_train is: (10922, 29, 1)\n",
      "The shape of y_train is: (10922, 1)\n",
      "The shape of x_test is: (30, 29, 1)\n",
      "The shape of y_test is: (30, 1)\n"
     ]
    }
   ],
   "source": [
    "SEQ_LEN = 30\n",
    "\n",
    "def create_sequences(data, seq_length):\n",
    "    xs = []\n",
    "    ys = []\n",
    "\n",
    "    for i in range(len(data)-seq_length+1):\n",
    "        x = data[i:i+seq_length-1]\n",
    "        y = data[i+seq_length-1]\n",
    "        xs.append(x)\n",
    "        ys.append(y)\n",
    "\n",
    "    return np.array(xs), np.array(ys)\n",
    "\n",
    "x_train, y_train = create_sequences(train_data, SEQ_LEN)\n",
    "test_data = np.concatenate((train_data[-SEQ_LEN+1:],test_data),axis=0)\n",
    "x_test, y_test = create_sequences(test_data, SEQ_LEN)\n",
    "\n",
    "print(\"The shape of x_train is: %s\"%str(x_train.shape))\n",
    "print(\"The shape of y_train is: %s\"%str(y_train.shape))\n",
    "print(\"The shape of x_test is: %s\"%str(x_test.shape))\n",
    "print(\"The shape of y_test is: %s\"%str(y_test.shape))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "数据集处理完毕，将数据集封装到CovidDataset，以便模型训练、预测时调用。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "class CovidDataset(paddle.io.Dataset):\n",
    "    def __init__(self, feature, label):\n",
    "        self.feature = feature\n",
    "        self.label = label\n",
    "        super(CovidDataset, self).__init__()\n",
    "\n",
    "    def __len__(self):\n",
    "        return len(self.label)\n",
    "\n",
    "    def __getitem__(self, index):\n",
    "        return [self.feature[index], self.label[index]]\n",
    "\n",
    "train_dataset = CovidDataset(x_train, y_train)\n",
    "test_dataset = CovidDataset(x_test, y_test)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "## 组网\n",
    "\n",
    "现在开始组建模型网络，我们采用时间卷积网络TCN作为特征提取器，将提取到的时序信息传送给全连接层获得最终的预测结果。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "class TimeSeriesNetwork(nn.Layer):\n",
    "\n",
    "  def __init__(self, input_size, next_k=1, num_channels=[64,128,256]):\n",
    "    super(TimeSeriesNetwork, self).__init__()\n",
    "\n",
    "    self.last_num_channel = num_channels[-1]\n",
    "\n",
    "    self.tcn = TCNEncoder(\n",
    "      input_size=input_size,\n",
    "      num_channels=num_channels,\n",
    "      kernel_size=2, \n",
    "      dropout=0.2\n",
    "    )\n",
    "\n",
    "    self.linear = nn.Linear(in_features= self.last_num_channel, out_features=next_k)\n",
    "\n",
    "  def forward(self, x):\n",
    "    tcn_out = self.tcn(x)\n",
    "    y_pred = self.linear(tcn_out)\n",
    "    return y_pred\n",
    "\n",
    "network = TimeSeriesNetwork(input_size=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "## 定义优化器、损失函数\n",
    "\n",
    "在这里我们使用Adam优化器、均方差损失函数，为启动训练做最后的准备。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "LR = 1e-3\n",
    "\n",
    "model = paddle.Model(network)\n",
    "\n",
    "optimizer = paddle.optimizer.Adam(\n",
    "        learning_rate=LR, parameters=model.parameters())\n",
    "\n",
    "loss = paddle.nn.MSELoss(reduction='sum')\n",
    "\n",
    "model.prepare(optimizer, loss)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "## 训练\n",
    "\n",
    "配置必要的超参数，启动训练。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The loss value printed in the log is the current step, and the metric is the average value of previous step.\n",
      "Epoch 1/100\n",
      "step 341/341 [==============================] - loss: 0.0179 - ETA: 6s - 20ms/st - loss: 0.0214 - ETA: 5s - 18ms/st - loss: 0.0175 - ETA: 5s - 17ms/st - loss: 0.0270 - ETA: 5s - 17ms/st - loss: 0.0078 - ETA: 4s - 17ms/st - loss: 0.0151 - ETA: 4s - 16ms/st - loss: 0.0599 - ETA: 4s - 16ms/st - loss: 0.0106 - ETA: 4s - 16ms/st - loss: 0.0171 - ETA: 4s - 16ms/st - loss: 0.0264 - ETA: 3s - 16ms/st - loss: 0.0189 - ETA: 3s - 16ms/st - loss: 0.0147 - ETA: 3s - 16ms/st - loss: 0.0106 - ETA: 3s - 16ms/st - loss: 0.0335 - ETA: 3s - 16ms/st - loss: 0.0085 - ETA: 3s - 16ms/st - loss: 0.0222 - ETA: 2s - 16ms/st - loss: 0.0153 - ETA: 2s - 16ms/st - loss: 0.0237 - ETA: 2s - 16ms/st - loss: 0.0290 - ETA: 2s - 16ms/st - loss: 0.0265 - ETA: 2s - 16ms/st - loss: 0.0136 - ETA: 2s - 16ms/st - loss: 0.0310 - ETA: 1s - 16ms/st - loss: 0.0176 - ETA: 1s - 16ms/st - loss: 0.0595 - ETA: 1s - 16ms/st - loss: 0.0151 - ETA: 1s - 16ms/st - loss: 0.0280 - ETA: 1s - 16ms/st - loss: 0.0106 - ETA: 1s - 16ms/st - loss: 0.0373 - ETA: 0s - 16ms/st - loss: 0.0063 - ETA: 0s - 16ms/st - loss: 0.0357 - ETA: 0s - 16ms/st - loss: 0.0109 - ETA: 0s - 16ms/st - loss: 0.0442 - ETA: 0s - 16ms/st - loss: 0.0337 - ETA: 0s - 16ms/st - loss: 0.0134 - ETA: 0s - 16ms/st - loss: 0.0142 - 16ms/step          \n",
      "save checkpoint at /home/aistudio/save_dir/0\n",
      "Epoch 2/100\n",
      "step 341/341 [==============================] - loss: 0.0179 - ETA: 5s - 16ms/st - loss: 0.0096 - ETA: 4s - 15ms/st - loss: 0.0084 - ETA: 4s - 15ms/st - loss: 0.0153 - ETA: 4s - 15ms/st - loss: 0.0252 - ETA: 4s - 15ms/st - loss: 0.0328 - ETA: 4s - 15ms/st - loss: 0.0347 - ETA: 4s - 15ms/st - loss: 0.0457 - ETA: 4s - 15ms/st - loss: 0.0152 - ETA: 3s - 15ms/st - loss: 0.0149 - ETA: 3s - 15ms/st - loss: 0.0181 - ETA: 3s - 15ms/st - loss: 0.0133 - ETA: 3s - 15ms/st - loss: 0.0184 - ETA: 3s - 15ms/st - loss: 0.0085 - ETA: 3s - 15ms/st - loss: 0.0279 - ETA: 2s - 16ms/st - loss: 0.0157 - ETA: 2s - 16ms/st - loss: 0.0257 - ETA: 2s - 16ms/st - loss: 0.0132 - ETA: 2s - 16ms/st - loss: 0.0169 - ETA: 2s - 16ms/st - loss: 0.0128 - ETA: 2s - 16ms/st - loss: 0.0098 - ETA: 2s - 16ms/st - loss: 0.0106 - ETA: 1s - 16ms/st - loss: 0.0129 - ETA: 1s - 16ms/st - loss: 0.0094 - ETA: 1s - 16ms/st - loss: 0.0424 - ETA: 1s - 16ms/st - loss: 0.0059 - ETA: 1s - 16ms/st - loss: 0.0255 - ETA: 1s - 16ms/st - loss: 0.0167 - ETA: 0s - 16ms/st - loss: 0.0088 - ETA: 0s - 16ms/st - loss: 0.0165 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/st - loss: 0.0209 - ETA: 0s - 16ms/st - loss: 0.0086 - ETA: 0s - 16ms/st - loss: 0.0104 - ETA: 0s - 16ms/st - loss: 0.0270 - 16ms/step          \n",
      "Epoch 3/100\n",
      "step 341/341 [==============================] - loss: 0.0145 - ETA: 5s - 16ms/st - loss: 0.0160 - ETA: 5s - 16ms/st - loss: 0.0270 - ETA: 4s - 16ms/st - loss: 0.0252 - ETA: 4s - 16ms/st - loss: 0.0151 - ETA: 4s - 16ms/st - loss: 0.0085 - ETA: 4s - 16ms/st - loss: 0.0195 - ETA: 4s - 16ms/st - loss: 0.0152 - ETA: 4s - 16ms/st - loss: 0.0173 - ETA: 3s - 16ms/st - loss: 0.0141 - ETA: 3s - 16ms/st - loss: 0.0151 - ETA: 3s - 16ms/st - loss: 0.0129 - ETA: 3s - 16ms/st - loss: 0.0247 - ETA: 3s - 16ms/st - loss: 0.0132 - ETA: 3s - 16ms/st - loss: 0.0369 - ETA: 2s - 16ms/st - loss: 0.0173 - ETA: 2s - 16ms/st - loss: 0.0112 - ETA: 2s - 16ms/st - loss: 0.0366 - ETA: 2s - 16ms/st - loss: 0.0045 - ETA: 2s - 16ms/st - loss: 0.0271 - ETA: 2s - 16ms/st - loss: 0.0345 - ETA: 2s - 16ms/st - loss: 0.0294 - ETA: 1s - 16ms/st - loss: 0.0165 - ETA: 1s - 16ms/st - loss: 0.0202 - ETA: 1s - 16ms/st - loss: 0.0132 - ETA: 1s - 16ms/st - loss: 0.0061 - ETA: 1s - 16ms/st - loss: 0.0138 - ETA: 1s - 16ms/st - loss: 0.0077 - ETA: 0s - 16ms/st - loss: 0.0078 - ETA: 0s - 16ms/st - loss: 0.0165 - ETA: 0s - 16ms/st - loss: 0.0032 - ETA: 0s - 16ms/st - loss: 0.0077 - ETA: 0s - 16ms/st - loss: 0.0236 - ETA: 0s - 16ms/st - loss: 0.0075 - ETA: 0s - 16ms/st - loss: 0.0085 - 16ms/step          \n",
      "Epoch 4/100\n",
      "step 341/341 [==============================] - loss: 0.0136 - ETA: 5s - 16ms/st - loss: 0.0148 - ETA: 5s - 16ms/st - loss: 0.0161 - ETA: 4s - 16ms/st - loss: 0.0070 - ETA: 4s - 16ms/st - loss: 0.0198 - ETA: 4s - 16ms/st - loss: 0.0151 - ETA: 4s - 16ms/st - loss: 0.0136 - ETA: 4s - 16ms/st - loss: 0.0108 - ETA: 4s - 16ms/st - loss: 0.0049 - ETA: 3s - 16ms/st - loss: 0.0176 - ETA: 3s - 16ms/st - loss: 0.0058 - ETA: 3s - 16ms/st - loss: 0.0190 - ETA: 3s - 16ms/st - loss: 0.0086 - ETA: 3s - 16ms/st - loss: 0.0266 - ETA: 3s - 16ms/st - loss: 0.0153 - ETA: 2s - 16ms/st - loss: 0.0061 - ETA: 2s - 16ms/st - loss: 0.0124 - ETA: 2s - 16ms/st - loss: 0.0129 - ETA: 2s - 16ms/st - loss: 0.0106 - ETA: 2s - 16ms/st - loss: 0.0320 - ETA: 2s - 16ms/st - loss: 0.0230 - ETA: 2s - 16ms/st - loss: 0.0094 - ETA: 1s - 16ms/st - loss: 0.0114 - ETA: 1s - 16ms/st - loss: 0.0104 - ETA: 1s - 16ms/st - loss: 0.0240 - ETA: 1s - 16ms/st - loss: 0.0249 - ETA: 1s - 16ms/st - loss: 0.0162 - ETA: 1s - 16ms/st - loss: 0.0393 - ETA: 0s - 16ms/st - loss: 0.0188 - ETA: 0s - 16ms/st - loss: 0.0133 - ETA: 0s - 16ms/st - loss: 0.0205 - ETA: 0s - 16ms/st - loss: 0.0227 - ETA: 0s - 16ms/st - loss: 0.0150 - ETA: 0s - 16ms/st - loss: 0.0071 - ETA: 0s - 16ms/st - loss: 0.0204 - 16ms/step          \n",
      "Epoch 5/100\n",
      "step 341/341 [==============================] - loss: 0.0284 - ETA: 5s - 16ms/st - loss: 0.0126 - ETA: 5s - 16ms/st - loss: 0.0176 - ETA: 4s - 16ms/st - loss: 0.0236 - ETA: 4s - 16ms/st - loss: 0.0115 - ETA: 4s - 16ms/st - loss: 0.0086 - ETA: 4s - 16ms/st - loss: 0.0240 - ETA: 4s - 16ms/st - loss: 0.0109 - ETA: 4s - 16ms/st - loss: 0.0055 - ETA: 3s - 16ms/st - loss: 0.0283 - ETA: 3s - 16ms/st - loss: 0.0153 - ETA: 3s - 16ms/st - loss: 0.0073 - ETA: 3s - 16ms/st - loss: 0.0126 - ETA: 3s - 16ms/st - loss: 0.0115 - ETA: 3s - 16ms/st - loss: 0.0130 - ETA: 2s - 16ms/st - loss: 0.0057 - ETA: 2s - 16ms/st - loss: 0.0152 - ETA: 2s - 16ms/st - loss: 0.0074 - ETA: 2s - 16ms/st - loss: 0.0218 - ETA: 2s - 16ms/st - loss: 0.0195 - ETA: 2s - 16ms/st - loss: 0.0102 - ETA: 2s - 16ms/st - loss: 0.0172 - ETA: 1s - 16ms/st - loss: 0.0074 - ETA: 1s - 16ms/st - loss: 0.0091 - ETA: 1s - 16ms/st - loss: 0.0125 - ETA: 1s - 16ms/st - loss: 0.0095 - ETA: 1s - 16ms/st - loss: 0.0037 - ETA: 1s - 16ms/st - loss: 0.0041 - ETA: 0s - 16ms/st - loss: 0.0092 - ETA: 0s - 16ms/st - loss: 0.0094 - ETA: 0s - 16ms/st - loss: 0.0216 - ETA: 0s - 16ms/st - loss: 0.0061 - ETA: 0s - 16ms/st - loss: 0.0078 - ETA: 0s - 16ms/st - loss: 0.0177 - ETA: 0s - 16ms/st - loss: 0.0101 - 16ms/step          \n",
      "Epoch 6/100\n",
      "step 341/341 [==============================] - loss: 0.0058 - ETA: 5s - 16ms/st - loss: 0.0052 - ETA: 4s - 16ms/st - loss: 0.0141 - ETA: 4s - 16ms/st - loss: 0.0166 - ETA: 4s - 16ms/st - loss: 0.0169 - ETA: 4s - 16ms/st - loss: 0.0077 - ETA: 4s - 16ms/st - loss: 0.0130 - ETA: 4s - 16ms/st - loss: 0.0125 - ETA: 4s - 16ms/st - loss: 0.0147 - ETA: 3s - 16ms/st - loss: 0.0040 - ETA: 3s - 16ms/st - loss: 0.0124 - ETA: 3s - 16ms/st - loss: 0.0048 - ETA: 3s - 16ms/st - loss: 0.0126 - ETA: 3s - 16ms/st - loss: 0.0106 - ETA: 3s - 16ms/st - loss: 0.0165 - ETA: 2s - 16ms/st - loss: 0.0086 - ETA: 2s - 16ms/st - loss: 0.0080 - ETA: 2s - 16ms/st - loss: 0.0165 - ETA: 2s - 16ms/st - loss: 0.0099 - ETA: 2s - 16ms/st - loss: 0.0061 - ETA: 2s - 16ms/st - loss: 0.0046 - ETA: 2s - 16ms/st - loss: 0.0069 - ETA: 1s - 16ms/st - loss: 0.0068 - ETA: 1s - 16ms/st - loss: 0.0035 - ETA: 1s - 16ms/st - loss: 0.0041 - ETA: 1s - 16ms/st - loss: 0.0056 - ETA: 1s - 16ms/st - loss: 0.0075 - ETA: 1s - 16ms/st - loss: 0.0109 - ETA: 0s - 16ms/st - loss: 0.0062 - ETA: 0s - 16ms/st - loss: 0.0074 - ETA: 0s - 16ms/st - loss: 0.0056 - ETA: 0s - 16ms/st - loss: 0.0090 - ETA: 0s - 16ms/st - loss: 0.0076 - ETA: 0s - 16ms/st - loss: 0.0101 - ETA: 0s - 16ms/st - loss: 0.0118 - 16ms/step          \n",
      "Epoch 7/100\n",
      "step 341/341 [==============================] - loss: 0.0118 - ETA: 5s - 16ms/st - loss: 0.0108 - ETA: 5s - 16ms/st - loss: 0.0107 - ETA: 4s - 16ms/st - loss: 0.0091 - ETA: 4s - 16ms/st - loss: 0.0056 - ETA: 4s - 16ms/st - loss: 0.0080 - ETA: 4s - 16ms/st - loss: 0.0065 - ETA: 4s - 16ms/st - loss: 0.0079 - ETA: 4s - 16ms/st - loss: 0.0068 - ETA: 3s - 16ms/st - loss: 0.0043 - ETA: 3s - 16ms/st - loss: 0.0066 - ETA: 3s - 16ms/st - loss: 0.0070 - ETA: 3s - 16ms/st - loss: 0.0030 - ETA: 3s - 16ms/st - loss: 0.0148 - ETA: 3s - 16ms/st - loss: 0.0106 - ETA: 2s - 16ms/st - loss: 0.0089 - ETA: 2s - 16ms/st - loss: 0.0181 - ETA: 2s - 16ms/st - loss: 0.0032 - ETA: 2s - 16ms/st - loss: 0.0100 - ETA: 2s - 16ms/st - loss: 0.0054 - ETA: 2s - 16ms/st - loss: 0.0074 - ETA: 2s - 16ms/st - loss: 0.0026 - ETA: 1s - 16ms/st - loss: 0.0114 - ETA: 1s - 16ms/st - loss: 0.0190 - ETA: 1s - 16ms/st - loss: 0.0015 - ETA: 1s - 16ms/st - loss: 0.0010 - ETA: 1s - 16ms/st - loss: 0.0032 - ETA: 1s - 16ms/st - loss: 0.0125 - ETA: 0s - 16ms/st - loss: 0.0141 - ETA: 0s - 16ms/st - loss: 0.0043 - ETA: 0s - 16ms/st - loss: 0.0038 - ETA: 0s - 16ms/st - loss: 0.0083 - ETA: 0s - 16ms/st - loss: 0.0036 - ETA: 0s - 16ms/st - loss: 0.0069 - ETA: 0s - 16ms/st - loss: 0.0092 - 16ms/step          \n",
      "Epoch 8/100\n",
      "step 341/341 [==============================] - loss: 0.0088 - ETA: 5s - 15ms/st - loss: 0.0057 - ETA: 4s - 15ms/st - loss: 0.0087 - ETA: 4s - 15ms/st - loss: 0.0081 - ETA: 4s - 15ms/st - loss: 0.0062 - ETA: 4s - 15ms/st - loss: 0.0021 - ETA: 4s - 15ms/st - loss: 0.0058 - ETA: 4s - 15ms/st - loss: 0.0034 - ETA: 4s - 15ms/st - loss: 0.0025 - ETA: 3s - 16ms/st - loss: 0.0040 - ETA: 3s - 16ms/st - loss: 0.0065 - ETA: 3s - 16ms/st - loss: 0.0082 - ETA: 3s - 16ms/st - loss: 0.0034 - ETA: 3s - 16ms/st - loss: 0.0067 - ETA: 3s - 16ms/st - loss: 0.0016 - ETA: 2s - 16ms/st - loss: 0.0019 - ETA: 2s - 16ms/st - loss: 0.0087 - ETA: 2s - 16ms/st - loss: 0.0125 - ETA: 2s - 16ms/st - loss: 0.0025 - ETA: 2s - 16ms/st - loss: 0.0068 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/st - loss: 0.0029 - ETA: 1s - 16ms/st - loss: 0.0028 - ETA: 1s - 16ms/st - loss: 0.0064 - ETA: 1s - 16ms/st - loss: 0.0031 - ETA: 1s - 16ms/st - loss: 0.0066 - ETA: 1s - 16ms/st - loss: 0.0068 - ETA: 1s - 16ms/st - loss: 0.0056 - ETA: 0s - 16ms/st - loss: 0.0032 - ETA: 0s - 16ms/st - loss: 0.0057 - ETA: 0s - 16ms/st - loss: 0.0055 - ETA: 0s - 16ms/st - loss: 0.0019 - ETA: 0s - 16ms/st - loss: 0.0072 - ETA: 0s - 16ms/st - loss: 0.0017 - ETA: 0s - 16ms/st - loss: 0.0041 - 16ms/step          \n",
      "Epoch 9/100\n",
      "step 341/341 [==============================] - loss: 0.0037 - ETA: 5s - 16ms/st - loss: 0.0025 - ETA: 4s - 16ms/st - loss: 0.0029 - ETA: 4s - 16ms/st - loss: 0.0037 - ETA: 4s - 16ms/st - loss: 0.0066 - ETA: 4s - 16ms/st - loss: 0.0035 - ETA: 4s - 16ms/st - loss: 0.0020 - ETA: 4s - 16ms/st - loss: 0.0042 - ETA: 4s - 16ms/st - loss: 0.0015 - ETA: 3s - 16ms/st - loss: 0.0100 - ETA: 3s - 16ms/st - loss: 0.0021 - ETA: 3s - 16ms/st - loss: 0.0018 - ETA: 3s - 16ms/st - loss: 0.0049 - ETA: 3s - 16ms/st - loss: 0.0076 - ETA: 3s - 16ms/st - loss: 0.0017 - ETA: 2s - 16ms/st - loss: 0.0022 - ETA: 2s - 16ms/st - loss: 0.0069 - ETA: 2s - 16ms/st - loss: 0.0044 - ETA: 2s - 16ms/st - loss: 0.0015 - ETA: 2s - 16ms/st - loss: 0.0013 - ETA: 2s - 16ms/st - loss: 0.0019 - ETA: 2s - 16ms/st - loss: 0.0036 - ETA: 1s - 16ms/st - loss: 0.0030 - ETA: 1s - 16ms/st - loss: 0.0068 - ETA: 1s - 16ms/st - loss: 0.0077 - ETA: 1s - 16ms/st - loss: 0.0022 - ETA: 1s - 16ms/st - loss: 0.0066 - ETA: 1s - 16ms/st - loss: 0.0020 - ETA: 0s - 16ms/st - loss: 0.0032 - ETA: 0s - 16ms/st - loss: 0.0054 - ETA: 0s - 16ms/st - loss: 0.0019 - ETA: 0s - 16ms/st - loss: 0.0059 - ETA: 0s - 16ms/st - loss: 0.0017 - ETA: 0s - 16ms/st - loss: 0.0029 - ETA: 0s - 16ms/st - loss: 0.0040 - 16ms/step          \n",
      "Epoch 10/100\n",
      "step 341/341 [==============================] - loss: 0.0046 - ETA: 5s - 16ms/st - loss: 0.0021 - ETA: 4s - 16ms/st - loss: 0.0054 - ETA: 4s - 16ms/st - loss: 0.0024 - ETA: 4s - 16ms/st - loss: 0.0023 - ETA: 4s - 16ms/st - loss: 0.0015 - ETA: 4s - 16ms/st - loss: 0.0027 - ETA: 4s - 16ms/st - loss: 0.0017 - ETA: 4s - 16ms/st - loss: 0.0046 - ETA: 3s - 16ms/st - loss: 0.0015 - ETA: 3s - 16ms/st - loss: 0.0054 - ETA: 3s - 16ms/st - loss: 0.0036 - ETA: 3s - 16ms/st - loss: 9.4760e-04 - ETA: 3s - 16ms/st - loss: 0.0079 - ETA: 3s - 16ms/step   - loss: 0.0029 - ETA: 2s - 16ms/st - loss: 0.0029 - ETA: 2s - 16ms/st - loss: 0.0028 - ETA: 2s - 16ms/st - loss: 0.0079 - ETA: 2s - 16ms/st - loss: 0.0019 - ETA: 2s - 16ms/st - loss: 0.0014 - ETA: 2s - 16ms/st - loss: 0.0022 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 1s - 16ms/st - loss: 0.0026 - ETA: 1s - 16ms/st - loss: 0.0021 - ETA: 1s - 16ms/st - loss: 7.2506e-04 - ETA: 1s - 16ms/st - loss: 0.0024 - ETA: 1s - 16ms/step   - loss: 6.7125e-04 - ETA: 1s - 16ms/st - loss: 9.4087e-04 - ETA: 0s - 16ms/st - loss: 0.0018 - ETA: 0s - 16ms/step   - loss: 0.0010 - ETA: 0s - 16ms/st - loss: 0.0039 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/st - loss: 0.0062 - ETA: 0s - 16ms/st - loss: 0.0015 - ETA: 0s - 16ms/st - loss: 0.0014 - 16ms/step          \n",
      "Epoch 11/100\n",
      "step 341/341 [==============================] - loss: 0.0011 - ETA: 5s - 15ms/st - loss: 0.0025 - ETA: 5s - 16ms/st - loss: 0.0015 - ETA: 4s - 16ms/st - loss: 0.0027 - ETA: 4s - 16ms/st - loss: 0.0027 - ETA: 4s - 16ms/st - loss: 0.0030 - ETA: 4s - 16ms/st - loss: 0.0031 - ETA: 4s - 16ms/st - loss: 0.0021 - ETA: 4s - 16ms/st - loss: 0.0016 - ETA: 3s - 16ms/st - loss: 0.0014 - ETA: 3s - 16ms/st - loss: 7.5797e-04 - ETA: 3s - 16ms/st - loss: 0.0014 - ETA: 3s - 16ms/step   - loss: 0.0034 - ETA: 3s - 16ms/st - loss: 0.0031 - ETA: 3s - 16ms/st - loss: 9.2804e-04 - ETA: 2s - 16ms/st - loss: 0.0026 - ETA: 2s - 16ms/step   - loss: 0.0015 - ETA: 2s - 16ms/st - loss: 0.0010 - ETA: 2s - 16ms/st - loss: 0.0014 - ETA: 2s - 16ms/st - loss: 0.0015 - ETA: 2s - 16ms/st - loss: 0.0023 - ETA: 2s - 16ms/st - loss: 0.0025 - ETA: 1s - 16ms/st - loss: 7.5453e-04 - ETA: 1s - 16ms/st - loss: 0.0036 - ETA: 1s - 16ms/step   - loss: 0.0016 - ETA: 1s - 16ms/st - loss: 7.8989e-04 - ETA: 1s - 16ms/st - loss: 8.5708e-04 - ETA: 1s - 16ms/st - loss: 9.1637e-04 - ETA: 0s - 16ms/st - loss: 0.0013 - ETA: 0s - 16ms/step   - loss: 0.0029 - ETA: 0s - 16ms/st - loss: 8.5644e-04 - ETA: 0s - 16ms/st - loss: 0.0020 - ETA: 0s - 16ms/step   - loss: 0.0013 - ETA: 0s - 16ms/st - loss: 0.0020 - ETA: 0s - 16ms/st - loss: 0.0018 - 16ms/step          \n",
      "save checkpoint at /home/aistudio/save_dir/10\n",
      "Epoch 12/100\n",
      "step 341/341 [==============================] - loss: 9.8885e-04 - ETA: 5s - 16ms/st - loss: 0.0042 - ETA: 5s - 16ms/step   - loss: 0.0012 - ETA: 4s - 16ms/st - loss: 0.0023 - ETA: 4s - 16ms/st - loss: 0.0018 - ETA: 4s - 16ms/st - loss: 0.0018 - ETA: 4s - 16ms/st - loss: 8.6822e-04 - ETA: 4s - 16ms/st - loss: 6.3628e-04 - ETA: 4s - 16ms/st - loss: 4.0006e-04 - ETA: 3s - 16ms/st - loss: 0.0015 - ETA: 3s - 16ms/step   - loss: 0.0013 - ETA: 3s - 16ms/st - loss: 0.0012 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 3s - 16ms/st - loss: 0.0025 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/st - loss: 3.3379e-04 - ETA: 2s - 16ms/st - loss: 0.0016 - ETA: 2s - 16ms/step   - loss: 0.0012 - ETA: 2s - 16ms/st - loss: 0.0024 - ETA: 2s - 16ms/st - loss: 0.0025 - ETA: 2s - 16ms/st - loss: 0.0021 - ETA: 1s - 16ms/st - loss: 0.0080 - ETA: 1s - 16ms/st - loss: 0.0013 - ETA: 1s - 16ms/st - loss: 0.0012 - ETA: 1s - 16ms/st - loss: 0.0013 - ETA: 1s - 16ms/st - loss: 0.0013 - ETA: 1s - 16ms/st - loss: 0.0024 - ETA: 0s - 16ms/st - loss: 0.0015 - ETA: 0s - 16ms/st - loss: 0.0029 - ETA: 0s - 16ms/st - loss: 9.5038e-04 - ETA: 0s - 16ms/st - loss: 7.2304e-04 - ETA: 0s - 16ms/st - loss: 0.0014 - ETA: 0s - 16ms/step   - loss: 0.0023 - ETA: 0s - 16ms/st - loss: 0.0035 - 16ms/step          \n",
      "Epoch 13/100\n",
      "step 341/341 [==============================] - loss: 0.0030 - ETA: 5s - 15ms/st - loss: 0.0017 - ETA: 4s - 15ms/st - loss: 5.0882e-04 - ETA: 4s - 15ms/st - loss: 0.0013 - ETA: 4s - 15ms/step   - loss: 0.0010 - ETA: 4s - 15ms/st - loss: 0.0014 - ETA: 4s - 15ms/st - loss: 7.6955e-04 - ETA: 4s - 15ms/st - loss: 6.7193e-04 - ETA: 4s - 15ms/st - loss: 0.0016 - ETA: 3s - 15ms/step   - loss: 0.0016 - ETA: 3s - 16ms/st - loss: 0.0030 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 3s - 16ms/st - loss: 0.0025 - ETA: 3s - 16ms/st - loss: 7.8957e-04 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 0.0018 - ETA: 2s - 16ms/st - loss: 0.0012 - ETA: 2s - 16ms/st - loss: 0.0015 - ETA: 2s - 16ms/st - loss: 6.6436e-04 - ETA: 2s - 16ms/st - loss: 9.9506e-04 - ETA: 2s - 16ms/st - loss: 0.0030 - ETA: 2s - 16ms/step   - loss: 0.0012 - ETA: 1s - 16ms/st - loss: 0.0014 - ETA: 1s - 16ms/st - loss: 7.9495e-04 - ETA: 1s - 16ms/st - loss: 0.0013 - ETA: 1s - 16ms/step   - loss: 0.0019 - ETA: 1s - 16ms/st - loss: 0.0020 - ETA: 1s - 16ms/st - loss: 0.0015 - ETA: 0s - 16ms/st - loss: 8.7309e-04 - ETA: 0s - 16ms/st - loss: 0.0023 - ETA: 0s - 16ms/step   - loss: 0.0028 - ETA: 0s - 16ms/st - loss: 0.0017 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/st - loss: 9.1351e-04 - ETA: 0s - 16ms/st - loss: 0.0010 - 16ms/step              \n",
      "Epoch 14/100\n",
      "step 341/341 [==============================] - loss: 4.9117e-04 - ETA: 5s - 16ms/st - loss: 4.8690e-04 - ETA: 4s - 16ms/st - loss: 0.0013 - ETA: 4s - 16ms/step   - loss: 0.0028 - ETA: 4s - 16ms/st - loss: 0.0022 - ETA: 4s - 16ms/st - loss: 0.0024 - ETA: 4s - 16ms/st - loss: 0.0023 - ETA: 4s - 16ms/st - loss: 7.0286e-04 - ETA: 4s - 16ms/st - loss: 8.9563e-04 - ETA: 3s - 16ms/st - loss: 0.0013 - ETA: 3s - 16ms/step   - loss: 0.0026 - ETA: 3s - 16ms/st - loss: 2.6010e-04 - ETA: 3s - 16ms/st - loss: 0.0025 - ETA: 3s - 16ms/step   - loss: 0.0024 - ETA: 3s - 16ms/st - loss: 0.0013 - ETA: 2s - 16ms/st - loss: 4.4975e-04 - ETA: 2s - 16ms/st - loss: 7.6760e-04 - ETA: 2s - 16ms/st - loss: 0.0012 - ETA: 2s - 16ms/step   - loss: 6.4542e-04 - ETA: 2s - 16ms/st - loss: 0.0016 - ETA: 2s - 16ms/step   - loss: 0.0017 - ETA: 2s - 16ms/st - loss: 7.6120e-04 - ETA: 1s - 16ms/st - loss: 0.0012 - ETA: 1s - 16ms/step   - loss: 0.0024 - ETA: 1s - 16ms/st - loss: 0.0013 - ETA: 1s - 16ms/st - loss: 0.0014 - ETA: 1s - 16ms/st - loss: 0.0015 - ETA: 1s - 16ms/st - loss: 0.0019 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/st - loss: 0.0014 - ETA: 0s - 16ms/st - loss: 0.0038 - ETA: 0s - 16ms/st - loss: 7.2185e-04 - ETA: 0s - 16ms/st - loss: 6.8203e-04 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/step   - loss: 9.6902e-04 - 16ms/step      \n",
      "Epoch 15/100\n",
      "step 341/341 [==============================] - loss: 9.1920e-04 - ETA: 5s - 15ms/st - loss: 9.6180e-04 - ETA: 4s - 15ms/st - loss: 9.7021e-04 - ETA: 4s - 15ms/st - loss: 0.0015 - ETA: 4s - 15ms/step   - loss: 0.0080 - ETA: 4s - 15ms/st - loss: 8.7014e-04 - ETA: 4s - 15ms/st - loss: 0.0025 - ETA: 4s - 15ms/step   - loss: 9.4982e-04 - ETA: 4s - 15ms/st - loss: 4.4645e-04 - ETA: 3s - 15ms/st - loss: 8.4966e-04 - ETA: 3s - 15ms/st - loss: 6.7233e-04 - ETA: 3s - 15ms/st - loss: 3.2514e-04 - ETA: 3s - 15ms/st - loss: 0.0014 - ETA: 3s - 15ms/step   - loss: 0.0022 - ETA: 3s - 15ms/st - loss: 0.0029 - ETA: 2s - 15ms/st - loss: 0.0013 - ETA: 2s - 15ms/st - loss: 8.9102e-04 - ETA: 2s - 15ms/st - loss: 7.1691e-04 - ETA: 2s - 15ms/st - loss: 8.1549e-04 - ETA: 2s - 15ms/st - loss: 0.0014 - ETA: 2s - 15ms/step   - loss: 0.0011 - ETA: 2s - 15ms/st - loss: 0.0018 - ETA: 1s - 15ms/st - loss: 0.0058 - ETA: 1s - 15ms/st - loss: 0.0018 - ETA: 1s - 15ms/st - loss: 7.7415e-04 - ETA: 1s - 16ms/st - loss: 4.0641e-04 - ETA: 1s - 16ms/st - loss: 0.0014 - ETA: 1s - 16ms/step   - loss: 0.0021 - ETA: 0s - 15ms/st - loss: 0.0012 - ETA: 0s - 16ms/st - loss: 7.7861e-04 - ETA: 0s - 16ms/st - loss: 0.0014 - ETA: 0s - 16ms/step   - loss: 9.0055e-04 - ETA: 0s - 16ms/st - loss: 6.4088e-04 - ETA: 0s - 16ms/st - loss: 0.0023 - ETA: 0s - 16ms/step   - loss: 0.0013 - 16ms/step          \n",
      "Epoch 16/100\n",
      "step 341/341 [==============================] - loss: 8.3891e-04 - ETA: 5s - 15ms/st - loss: 0.0013 - ETA: 4s - 15ms/step   - loss: 0.0023 - ETA: 4s - 15ms/st - loss: 0.0029 - ETA: 4s - 15ms/st - loss: 7.0533e-04 - ETA: 4s - 15ms/st - loss: 0.0016 - ETA: 4s - 15ms/step   - loss: 0.0012 - ETA: 4s - 15ms/st - loss: 6.9801e-04 - ETA: 4s - 15ms/st - loss: 8.7542e-04 - ETA: 3s - 15ms/st - loss: 4.1982e-04 - ETA: 3s - 15ms/st - loss: 7.4219e-04 - ETA: 3s - 15ms/st - loss: 2.4126e-04 - ETA: 3s - 15ms/st - loss: 0.0014 - ETA: 3s - 15ms/step   - loss: 6.5107e-04 - ETA: 3s - 15ms/st - loss: 7.1300e-04 - ETA: 2s - 15ms/st - loss: 3.6040e-04 - ETA: 2s - 15ms/st - loss: 0.0017 - ETA: 2s - 15ms/step   - loss: 0.0013 - ETA: 2s - 15ms/st - loss: 3.2331e-04 - ETA: 2s - 15ms/st - loss: 0.0016 - ETA: 2s - 15ms/step   - loss: 0.0010 - ETA: 2s - 15ms/st - loss: 0.0014 - ETA: 1s - 15ms/st - loss: 3.1158e-04 - ETA: 1s - 15ms/st - loss: 0.0023 - ETA: 1s - 15ms/step   - loss: 0.0011 - ETA: 1s - 15ms/st - loss: 0.0030 - ETA: 1s - 15ms/st - loss: 0.0015 - ETA: 1s - 15ms/st - loss: 0.0011 - ETA: 0s - 15ms/st - loss: 9.9608e-04 - ETA: 0s - 15ms/st - loss: 8.6730e-04 - ETA: 0s - 15ms/st - loss: 6.4433e-04 - ETA: 0s - 15ms/st - loss: 0.0017 - ETA: 0s - 15ms/step   - loss: 0.0010 - ETA: 0s - 15ms/st - loss: 0.0016 - ETA: 0s - 15ms/st - loss: 8.3307e-04 - 15ms/step      \n",
      "Epoch 17/100\n",
      "step 341/341 [==============================] - loss: 9.9677e-04 - ETA: 5s - 15ms/st - loss: 5.9312e-04 - ETA: 4s - 15ms/st - loss: 0.0021 - ETA: 5s - 16ms/step   - loss: 7.1672e-04 - ETA: 4s - 16ms/st - loss: 7.3383e-04 - ETA: 4s - 16ms/st - loss: 0.0033 - ETA: 4s - 16ms/step   - loss: 0.0013 - ETA: 4s - 16ms/st - loss: 3.9457e-04 - ETA: 4s - 16ms/st - loss: 0.0025 - ETA: 3s - 16ms/step   - loss: 0.0013 - ETA: 3s - 16ms/st - loss: 6.6422e-04 - ETA: 3s - 16ms/st - loss: 0.0020 - ETA: 3s - 16ms/step   - loss: 0.0033 - ETA: 3s - 16ms/st - loss: 0.0017 - ETA: 3s - 16ms/st - loss: 6.3900e-04 - ETA: 2s - 16ms/st - loss: 5.5199e-04 - ETA: 2s - 16ms/st - loss: 6.6958e-04 - ETA: 2s - 16ms/st - loss: 3.7790e-04 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 7.2245e-04 - ETA: 2s - 16ms/st - loss: 3.4602e-04 - ETA: 2s - 16ms/st - loss: 5.6828e-04 - ETA: 1s - 16ms/st - loss: 6.9670e-04 - ETA: 1s - 16ms/st - loss: 0.0011 - ETA: 1s - 16ms/step   - loss: 8.3317e-04 - ETA: 1s - 16ms/st - loss: 0.0064 - ETA: 1s - 16ms/step   - loss: 3.8701e-04 - ETA: 1s - 16ms/st - loss: 2.7078e-04 - ETA: 0s - 16ms/st - loss: 8.1025e-04 - ETA: 0s - 16ms/st - loss: 0.0019 - ETA: 0s - 16ms/step   - loss: 3.8138e-04 - ETA: 0s - 16ms/st - loss: 7.5610e-04 - ETA: 0s - 16ms/st - loss: 6.1817e-04 - ETA: 0s - 16ms/st - loss: 5.4907e-04 - ETA: 0s - 16ms/st - loss: 8.4313e-04 - 16ms/step          \n",
      "Epoch 18/100\n",
      "step 341/341 [==============================] - loss: 9.9427e-04 - ETA: 5s - 16ms/st - loss: 6.7397e-04 - ETA: 5s - 16ms/st - loss: 0.0017 - ETA: 4s - 16ms/step   - loss: 6.5261e-04 - ETA: 4s - 16ms/st - loss: 0.0012 - ETA: 4s - 16ms/step   - loss: 0.0012 - ETA: 4s - 16ms/st - loss: 0.0021 - ETA: 4s - 16ms/st - loss: 8.0346e-04 - ETA: 4s - 16ms/st - loss: 0.0013 - ETA: 3s - 16ms/step   - loss: 4.1162e-04 - ETA: 3s - 16ms/st - loss: 0.0041 - ETA: 3s - 16ms/step   - loss: 8.6492e-04 - ETA: 3s - 16ms/st - loss: 5.1472e-04 - ETA: 3s - 16ms/st - loss: 8.1766e-04 - ETA: 3s - 16ms/st - loss: 0.0012 - ETA: 3s - 16ms/step   - loss: 0.0040 - ETA: 2s - 16ms/st - loss: 0.0023 - ETA: 2s - 16ms/st - loss: 0.0017 - ETA: 2s - 16ms/st - loss: 6.1930e-04 - ETA: 2s - 16ms/st - loss: 5.6088e-04 - ETA: 2s - 16ms/st - loss: 8.9936e-04 - ETA: 2s - 16ms/st - loss: 3.0923e-04 - ETA: 1s - 16ms/st - loss: 0.0017 - ETA: 1s - 16ms/step   - loss: 0.0014 - ETA: 1s - 16ms/st - loss: 0.0017 - ETA: 1s - 16ms/st - loss: 0.0011 - ETA: 1s - 16ms/st - loss: 6.7000e-04 - ETA: 1s - 16ms/st - loss: 0.0024 - ETA: 0s - 16ms/step   - loss: 0.0012 - ETA: 0s - 16ms/st - loss: 0.0018 - ETA: 0s - 16ms/st - loss: 0.0020 - ETA: 0s - 16ms/st - loss: 0.0024 - ETA: 0s - 16ms/st - loss: 0.0018 - ETA: 0s - 16ms/st - loss: 4.0708e-04 - ETA: 0s - 16ms/st - loss: 7.1905e-04 - 16ms/step          \n",
      "Epoch 19/100\n",
      "step 341/341 [==============================] - loss: 0.0021 - ETA: 5s - 16ms/st - loss: 6.9711e-04 - ETA: 5s - 16ms/st - loss: 0.0011 - ETA: 4s - 16ms/step   - loss: 0.0011 - ETA: 4s - 16ms/st - loss: 8.9232e-04 - ETA: 4s - 16ms/st - loss: 0.0043 - ETA: 4s - 16ms/step   - loss: 8.2857e-04 - ETA: 4s - 16ms/st - loss: 5.9678e-04 - ETA: 4s - 16ms/st - loss: 0.0012 - ETA: 3s - 16ms/step   - loss: 0.0019 - ETA: 3s - 16ms/st - loss: 0.0022 - ETA: 3s - 16ms/st - loss: 0.0015 - ETA: 3s - 16ms/st - loss: 7.1012e-04 - ETA: 3s - 16ms/st - loss: 8.0836e-04 - ETA: 3s - 16ms/st - loss: 0.0019 - ETA: 2s - 16ms/step   - loss: 8.4673e-04 - ETA: 2s - 16ms/st - loss: 0.0015 - ETA: 2s - 16ms/step   - loss: 6.5124e-04 - ETA: 2s - 16ms/st - loss: 0.0022 - ETA: 2s - 16ms/step   - loss: 0.0020 - ETA: 2s - 16ms/st - loss: 7.5255e-04 - ETA: 2s - 16ms/st - loss: 5.7582e-04 - ETA: 1s - 16ms/st - loss: 0.0015 - ETA: 1s - 16ms/step   - loss: 9.2380e-04 - ETA: 1s - 16ms/st - loss: 8.9005e-04 - ETA: 1s - 16ms/st - loss: 0.0022 - ETA: 1s - 16ms/step   - loss: 7.2383e-04 - ETA: 1s - 16ms/st - loss: 5.2377e-04 - ETA: 0s - 16ms/st - loss: 9.8523e-04 - ETA: 0s - 16ms/st - loss: 0.0015 - ETA: 0s - 16ms/step   - loss: 8.8900e-04 - ETA: 0s - 16ms/st - loss: 0.0013 - ETA: 0s - 16ms/step   - loss: 9.9508e-04 - ETA: 0s - 16ms/st - loss: 0.0044 - ETA: 0s - 16ms/step   - loss: 0.0011 - 16ms/step          \n",
      "Epoch 20/100\n",
      "step 341/341 [==============================] - loss: 4.9716e-04 - ETA: 5s - 15ms/st - loss: 7.2799e-04 - ETA: 4s - 16ms/st - loss: 8.9539e-04 - ETA: 4s - 16ms/st - loss: 0.0013 - ETA: 4s - 16ms/step   - loss: 0.0030 - ETA: 4s - 16ms/st - loss: 6.0361e-04 - ETA: 4s - 16ms/st - loss: 8.4201e-04 - ETA: 4s - 16ms/st - loss: 8.9321e-04 - ETA: 4s - 16ms/st - loss: 6.5774e-04 - ETA: 3s - 16ms/st - loss: 0.0013 - ETA: 3s - 16ms/step   - loss: 9.1679e-04 - ETA: 3s - 16ms/st - loss: 9.9702e-04 - ETA: 3s - 16ms/st - loss: 0.0065 - ETA: 3s - 16ms/step   - loss: 0.0019 - ETA: 3s - 16ms/st - loss: 0.0012 - ETA: 3s - 17ms/st - loss: 8.8860e-04 - ETA: 3s - 17ms/st - loss: 0.0018 - ETA: 2s - 17ms/step   - loss: 0.0013 - ETA: 2s - 17ms/st - loss: 5.1146e-04 - ETA: 2s - 17ms/st - loss: 0.0013 - ETA: 2s - 17ms/step   - loss: 0.0018 - ETA: 2s - 17ms/st - loss: 0.0020 - ETA: 2s - 17ms/st - loss: 4.8382e-04 - ETA: 1s - 17ms/st - loss: 0.0010 - ETA: 1s - 18ms/step   - loss: 0.0019 - ETA: 1s - 18ms/st - loss: 7.0167e-04 - ETA: 1s - 18ms/st - loss: 6.4808e-04 - ETA: 1s - 17ms/st - loss: 0.0014 - ETA: 1s - 17ms/step   - loss: 0.0022 - ETA: 0s - 17ms/st - loss: 5.9684e-04 - ETA: 0s - 17ms/st - loss: 0.0026 - ETA: 0s - 17ms/step   - loss: 0.0019 - ETA: 0s - 17ms/st - loss: 0.0011 - ETA: 0s - 17ms/st - loss: 5.1246e-04 - ETA: 0s - 17ms/st - loss: 0.0012 - 17ms/step              \n",
      "Epoch 21/100\n",
      "step 341/341 [==============================] - loss: 7.4693e-04 - ETA: 5s - 15ms/st - loss: 0.0030 - ETA: 4s - 15ms/step   - loss: 0.0025 - ETA: 4s - 15ms/st - loss: 0.0019 - ETA: 4s - 15ms/st - loss: 0.0018 - ETA: 4s - 15ms/st - loss: 5.9156e-04 - ETA: 4s - 15ms/st - loss: 7.9818e-04 - ETA: 4s - 16ms/st - loss: 0.0012 - ETA: 4s - 16ms/step   - loss: 3.9550e-04 - ETA: 3s - 16ms/st - loss: 0.0034 - ETA: 3s - 16ms/step   - loss: 0.0043 - ETA: 3s - 16ms/st - loss: 0.0012 - ETA: 3s - 16ms/st - loss: 7.6256e-04 - ETA: 3s - 16ms/st - loss: 0.0017 - ETA: 3s - 16ms/step   - loss: 0.0016 - ETA: 2s - 16ms/st - loss: 3.3623e-04 - ETA: 2s - 16ms/st - loss: 6.6705e-04 - ETA: 2s - 16ms/st - loss: 0.0017 - ETA: 2s - 16ms/step   - loss: 0.0013 - ETA: 2s - 16ms/st - loss: 9.6155e-04 - ETA: 2s - 16ms/st - loss: 0.0040 - ETA: 2s - 16ms/step   - loss: 0.0073 - ETA: 1s - 16ms/st - loss: 0.0014 - ETA: 1s - 16ms/st - loss: 0.0013 - ETA: 1s - 16ms/st - loss: 4.7452e-04 - ETA: 1s - 16ms/st - loss: 7.8284e-04 - ETA: 1s - 16ms/st - loss: 0.0016 - ETA: 1s - 16ms/step   - loss: 0.0020 - ETA: 0s - 16ms/st - loss: 0.0014 - ETA: 0s - 16ms/st - loss: 0.0017 - ETA: 0s - 16ms/st - loss: 3.6046e-04 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/step   - loss: 0.0012 - ETA: 0s - 16ms/st - loss: 9.7390e-04 - ETA: 0s - 16ms/st - loss: 5.2016e-04 - 16ms/step          \n",
      "save checkpoint at /home/aistudio/save_dir/20\n",
      "Epoch 22/100\n",
      "step 341/341 [==============================] - loss: 6.4413e-04 - ETA: 5s - 16ms/st - loss: 2.4242e-04 - ETA: 5s - 16ms/st - loss: 0.0037 - ETA: 4s - 16ms/step   - loss: 4.0556e-04 - ETA: 4s - 16ms/st - loss: 0.0021 - ETA: 4s - 16ms/step   - loss: 9.5721e-04 - ETA: 4s - 16ms/st - loss: 0.0010 - ETA: 4s - 16ms/step   - loss: 5.6114e-04 - ETA: 4s - 16ms/st - loss: 0.0035 - ETA: 3s - 16ms/step   - loss: 0.0020 - ETA: 3s - 16ms/st - loss: 0.0010 - ETA: 3s - 16ms/st - loss: 8.8910e-04 - ETA: 3s - 16ms/st - loss: 6.9429e-04 - ETA: 3s - 16ms/st - loss: 7.6949e-04 - ETA: 3s - 16ms/st - loss: 4.4643e-04 - ETA: 3s - 16ms/st - loss: 7.0806e-04 - ETA: 2s - 16ms/st - loss: 8.2727e-04 - ETA: 2s - 16ms/st - loss: 0.0012 - ETA: 2s - 16ms/step   - loss: 0.0027 - ETA: 2s - 16ms/st - loss: 7.0525e-04 - ETA: 2s - 16ms/st - loss: 0.0018 - ETA: 2s - 16ms/step   - loss: 9.7288e-04 - ETA: 1s - 16ms/st - loss: 7.2164e-04 - ETA: 1s - 16ms/st - loss: 0.0016 - ETA: 1s - 16ms/step   - loss: 7.7439e-04 - ETA: 1s - 16ms/st - loss: 4.8986e-04 - ETA: 1s - 16ms/st - loss: 6.2089e-04 - ETA: 1s - 16ms/st - loss: 8.3720e-04 - ETA: 0s - 16ms/st - loss: 0.0031 - ETA: 0s - 16ms/step   - loss: 0.0013 - ETA: 0s - 16ms/st - loss: 8.7472e-04 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 6.3540e-04 - ETA: 0s - 16ms/st - loss: 8.9754e-04 - ETA: 0s - 16ms/st - loss: 0.0017 - 16ms/step              \n",
      "Epoch 23/100\n",
      "step 341/341 [==============================] - loss: 0.0011 - ETA: 5s - 16ms/st - loss: 2.4463e-04 - ETA: 5s - 16ms/st - loss: 5.8332e-04 - ETA: 5s - 16ms/st - loss: 6.3135e-04 - ETA: 4s - 16ms/st - loss: 5.7410e-04 - ETA: 4s - 16ms/st - loss: 8.8900e-04 - ETA: 4s - 16ms/st - loss: 0.0015 - ETA: 4s - 16ms/step   - loss: 6.4330e-04 - ETA: 4s - 16ms/st - loss: 7.5529e-04 - ETA: 3s - 16ms/st - loss: 9.4387e-04 - ETA: 3s - 16ms/st - loss: 8.0977e-04 - ETA: 3s - 16ms/st - loss: 3.0383e-04 - ETA: 3s - 16ms/st - loss: 0.0027 - ETA: 3s - 16ms/step   - loss: 0.0012 - ETA: 3s - 16ms/st - loss: 0.0016 - ETA: 2s - 16ms/st - loss: 0.0023 - ETA: 2s - 16ms/st - loss: 8.2813e-04 - ETA: 2s - 16ms/st - loss: 6.2370e-04 - ETA: 2s - 16ms/st - loss: 0.0014 - ETA: 2s - 16ms/step   - loss: 0.0012 - ETA: 2s - 16ms/st - loss: 2.4044e-04 - ETA: 2s - 16ms/st - loss: 4.4662e-04 - ETA: 1s - 16ms/st - loss: 5.8786e-04 - ETA: 1s - 16ms/st - loss: 0.0061 - ETA: 1s - 16ms/step   - loss: 0.0026 - ETA: 1s - 16ms/st - loss: 8.4134e-04 - ETA: 1s - 16ms/st - loss: 7.3361e-04 - ETA: 1s - 16ms/st - loss: 0.0046 - ETA: 0s - 16ms/step   - loss: 6.2117e-04 - ETA: 0s - 16ms/st - loss: 5.6784e-04 - ETA: 0s - 16ms/st - loss: 0.0010 - ETA: 0s - 16ms/step   - loss: 0.0018 - ETA: 0s - 16ms/st - loss: 0.0010 - ETA: 0s - 16ms/st - loss: 4.4947e-04 - ETA: 0s - 16ms/st - loss: 7.5004e-04 - 16ms/step          \n",
      "Epoch 24/100\n",
      "step 341/341 [==============================] - loss: 5.7945e-04 - ETA: 5s - 15ms/st - loss: 0.0019 - ETA: 4s - 15ms/step   - loss: 0.0016 - ETA: 4s - 15ms/st - loss: 0.0020 - ETA: 4s - 15ms/st - loss: 8.5885e-04 - ETA: 4s - 15ms/st - loss: 0.0017 - ETA: 4s - 15ms/step   - loss: 6.9158e-04 - ETA: 4s - 15ms/st - loss: 0.0012 - ETA: 4s - 15ms/step   - loss: 0.0016 - ETA: 3s - 15ms/st - loss: 0.0011 - ETA: 3s - 15ms/st - loss: 0.0015 - ETA: 3s - 15ms/st - loss: 8.1250e-04 - ETA: 3s - 15ms/st - loss: 7.2312e-04 - ETA: 3s - 15ms/st - loss: 8.7427e-04 - ETA: 3s - 15ms/st - loss: 4.3758e-04 - ETA: 2s - 15ms/st - loss: 0.0012 - ETA: 2s - 15ms/step   - loss: 0.0038 - ETA: 2s - 15ms/st - loss: 0.0041 - ETA: 2s - 15ms/st - loss: 0.0021 - ETA: 2s - 15ms/st - loss: 6.4972e-04 - ETA: 2s - 15ms/st - loss: 7.8281e-04 - ETA: 2s - 15ms/st - loss: 9.2370e-04 - ETA: 1s - 15ms/st - loss: 7.5162e-04 - ETA: 1s - 15ms/st - loss: 0.0011 - ETA: 1s - 15ms/step   - loss: 4.6309e-04 - ETA: 1s - 15ms/st - loss: 4.1721e-04 - ETA: 1s - 15ms/st - loss: 0.0023 - ETA: 1s - 15ms/step   - loss: 0.0039 - ETA: 0s - 15ms/st - loss: 0.0011 - ETA: 0s - 15ms/st - loss: 3.8238e-04 - ETA: 0s - 15ms/st - loss: 7.0319e-04 - ETA: 0s - 15ms/st - loss: 0.0017 - ETA: 0s - 15ms/step   - loss: 6.4183e-04 - ETA: 0s - 15ms/st - loss: 7.9711e-04 - ETA: 0s - 15ms/st - loss: 6.5834e-04 - 15ms/step          \n",
      "Epoch 25/100\n",
      "step 341/341 [==============================] - loss: 0.0026 - ETA: 5s - 16ms/st - loss: 0.0013 - ETA: 4s - 16ms/st - loss: 7.7609e-04 - ETA: 4s - 16ms/st - loss: 0.0012 - ETA: 4s - 16ms/step   - loss: 9.9289e-04 - ETA: 4s - 16ms/st - loss: 0.0010 - ETA: 4s - 16ms/step   - loss: 0.0012 - ETA: 4s - 16ms/st - loss: 0.0023 - ETA: 4s - 15ms/st - loss: 0.0011 - ETA: 3s - 15ms/st - loss: 5.8841e-04 - ETA: 3s - 16ms/st - loss: 0.0017 - ETA: 3s - 16ms/step   - loss: 7.6476e-04 - ETA: 3s - 16ms/st - loss: 0.0015 - ETA: 3s - 16ms/step   - loss: 0.0010 - ETA: 3s - 16ms/st - loss: 7.1901e-04 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 7.2206e-04 - ETA: 2s - 16ms/st - loss: 4.9319e-04 - ETA: 2s - 16ms/st - loss: 0.0017 - ETA: 2s - 16ms/step   - loss: 0.0038 - ETA: 2s - 16ms/st - loss: 0.0017 - ETA: 2s - 16ms/st - loss: 0.0016 - ETA: 1s - 16ms/st - loss: 0.0013 - ETA: 1s - 16ms/st - loss: 0.0015 - ETA: 1s - 16ms/st - loss: 8.1140e-04 - ETA: 1s - 16ms/st - loss: 0.0012 - ETA: 1s - 16ms/step   - loss: 4.7769e-04 - ETA: 1s - 16ms/st - loss: 7.6247e-04 - ETA: 0s - 16ms/st - loss: 9.4789e-04 - ETA: 0s - 16ms/st - loss: 7.2034e-04 - ETA: 0s - 16ms/st - loss: 7.0639e-04 - ETA: 0s - 16ms/st - loss: 9.2990e-04 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/step   - loss: 0.0011 - ETA: 0s - 16ms/st - loss: 8.9724e-04 - 16ms/step      \n",
      "Epoch 26/100\n",
      "step 341/341 [==============================] - loss: 0.0011 - ETA: 5s - 15ms/st - loss: 0.0019 - ETA: 4s - 15ms/st - loss: 8.9719e-04 - ETA: 4s - 15ms/st - loss: 6.8098e-04 - ETA: 4s - 15ms/st - loss: 0.0015 - ETA: 4s - 15ms/step   - loss: 8.9158e-04 - ETA: 4s - 15ms/st - loss: 8.6160e-04 - ETA: 4s - 15ms/st - loss: 6.2211e-04 - ETA: 4s - 15ms/st - loss: 3.9737e-04 - ETA: 3s - 15ms/st - loss: 3.6280e-04 - ETA: 3s - 15ms/st - loss: 2.2676e-04 - ETA: 3s - 15ms/st - loss: 3.0689e-04 - ETA: 3s - 15ms/st - loss: 0.0021 - ETA: 3s - 15ms/step   - loss: 0.0016 - ETA: 3s - 15ms/st - loss: 7.3226e-04 - ETA: 2s - 15ms/st - loss: 0.0011 - ETA: 2s - 15ms/step   - loss: 0.0023 - ETA: 2s - 15ms/st - loss: 8.6647e-04 - ETA: 2s - 15ms/st - loss: 5.4928e-04 - ETA: 2s - 15ms/st - loss: 7.7150e-04 - ETA: 2s - 15ms/st - loss: 0.0011 - ETA: 2s - 15ms/step   - loss: 5.2601e-04 - ETA: 1s - 15ms/st - loss: 0.0021 - ETA: 1s - 15ms/step   - loss: 0.0015 - ETA: 1s - 15ms/st - loss: 2.6512e-04 - ETA: 1s - 15ms/st - loss: 0.0022 - ETA: 1s - 15ms/step   - loss: 0.0018 - ETA: 1s - 15ms/st - loss: 8.8263e-04 - ETA: 0s - 15ms/st - loss: 5.9982e-04 - ETA: 0s - 15ms/st - loss: 6.4646e-04 - ETA: 0s - 15ms/st - loss: 4.6052e-04 - ETA: 0s - 15ms/st - loss: 0.0015 - ETA: 0s - 15ms/step   - loss: 7.4767e-04 - ETA: 0s - 15ms/st - loss: 0.0021 - ETA: 0s - 15ms/step   - loss: 0.0027 - 15ms/step          \n",
      "Epoch 27/100\n",
      "step 341/341 [==============================] - loss: 0.0018 - ETA: 5s - 16ms/st - loss: 0.0017 - ETA: 4s - 16ms/st - loss: 0.0021 - ETA: 4s - 16ms/st - loss: 0.0010 - ETA: 4s - 16ms/st - loss: 8.2396e-04 - ETA: 4s - 15ms/st - loss: 0.0014 - ETA: 4s - 16ms/step   - loss: 4.8178e-04 - ETA: 4s - 15ms/st - loss: 6.6827e-04 - ETA: 4s - 15ms/st - loss: 9.8379e-04 - ETA: 3s - 15ms/st - loss: 0.0014 - ETA: 3s - 15ms/step   - loss: 9.4527e-04 - ETA: 3s - 15ms/st - loss: 7.3596e-04 - ETA: 3s - 15ms/st - loss: 6.6090e-04 - ETA: 3s - 15ms/st - loss: 0.0022 - ETA: 3s - 15ms/step   - loss: 5.6784e-04 - ETA: 2s - 15ms/st - loss: 6.9984e-04 - ETA: 2s - 15ms/st - loss: 0.0019 - ETA: 2s - 15ms/step   - loss: 0.0019 - ETA: 2s - 15ms/st - loss: 5.7418e-04 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 9.8887e-04 - ETA: 2s - 15ms/st - loss: 7.4197e-04 - ETA: 1s - 15ms/st - loss: 2.3453e-04 - ETA: 1s - 15ms/st - loss: 0.0023 - ETA: 1s - 15ms/step   - loss: 5.1080e-04 - ETA: 1s - 15ms/st - loss: 0.0025 - ETA: 1s - 15ms/step   - loss: 0.0011 - ETA: 1s - 15ms/st - loss: 0.0014 - ETA: 0s - 15ms/st - loss: 0.0015 - ETA: 0s - 15ms/st - loss: 9.7634e-04 - ETA: 0s - 15ms/st - loss: 0.0015 - ETA: 0s - 15ms/step   - loss: 0.0012 - ETA: 0s - 15ms/st - loss: 5.0330e-04 - ETA: 0s - 15ms/st - loss: 2.9113e-04 - ETA: 0s - 15ms/st - loss: 5.1766e-04 - 15ms/step          \n",
      "Epoch 28/100\n",
      "step 341/341 [==============================] - loss: 7.4726e-04 - ETA: 5s - 15ms/st - loss: 0.0011 - ETA: 4s - 15ms/step   - loss: 0.0019 - ETA: 4s - 15ms/st - loss: 5.7254e-04 - ETA: 4s - 15ms/st - loss: 0.0012 - ETA: 4s - 15ms/step   - loss: 7.4885e-04 - ETA: 4s - 15ms/st - loss: 4.4501e-04 - ETA: 4s - 15ms/st - loss: 5.2714e-04 - ETA: 4s - 15ms/st - loss: 5.2509e-04 - ETA: 3s - 15ms/st - loss: 0.0016 - ETA: 3s - 15ms/step   - loss: 7.4128e-04 - ETA: 3s - 15ms/st - loss: 6.6171e-04 - ETA: 3s - 15ms/st - loss: 0.0016 - ETA: 3s - 15ms/step   - loss: 0.0013 - ETA: 3s - 15ms/st - loss: 0.0018 - ETA: 2s - 15ms/st - loss: 0.0015 - ETA: 2s - 15ms/st - loss: 7.6741e-04 - ETA: 2s - 15ms/st - loss: 7.8552e-04 - ETA: 2s - 15ms/st - loss: 8.0248e-04 - ETA: 2s - 15ms/st - loss: 4.5656e-04 - ETA: 2s - 15ms/st - loss: 1.9632e-04 - ETA: 2s - 15ms/st - loss: 6.4299e-04 - ETA: 1s - 15ms/st - loss: 6.5736e-04 - ETA: 1s - 15ms/st - loss: 7.9970e-04 - ETA: 1s - 15ms/st - loss: 0.0012 - ETA: 1s - 15ms/step   - loss: 5.5440e-04 - ETA: 1s - 15ms/st - loss: 5.2643e-04 - ETA: 1s - 15ms/st - loss: 0.0026 - ETA: 0s - 15ms/step   - loss: 4.0859e-04 - ETA: 0s - 15ms/st - loss: 0.0014 - ETA: 0s - 15ms/step   - loss: 7.0554e-04 - ETA: 0s - 15ms/st - loss: 9.2472e-04 - ETA: 0s - 15ms/st - loss: 0.0021 - ETA: 0s - 15ms/step   - loss: 7.8773e-04 - ETA: 0s - 15ms/st - loss: 0.0025 - 16ms/step              \n",
      "Epoch 29/100\n",
      "step 341/341 [==============================] - loss: 6.6524e-04 - ETA: 5s - 16ms/st - loss: 9.8127e-04 - ETA: 5s - 16ms/st - loss: 7.2055e-04 - ETA: 4s - 16ms/st - loss: 9.0893e-04 - ETA: 4s - 16ms/st - loss: 0.0020 - ETA: 4s - 16ms/step   - loss: 8.2530e-04 - ETA: 4s - 16ms/st - loss: 7.7772e-04 - ETA: 4s - 16ms/st - loss: 1.9463e-04 - ETA: 4s - 16ms/st - loss: 5.1418e-04 - ETA: 3s - 16ms/st - loss: 0.0014 - ETA: 3s - 16ms/step   - loss: 0.0014 - ETA: 3s - 16ms/st - loss: 5.9352e-04 - ETA: 3s - 16ms/st - loss: 9.7645e-04 - ETA: 3s - 16ms/st - loss: 0.0016 - ETA: 3s - 16ms/step   - loss: 8.7306e-04 - ETA: 2s - 16ms/st - loss: 2.9900e-04 - ETA: 2s - 16ms/st - loss: 8.3788e-04 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 0.0015 - ETA: 2s - 16ms/st - loss: 6.0695e-04 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 0.0013 - ETA: 1s - 16ms/st - loss: 6.3725e-04 - ETA: 1s - 16ms/st - loss: 5.5990e-04 - ETA: 1s - 16ms/st - loss: 9.0266e-04 - ETA: 1s - 16ms/st - loss: 0.0017 - ETA: 1s - 16ms/step   - loss: 0.0011 - ETA: 1s - 16ms/st - loss: 7.0969e-04 - ETA: 0s - 16ms/st - loss: 8.3854e-04 - ETA: 0s - 16ms/st - loss: 0.0010 - ETA: 0s - 16ms/step   - loss: 4.6480e-04 - ETA: 0s - 16ms/st - loss: 5.1610e-04 - ETA: 0s - 16ms/st - loss: 6.3113e-04 - ETA: 0s - 16ms/st - loss: 0.0015 - ETA: 0s - 16ms/step   - loss: 3.3414e-04 - 16ms/step      \n",
      "Epoch 30/100\n",
      "step 341/341 [==============================] - loss: 0.0045 - ETA: 5s - 15ms/st - loss: 0.0015 - ETA: 4s - 15ms/st - loss: 6.1383e-04 - ETA: 4s - 15ms/st - loss: 7.3162e-04 - ETA: 4s - 16ms/st - loss: 7.3161e-04 - ETA: 4s - 16ms/st - loss: 7.2091e-04 - ETA: 4s - 16ms/st - loss: 8.2301e-04 - ETA: 4s - 16ms/st - loss: 0.0011 - ETA: 4s - 16ms/step   - loss: 0.0043 - ETA: 3s - 16ms/st - loss: 0.0014 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 3s - 16ms/st - loss: 6.7875e-04 - ETA: 3s - 16ms/st - loss: 8.8017e-04 - ETA: 3s - 15ms/st - loss: 9.5793e-04 - ETA: 3s - 15ms/st - loss: 6.3137e-04 - ETA: 2s - 15ms/st - loss: 0.0026 - ETA: 2s - 15ms/step   - loss: 0.0012 - ETA: 2s - 15ms/st - loss: 7.0651e-04 - ETA: 2s - 15ms/st - loss: 4.7430e-04 - ETA: 2s - 15ms/st - loss: 3.7674e-04 - ETA: 2s - 16ms/st - loss: 6.3111e-04 - ETA: 2s - 16ms/st - loss: 9.2834e-04 - ETA: 1s - 16ms/st - loss: 7.1328e-04 - ETA: 1s - 16ms/st - loss: 0.0014 - ETA: 1s - 16ms/step   - loss: 3.5474e-04 - ETA: 1s - 16ms/st - loss: 7.1154e-04 - ETA: 1s - 16ms/st - loss: 7.4774e-04 - ETA: 1s - 16ms/st - loss: 0.0020 - ETA: 0s - 16ms/step   - loss: 0.0013 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/st - loss: 3.9183e-04 - ETA: 0s - 16ms/st - loss: 8.0777e-04 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 0.0033 - ETA: 0s - 16ms/st - loss: 0.0011 - 16ms/step          \n",
      "Epoch 31/100\n",
      "step 341/341 [==============================] - loss: 8.3034e-04 - ETA: 5s - 16ms/st - loss: 0.0014 - ETA: 5s - 16ms/step   - loss: 8.1950e-04 - ETA: 4s - 16ms/st - loss: 8.8579e-04 - ETA: 4s - 16ms/st - loss: 9.8308e-04 - ETA: 4s - 16ms/st - loss: 0.0011 - ETA: 4s - 16ms/step   - loss: 6.2418e-04 - ETA: 4s - 16ms/st - loss: 9.7934e-04 - ETA: 4s - 16ms/st - loss: 0.0021 - ETA: 3s - 16ms/step   - loss: 0.0013 - ETA: 3s - 16ms/st - loss: 0.0022 - ETA: 3s - 16ms/st - loss: 3.5284e-04 - ETA: 3s - 16ms/st - loss: 0.0026 - ETA: 3s - 16ms/step   - loss: 0.0018 - ETA: 3s - 16ms/st - loss: 0.0016 - ETA: 2s - 16ms/st - loss: 0.0023 - ETA: 2s - 16ms/st - loss: 8.3305e-04 - ETA: 2s - 16ms/st - loss: 9.8447e-04 - ETA: 2s - 16ms/st - loss: 6.4609e-04 - ETA: 2s - 16ms/st - loss: 8.7917e-04 - ETA: 2s - 16ms/st - loss: 0.0014 - ETA: 2s - 16ms/step   - loss: 0.0014 - ETA: 1s - 16ms/st - loss: 0.0019 - ETA: 1s - 16ms/st - loss: 0.0025 - ETA: 1s - 16ms/st - loss: 0.0024 - ETA: 1s - 16ms/st - loss: 0.0015 - ETA: 1s - 16ms/st - loss: 0.0019 - ETA: 1s - 16ms/st - loss: 0.0015 - ETA: 0s - 16ms/st - loss: 7.5507e-04 - ETA: 0s - 16ms/st - loss: 6.5643e-04 - ETA: 0s - 16ms/st - loss: 6.4073e-04 - ETA: 0s - 16ms/st - loss: 0.0013 - ETA: 0s - 16ms/step   - loss: 0.0016 - ETA: 0s - 16ms/st - loss: 4.7456e-04 - ETA: 0s - 16ms/st - loss: 0.0016 - 16ms/step              \n",
      "save checkpoint at /home/aistudio/save_dir/30\n",
      "Epoch 32/100\n",
      "step 341/341 [==============================] - loss: 0.0011 - ETA: 5s - 16ms/st - loss: 6.7925e-04 - ETA: 5s - 16ms/st - loss: 5.7468e-04 - ETA: 4s - 16ms/st - loss: 0.0011 - ETA: 4s - 16ms/step   - loss: 2.9069e-04 - ETA: 4s - 16ms/st - loss: 0.0019 - ETA: 4s - 16ms/step   - loss: 4.9448e-04 - ETA: 4s - 16ms/st - loss: 0.0013 - ETA: 4s - 16ms/step   - loss: 0.0028 - ETA: 3s - 16ms/st - loss: 8.1990e-04 - ETA: 3s - 16ms/st - loss: 8.2443e-04 - ETA: 3s - 16ms/st - loss: 3.0864e-04 - ETA: 3s - 16ms/st - loss: 4.8303e-04 - ETA: 3s - 16ms/st - loss: 0.0014 - ETA: 3s - 16ms/step   - loss: 0.0030 - ETA: 3s - 16ms/st - loss: 1.9726e-04 - ETA: 2s - 16ms/st - loss: 6.9317e-04 - ETA: 2s - 16ms/st - loss: 0.0014 - ETA: 2s - 16ms/step   - loss: 3.9312e-04 - ETA: 2s - 16ms/st - loss: 6.1637e-04 - ETA: 2s - 16ms/st - loss: 8.2280e-04 - ETA: 2s - 16ms/st - loss: 3.4361e-04 - ETA: 1s - 16ms/st - loss: 6.2692e-04 - ETA: 1s - 16ms/st - loss: 7.1111e-04 - ETA: 1s - 16ms/st - loss: 0.0012 - ETA: 1s - 16ms/step   - loss: 5.3800e-04 - ETA: 1s - 16ms/st - loss: 3.0599e-04 - ETA: 1s - 16ms/st - loss: 7.9793e-04 - ETA: 0s - 16ms/st - loss: 9.0776e-04 - ETA: 0s - 16ms/st - loss: 0.0016 - ETA: 0s - 16ms/step   - loss: 0.0018 - ETA: 0s - 16ms/st - loss: 0.0050 - ETA: 0s - 16ms/st - loss: 0.0017 - ETA: 0s - 16ms/st - loss: 9.1336e-04 - ETA: 0s - 16ms/st - loss: 5.6291e-04 - 16ms/step          \n",
      "Epoch 33/100\n",
      "step 341/341 [==============================] - loss: 3.2529e-04 - ETA: 5s - 15ms/st - loss: 2.1428e-04 - ETA: 4s - 15ms/st - loss: 9.6432e-04 - ETA: 4s - 16ms/st - loss: 0.0012 - ETA: 4s - 16ms/step   - loss: 0.0023 - ETA: 4s - 16ms/st - loss: 6.1387e-04 - ETA: 4s - 16ms/st - loss: 5.4394e-04 - ETA: 4s - 16ms/st - loss: 0.0011 - ETA: 4s - 16ms/step   - loss: 3.5733e-04 - ETA: 3s - 16ms/st - loss: 0.0012 - ETA: 3s - 16ms/step   - loss: 0.0014 - ETA: 3s - 16ms/st - loss: 7.6947e-04 - ETA: 3s - 16ms/st - loss: 8.7896e-04 - ETA: 3s - 16ms/st - loss: 9.9019e-04 - ETA: 3s - 16ms/st - loss: 9.1398e-04 - ETA: 3s - 16ms/st - loss: 4.2043e-04 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 6.0723e-04 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 5.8118e-04 - ETA: 2s - 16ms/st - loss: 7.8243e-04 - ETA: 2s - 16ms/st - loss: 0.0016 - ETA: 1s - 16ms/step   - loss: 7.8427e-04 - ETA: 1s - 16ms/st - loss: 6.0945e-04 - ETA: 1s - 16ms/st - loss: 0.0011 - ETA: 1s - 16ms/step   - loss: 5.8647e-04 - ETA: 1s - 16ms/st - loss: 6.6888e-04 - ETA: 1s - 16ms/st - loss: 6.0519e-04 - ETA: 0s - 16ms/st - loss: 6.5589e-04 - ETA: 0s - 16ms/st - loss: 5.8212e-04 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/step   - loss: 2.5994e-04 - ETA: 0s - 16ms/st - loss: 0.0035 - ETA: 0s - 16ms/step   - loss: 0.0021 - ETA: 0s - 16ms/st - loss: 0.0015 - 16ms/step          \n",
      "Epoch 34/100\n",
      "step 341/341 [==============================] - loss: 0.0021 - ETA: 5s - 16ms/st - loss: 8.9067e-04 - ETA: 5s - 16ms/st - loss: 5.6791e-04 - ETA: 4s - 16ms/st - loss: 2.5231e-04 - ETA: 4s - 16ms/st - loss: 8.7748e-04 - ETA: 4s - 16ms/st - loss: 5.0210e-04 - ETA: 4s - 16ms/st - loss: 7.4486e-04 - ETA: 4s - 16ms/st - loss: 0.0022 - ETA: 4s - 16ms/step   - loss: 4.7776e-04 - ETA: 3s - 16ms/st - loss: 2.6941e-04 - ETA: 3s - 16ms/st - loss: 0.0016 - ETA: 3s - 16ms/step   - loss: 4.7412e-04 - ETA: 3s - 16ms/st - loss: 2.7817e-04 - ETA: 3s - 16ms/st - loss: 8.6970e-04 - ETA: 3s - 16ms/st - loss: 7.6264e-04 - ETA: 3s - 16ms/st - loss: 0.0014 - ETA: 2s - 16ms/step   - loss: 6.3566e-04 - ETA: 2s - 16ms/st - loss: 0.0010 - ETA: 2s - 16ms/step   - loss: 0.0060 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/st - loss: 7.0349e-04 - ETA: 2s - 16ms/st - loss: 0.0023 - ETA: 1s - 16ms/step   - loss: 5.5247e-04 - ETA: 1s - 16ms/st - loss: 6.4116e-04 - ETA: 1s - 16ms/st - loss: 0.0012 - ETA: 1s - 16ms/step   - loss: 0.0012 - ETA: 1s - 16ms/st - loss: 9.9480e-04 - ETA: 1s - 16ms/st - loss: 0.0020 - ETA: 0s - 16ms/step   - loss: 0.0022 - ETA: 0s - 16ms/st - loss: 8.5846e-04 - ETA: 0s - 16ms/st - loss: 0.0026 - ETA: 0s - 16ms/step   - loss: 0.0011 - ETA: 0s - 16ms/st - loss: 8.0813e-04 - ETA: 0s - 16ms/st - loss: 9.7180e-04 - ETA: 0s - 16ms/st - loss: 7.6753e-04 - 16ms/step          \n",
      "Epoch 35/100\n",
      "step 341/341 [==============================] - loss: 0.0042 - ETA: 5s - 16ms/st - loss: 7.1110e-04 - ETA: 5s - 16ms/st - loss: 6.8531e-04 - ETA: 4s - 16ms/st - loss: 0.0021 - ETA: 4s - 16ms/step   - loss: 5.1949e-04 - ETA: 4s - 16ms/st - loss: 4.2080e-04 - ETA: 4s - 16ms/st - loss: 9.0805e-04 - ETA: 4s - 16ms/st - loss: 6.1841e-04 - ETA: 4s - 16ms/st - loss: 9.4941e-04 - ETA: 3s - 15ms/st - loss: 0.0010 - ETA: 3s - 15ms/step   - loss: 0.0017 - ETA: 3s - 15ms/st - loss: 0.0012 - ETA: 3s - 16ms/st - loss: 0.0017 - ETA: 3s - 16ms/st - loss: 0.0019 - ETA: 3s - 16ms/st - loss: 7.3933e-04 - ETA: 2s - 16ms/st - loss: 6.6082e-04 - ETA: 2s - 16ms/st - loss: 9.5512e-04 - ETA: 2s - 16ms/st - loss: 0.0013 - ETA: 2s - 16ms/step   - loss: 0.0012 - ETA: 2s - 16ms/st - loss: 5.3317e-04 - ETA: 2s - 16ms/st - loss: 8.3875e-04 - ETA: 2s - 16ms/st - loss: 4.8929e-04 - ETA: 1s - 16ms/st - loss: 0.0028 - ETA: 1s - 16ms/step   - loss: 0.0011 - ETA: 1s - 16ms/st - loss: 0.0011 - ETA: 1s - 16ms/st - loss: 0.0014 - ETA: 1s - 16ms/st - loss: 9.7272e-04 - ETA: 1s - 16ms/st - loss: 6.5536e-04 - ETA: 0s - 16ms/st - loss: 3.9876e-04 - ETA: 0s - 16ms/st - loss: 3.9320e-04 - ETA: 0s - 16ms/st - loss: 7.9450e-04 - ETA: 0s - 16ms/st - loss: 3.5269e-04 - ETA: 0s - 16ms/st - loss: 7.9006e-04 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/step   - loss: 0.0012 - 16ms/step          \n",
      "Epoch 36/100\n",
      "step 341/341 [==============================] - loss: 2.4332e-04 - ETA: 5s - 15ms/st - loss: 3.5863e-04 - ETA: 4s - 15ms/st - loss: 0.0020 - ETA: 4s - 15ms/step   - loss: 0.0029 - ETA: 4s - 15ms/st - loss: 0.0012 - ETA: 4s - 16ms/st - loss: 0.0013 - ETA: 4s - 16ms/st - loss: 0.0015 - ETA: 4s - 16ms/st - loss: 5.7175e-04 - ETA: 4s - 16ms/st - loss: 8.4650e-04 - ETA: 3s - 16ms/st - loss: 0.0017 - ETA: 3s - 16ms/step   - loss: 9.8373e-04 - ETA: 3s - 16ms/st - loss: 0.0010 - ETA: 3s - 16ms/step   - loss: 5.3206e-04 - ETA: 3s - 16ms/st - loss: 2.0046e-04 - ETA: 3s - 16ms/st - loss: 8.7090e-04 - ETA: 2s - 16ms/st - loss: 0.0013 - ETA: 2s - 16ms/step   - loss: 7.0832e-04 - ETA: 2s - 16ms/st - loss: 0.0029 - ETA: 2s - 16ms/step   - loss: 0.0016 - ETA: 2s - 16ms/st - loss: 6.7486e-04 - ETA: 2s - 16ms/st - loss: 0.0015 - ETA: 2s - 16ms/step   - loss: 0.0010 - ETA: 1s - 16ms/st - loss: 5.2346e-04 - ETA: 1s - 16ms/st - loss: 0.0012 - ETA: 1s - 16ms/step   - loss: 0.0011 - ETA: 1s - 16ms/st - loss: 0.0022 - ETA: 1s - 16ms/st - loss: 0.0039 - ETA: 1s - 16ms/st - loss: 7.2657e-04 - ETA: 0s - 16ms/st - loss: 3.9335e-04 - ETA: 0s - 16ms/st - loss: 0.0015 - ETA: 0s - 16ms/step   - loss: 0.0014 - ETA: 0s - 16ms/st - loss: 0.0052 - ETA: 0s - 16ms/st - loss: 3.6454e-04 - ETA: 0s - 16ms/st - loss: 0.0010 - ETA: 0s - 16ms/step   - loss: 2.0281e-04 - 16ms/step      \n",
      "Epoch 37/100\n",
      "step 341/341 [==============================] - loss: 7.0488e-04 - ETA: 5s - 15ms/st - loss: 5.6891e-04 - ETA: 4s - 15ms/st - loss: 6.0522e-04 - ETA: 4s - 16ms/st - loss: 7.9710e-04 - ETA: 4s - 16ms/st - loss: 5.2231e-04 - ETA: 4s - 16ms/st - loss: 0.0011 - ETA: 4s - 16ms/step   - loss: 0.0014 - ETA: 4s - 16ms/st - loss: 5.1816e-04 - ETA: 4s - 16ms/st - loss: 0.0014 - ETA: 3s - 16ms/step   - loss: 0.0020 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 3s - 16ms/st - loss: 3.4134e-04 - ETA: 3s - 16ms/st - loss: 4.7115e-04 - ETA: 3s - 16ms/st - loss: 4.2120e-04 - ETA: 3s - 16ms/st - loss: 6.1675e-04 - ETA: 3s - 16ms/st - loss: 0.0010 - ETA: 2s - 16ms/step   - loss: 6.7909e-04 - ETA: 2s - 16ms/st - loss: 4.8711e-04 - ETA: 2s - 16ms/st - loss: 5.0130e-04 - ETA: 2s - 16ms/st - loss: 0.0021 - ETA: 2s - 16ms/step   - loss: 9.2489e-04 - ETA: 2s - 16ms/st - loss: 3.9583e-04 - ETA: 1s - 16ms/st - loss: 0.0035 - ETA: 1s - 16ms/step   - loss: 0.0015 - ETA: 1s - 16ms/st - loss: 5.6834e-04 - ETA: 1s - 16ms/st - loss: 8.0556e-04 - ETA: 1s - 16ms/st - loss: 8.2459e-04 - ETA: 1s - 16ms/st - loss: 5.0548e-04 - ETA: 0s - 16ms/st - loss: 6.5457e-04 - ETA: 0s - 16ms/st - loss: 8.2245e-04 - ETA: 0s - 16ms/st - loss: 0.0013 - ETA: 0s - 16ms/step   - loss: 4.7183e-04 - ETA: 0s - 16ms/st - loss: 3.3806e-04 - ETA: 0s - 16ms/st - loss: 8.0371e-04 - ETA: 0s - 16ms/st - loss: 9.7071e-04 - 16ms/step          \n",
      "Epoch 38/100\n",
      "step 341/341 [==============================] - loss: 0.0017 - ETA: 5s - 16ms/st - loss: 5.0845e-04 - ETA: 5s - 16ms/st - loss: 8.6164e-04 - ETA: 4s - 16ms/st - loss: 5.4116e-04 - ETA: 4s - 16ms/st - loss: 9.6372e-04 - ETA: 4s - 16ms/st - loss: 4.9773e-04 - ETA: 4s - 16ms/st - loss: 0.0023 - ETA: 4s - 16ms/step   - loss: 0.0037 - ETA: 4s - 16ms/st - loss: 3.5514e-04 - ETA: 4s - 16ms/st - loss: 3.2689e-04 - ETA: 3s - 16ms/st - loss: 0.0019 - ETA: 3s - 16ms/step   - loss: 4.0720e-04 - ETA: 3s - 16ms/st - loss: 0.0017 - ETA: 3s - 16ms/step   - loss: 6.4220e-04 - ETA: 3s - 16ms/st - loss: 7.6339e-04 - ETA: 3s - 16ms/st - loss: 6.4961e-04 - ETA: 2s - 16ms/st - loss: 3.7125e-04 - ETA: 2s - 16ms/st - loss: 9.7491e-04 - ETA: 2s - 16ms/st - loss: 7.4547e-04 - ETA: 2s - 16ms/st - loss: 9.9658e-04 - ETA: 2s - 16ms/st - loss: 4.3643e-04 - ETA: 2s - 16ms/st - loss: 5.7860e-04 - ETA: 1s - 16ms/st - loss: 9.4750e-04 - ETA: 1s - 16ms/st - loss: 0.0024 - ETA: 1s - 16ms/step   - loss: 0.0011 - ETA: 1s - 16ms/st - loss: 6.8584e-04 - ETA: 1s - 16ms/st - loss: 5.3014e-04 - ETA: 1s - 16ms/st - loss: 5.9235e-04 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 7.2450e-04 - ETA: 0s - 16ms/st - loss: 4.4567e-04 - ETA: 0s - 16ms/st - loss: 4.3006e-04 - ETA: 0s - 16ms/st - loss: 6.9923e-04 - ETA: 0s - 16ms/st - loss: 2.6629e-04 - ETA: 0s - 16ms/st - loss: 0.0019 - 16ms/step              \n",
      "Epoch 39/100\n",
      "step 341/341 [==============================] - loss: 3.8821e-04 - ETA: 5s - 16ms/st - loss: 0.0013 - ETA: 5s - 16ms/step   - loss: 0.0021 - ETA: 4s - 16ms/st - loss: 0.0023 - ETA: 4s - 16ms/st - loss: 0.0026 - ETA: 4s - 16ms/st - loss: 8.0718e-04 - ETA: 4s - 16ms/st - loss: 3.9716e-04 - ETA: 4s - 16ms/st - loss: 4.8999e-04 - ETA: 4s - 16ms/st - loss: 0.0011 - ETA: 3s - 16ms/step   - loss: 5.6942e-04 - ETA: 3s - 16ms/st - loss: 9.8809e-04 - ETA: 3s - 16ms/st - loss: 9.8233e-04 - ETA: 3s - 16ms/st - loss: 4.4072e-04 - ETA: 3s - 16ms/st - loss: 5.4895e-04 - ETA: 3s - 16ms/st - loss: 0.0013 - ETA: 2s - 16ms/step   - loss: 4.5366e-04 - ETA: 2s - 16ms/st - loss: 9.1716e-04 - ETA: 2s - 16ms/st - loss: 8.3468e-04 - ETA: 2s - 16ms/st - loss: 7.2146e-04 - ETA: 2s - 16ms/st - loss: 0.0030 - ETA: 2s - 16ms/step   - loss: 0.0014 - ETA: 2s - 16ms/st - loss: 7.1943e-04 - ETA: 1s - 16ms/st - loss: 4.4255e-04 - ETA: 1s - 16ms/st - loss: 2.7066e-04 - ETA: 1s - 16ms/st - loss: 6.4671e-04 - ETA: 1s - 16ms/st - loss: 2.8840e-04 - ETA: 1s - 16ms/st - loss: 8.1555e-04 - ETA: 1s - 16ms/st - loss: 5.5985e-04 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/step   - loss: 8.5001e-04 - ETA: 0s - 16ms/st - loss: 5.0204e-04 - ETA: 0s - 16ms/st - loss: 2.4486e-04 - ETA: 0s - 16ms/st - loss: 7.3855e-04 - ETA: 0s - 16ms/st - loss: 0.0016 - ETA: 0s - 16ms/step   - loss: 5.4673e-04 - 16ms/step      \n",
      "Epoch 40/100\n",
      "step 341/341 [==============================] - loss: 0.0017 - ETA: 5s - 16ms/st - loss: 7.7921e-04 - ETA: 5s - 16ms/st - loss: 5.6748e-04 - ETA: 4s - 16ms/st - loss: 9.7149e-04 - ETA: 4s - 16ms/st - loss: 1.4717e-04 - ETA: 4s - 16ms/st - loss: 6.6919e-04 - ETA: 4s - 16ms/st - loss: 0.0014 - ETA: 4s - 16ms/step   - loss: 0.0011 - ETA: 4s - 16ms/st - loss: 0.0017 - ETA: 3s - 16ms/st - loss: 5.8035e-04 - ETA: 3s - 16ms/st - loss: 0.0019 - ETA: 3s - 16ms/step   - loss: 5.6033e-04 - ETA: 3s - 16ms/st - loss: 3.2976e-04 - ETA: 3s - 16ms/st - loss: 0.0013 - ETA: 3s - 16ms/step   - loss: 0.0012 - ETA: 3s - 16ms/st - loss: 4.4304e-04 - ETA: 2s - 16ms/st - loss: 6.3498e-04 - ETA: 2s - 16ms/st - loss: 4.5045e-04 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 8.5079e-04 - ETA: 2s - 16ms/st - loss: 6.2777e-04 - ETA: 2s - 16ms/st - loss: 0.0016 - ETA: 1s - 16ms/step   - loss: 9.9974e-04 - ETA: 1s - 16ms/st - loss: 4.2719e-04 - ETA: 1s - 16ms/st - loss: 4.8788e-04 - ETA: 1s - 16ms/st - loss: 0.0016 - ETA: 1s - 16ms/step   - loss: 8.2794e-04 - ETA: 1s - 16ms/st - loss: 0.0024 - ETA: 0s - 16ms/step   - loss: 7.9779e-04 - ETA: 0s - 16ms/st - loss: 6.8507e-04 - ETA: 0s - 16ms/st - loss: 0.0015 - ETA: 0s - 16ms/step   - loss: 2.3972e-04 - ETA: 0s - 16ms/st - loss: 9.8320e-04 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/step   - loss: 9.8955e-04 - 16ms/step      \n",
      "Epoch 41/100\n",
      "step 341/341 [==============================] - loss: 0.0026 - ETA: 5s - 15ms/st - loss: 0.0022 - ETA: 5s - 16ms/st - loss: 0.0012 - ETA: 4s - 16ms/st - loss: 8.9456e-04 - ETA: 4s - 16ms/st - loss: 3.1681e-04 - ETA: 4s - 16ms/st - loss: 8.0941e-04 - ETA: 4s - 16ms/st - loss: 9.8832e-04 - ETA: 4s - 16ms/st - loss: 8.2450e-04 - ETA: 4s - 16ms/st - loss: 7.5002e-04 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 3s - 16ms/step   - loss: 3.8428e-04 - ETA: 3s - 16ms/st - loss: 3.2124e-04 - ETA: 3s - 16ms/st - loss: 0.0027 - ETA: 3s - 16ms/step   - loss: 8.4630e-04 - ETA: 3s - 16ms/st - loss: 5.8586e-04 - ETA: 2s - 16ms/st - loss: 4.6752e-04 - ETA: 2s - 16ms/st - loss: 9.7514e-04 - ETA: 2s - 16ms/st - loss: 8.9202e-04 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 4.2991e-04 - ETA: 2s - 16ms/st - loss: 5.3349e-04 - ETA: 2s - 16ms/st - loss: 0.0030 - ETA: 1s - 16ms/step   - loss: 5.6591e-04 - ETA: 1s - 16ms/st - loss: 0.0017 - ETA: 1s - 16ms/step   - loss: 6.5211e-04 - ETA: 1s - 16ms/st - loss: 7.8069e-04 - ETA: 1s - 16ms/st - loss: 4.7476e-04 - ETA: 1s - 16ms/st - loss: 4.1000e-04 - ETA: 0s - 16ms/st - loss: 7.7221e-04 - ETA: 0s - 16ms/st - loss: 9.3034e-04 - ETA: 0s - 16ms/st - loss: 4.6618e-04 - ETA: 0s - 16ms/st - loss: 0.0038 - ETA: 0s - 16ms/step   - loss: 0.0012 - ETA: 0s - 16ms/st - loss: 6.6580e-04 - ETA: 0s - 16ms/st - loss: 2.7150e-04 - 16ms/step          \n",
      "save checkpoint at /home/aistudio/save_dir/40\n",
      "Epoch 42/100\n",
      "step 341/341 [==============================] - loss: 0.0011 - ETA: 5s - 16ms/st - loss: 0.0013 - ETA: 5s - 16ms/st - loss: 0.0019 - ETA: 4s - 16ms/st - loss: 7.6350e-04 - ETA: 4s - 16ms/st - loss: 6.2871e-04 - ETA: 4s - 16ms/st - loss: 5.5383e-04 - ETA: 4s - 16ms/st - loss: 0.0019 - ETA: 4s - 16ms/step   - loss: 7.5838e-04 - ETA: 4s - 16ms/st - loss: 2.7557e-04 - ETA: 3s - 16ms/st - loss: 0.0025 - ETA: 3s - 16ms/step   - loss: 0.0017 - ETA: 3s - 16ms/st - loss: 7.9251e-04 - ETA: 3s - 16ms/st - loss: 0.0030 - ETA: 3s - 16ms/step   - loss: 0.0015 - ETA: 3s - 16ms/st - loss: 3.0185e-04 - ETA: 3s - 16ms/st - loss: 5.4388e-04 - ETA: 2s - 16ms/st - loss: 5.0287e-04 - ETA: 2s - 16ms/st - loss: 0.0013 - ETA: 2s - 16ms/step   - loss: 0.0018 - ETA: 2s - 16ms/st - loss: 0.0010 - ETA: 2s - 16ms/st - loss: 5.5475e-04 - ETA: 2s - 16ms/st - loss: 0.0014 - ETA: 1s - 16ms/step   - loss: 5.0338e-04 - ETA: 1s - 16ms/st - loss: 0.0011 - ETA: 1s - 16ms/step   - loss: 5.7915e-04 - ETA: 1s - 16ms/st - loss: 6.2891e-04 - ETA: 1s - 16ms/st - loss: 8.5011e-04 - ETA: 1s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 2.9679e-04 - ETA: 0s - 16ms/st - loss: 0.0014 - ETA: 0s - 16ms/step   - loss: 2.8219e-04 - ETA: 0s - 16ms/st - loss: 0.0014 - ETA: 0s - 16ms/step   - loss: 7.7646e-04 - ETA: 0s - 16ms/st - loss: 0.0014 - ETA: 0s - 16ms/step   - loss: 0.0011 - 16ms/step          \n",
      "Epoch 43/100\n",
      "step 341/341 [==============================] - loss: 7.2239e-04 - ETA: 5s - 15ms/st - loss: 4.0658e-04 - ETA: 4s - 15ms/st - loss: 0.0011 - ETA: 4s - 15ms/step   - loss: 5.6121e-04 - ETA: 4s - 16ms/st - loss: 5.4959e-04 - ETA: 4s - 16ms/st - loss: 0.0016 - ETA: 4s - 16ms/step   - loss: 0.0017 - ETA: 4s - 16ms/st - loss: 6.6464e-04 - ETA: 4s - 16ms/st - loss: 0.0014 - ETA: 3s - 16ms/step   - loss: 0.0025 - ETA: 3s - 16ms/st - loss: 2.3855e-04 - ETA: 3s - 16ms/st - loss: 5.1578e-04 - ETA: 3s - 16ms/st - loss: 0.0012 - ETA: 3s - 16ms/step   - loss: 0.0018 - ETA: 3s - 16ms/st - loss: 5.0149e-04 - ETA: 2s - 16ms/st - loss: 8.2128e-04 - ETA: 2s - 16ms/st - loss: 7.9973e-04 - ETA: 2s - 16ms/st - loss: 0.0024 - ETA: 2s - 16ms/step   - loss: 0.0029 - ETA: 2s - 16ms/st - loss: 8.0896e-04 - ETA: 2s - 16ms/st - loss: 4.7500e-04 - ETA: 2s - 16ms/st - loss: 0.0018 - ETA: 1s - 16ms/step   - loss: 0.0011 - ETA: 1s - 16ms/st - loss: 6.3436e-04 - ETA: 1s - 16ms/st - loss: 9.8127e-04 - ETA: 1s - 16ms/st - loss: 0.0016 - ETA: 1s - 16ms/step   - loss: 0.0019 - ETA: 1s - 16ms/st - loss: 0.0021 - ETA: 0s - 16ms/st - loss: 9.4910e-04 - ETA: 0s - 16ms/st - loss: 8.2443e-04 - ETA: 0s - 16ms/st - loss: 0.0016 - ETA: 0s - 16ms/step   - loss: 4.5537e-04 - ETA: 0s - 16ms/st - loss: 0.0017 - ETA: 0s - 16ms/step   - loss: 9.4665e-04 - ETA: 0s - 16ms/st - loss: 1.7922e-04 - 16ms/step          \n",
      "Epoch 44/100\n",
      "step 341/341 [==============================] - loss: 0.0014 - ETA: 5s - 16ms/st - loss: 7.6404e-04 - ETA: 5s - 16ms/st - loss: 3.8905e-04 - ETA: 4s - 16ms/st - loss: 7.1509e-04 - ETA: 4s - 16ms/st - loss: 6.8027e-04 - ETA: 4s - 16ms/st - loss: 2.0943e-04 - ETA: 4s - 16ms/st - loss: 0.0019 - ETA: 4s - 16ms/step   - loss: 3.5806e-04 - ETA: 4s - 16ms/st - loss: 8.3811e-04 - ETA: 3s - 16ms/st - loss: 5.8109e-04 - ETA: 3s - 16ms/st - loss: 0.0014 - ETA: 3s - 16ms/step   - loss: 0.0013 - ETA: 3s - 16ms/st - loss: 0.0019 - ETA: 3s - 16ms/st - loss: 6.3383e-04 - ETA: 3s - 16ms/st - loss: 0.0010 - ETA: 3s - 16ms/step   - loss: 7.0892e-04 - ETA: 2s - 16ms/st - loss: 9.0303e-04 - ETA: 2s - 16ms/st - loss: 0.0013 - ETA: 2s - 16ms/step   - loss: 7.2552e-04 - ETA: 2s - 16ms/st - loss: 0.0085 - ETA: 2s - 16ms/step   - loss: 5.8693e-04 - ETA: 2s - 16ms/st - loss: 7.4107e-04 - ETA: 1s - 16ms/st - loss: 3.2408e-04 - ETA: 1s - 16ms/st - loss: 0.0012 - ETA: 1s - 16ms/step   - loss: 9.0816e-04 - ETA: 1s - 16ms/st - loss: 6.1274e-04 - ETA: 1s - 16ms/st - loss: 5.5702e-04 - ETA: 1s - 16ms/st - loss: 7.3211e-04 - ETA: 0s - 16ms/st - loss: 5.5660e-04 - ETA: 0s - 16ms/st - loss: 0.0015 - ETA: 0s - 16ms/step   - loss: 4.7159e-04 - ETA: 0s - 16ms/st - loss: 8.1011e-04 - ETA: 0s - 16ms/st - loss: 9.3794e-04 - ETA: 0s - 16ms/st - loss: 2.0356e-04 - ETA: 0s - 16ms/st - loss: 5.1524e-04 - 16ms/step          \n",
      "Epoch 45/100\n",
      "step 341/341 [==============================] - loss: 6.1303e-04 - ETA: 5s - 16ms/st - loss: 0.0015 - ETA: 5s - 16ms/step   - loss: 9.7272e-04 - ETA: 4s - 16ms/st - loss: 9.5905e-04 - ETA: 4s - 16ms/st - loss: 4.5505e-04 - ETA: 4s - 16ms/st - loss: 6.5668e-04 - ETA: 4s - 16ms/st - loss: 0.0010 - ETA: 4s - 16ms/step   - loss: 3.4803e-04 - ETA: 4s - 16ms/st - loss: 8.6312e-04 - ETA: 3s - 16ms/st - loss: 8.3064e-04 - ETA: 3s - 16ms/st - loss: 7.0075e-04 - ETA: 3s - 16ms/st - loss: 0.0022 - ETA: 3s - 16ms/step   - loss: 0.0014 - ETA: 3s - 16ms/st - loss: 0.0019 - ETA: 3s - 16ms/st - loss: 8.7811e-04 - ETA: 3s - 16ms/st - loss: 7.5205e-04 - ETA: 2s - 16ms/st - loss: 2.3128e-04 - ETA: 2s - 16ms/st - loss: 6.0149e-04 - ETA: 2s - 16ms/st - loss: 0.0013 - ETA: 2s - 16ms/step   - loss: 5.6207e-04 - ETA: 2s - 16ms/st - loss: 0.0036 - ETA: 2s - 16ms/step   - loss: 2.0883e-04 - ETA: 1s - 16ms/st - loss: 7.9275e-04 - ETA: 1s - 16ms/st - loss: 0.0021 - ETA: 1s - 16ms/step   - loss: 8.3333e-04 - ETA: 1s - 16ms/st - loss: 0.0015 - ETA: 1s - 16ms/step   - loss: 0.0013 - ETA: 1s - 16ms/st - loss: 0.0024 - ETA: 0s - 16ms/st - loss: 8.3203e-04 - ETA: 0s - 16ms/st - loss: 0.0013 - ETA: 0s - 16ms/step   - loss: 4.4891e-04 - ETA: 0s - 16ms/st - loss: 0.0018 - ETA: 0s - 16ms/step   - loss: 0.0014 - ETA: 0s - 16ms/st - loss: 0.0026 - ETA: 0s - 16ms/st - loss: 3.4252e-04 - 16ms/step      \n",
      "Epoch 46/100\n",
      "step 341/341 [==============================] - loss: 9.9639e-04 - ETA: 5s - 15ms/st - loss: 8.9263e-04 - ETA: 4s - 16ms/st - loss: 5.4120e-04 - ETA: 4s - 16ms/st - loss: 0.0018 - ETA: 4s - 16ms/step   - loss: 0.0013 - ETA: 4s - 16ms/st - loss: 8.8307e-04 - ETA: 4s - 16ms/st - loss: 2.0458e-04 - ETA: 4s - 16ms/st - loss: 0.0011 - ETA: 4s - 16ms/step   - loss: 5.9554e-04 - ETA: 3s - 16ms/st - loss: 3.4578e-04 - ETA: 3s - 16ms/st - loss: 2.5756e-04 - ETA: 3s - 16ms/st - loss: 4.7514e-04 - ETA: 3s - 16ms/st - loss: 7.0467e-04 - ETA: 3s - 16ms/st - loss: 0.0016 - ETA: 3s - 16ms/step   - loss: 0.0043 - ETA: 3s - 16ms/st - loss: 4.9540e-04 - ETA: 2s - 16ms/st - loss: 8.1170e-04 - ETA: 2s - 16ms/st - loss: 6.5941e-04 - ETA: 2s - 16ms/st - loss: 8.8701e-04 - ETA: 2s - 16ms/st - loss: 6.2825e-04 - ETA: 2s - 16ms/st - loss: 2.5712e-04 - ETA: 2s - 16ms/st - loss: 5.5449e-04 - ETA: 1s - 16ms/st - loss: 6.6015e-04 - ETA: 1s - 16ms/st - loss: 4.7865e-04 - ETA: 1s - 16ms/st - loss: 6.0071e-04 - ETA: 1s - 16ms/st - loss: 8.3066e-04 - ETA: 1s - 16ms/st - loss: 0.0013 - ETA: 1s - 16ms/step   - loss: 4.0020e-04 - ETA: 0s - 16ms/st - loss: 5.6880e-04 - ETA: 0s - 16ms/st - loss: 5.5940e-04 - ETA: 0s - 16ms/st - loss: 4.4798e-04 - ETA: 0s - 16ms/st - loss: 8.6990e-04 - ETA: 0s - 16ms/st - loss: 6.9273e-04 - ETA: 0s - 16ms/st - loss: 9.8642e-04 - ETA: 0s - 16ms/st - loss: 5.9078e-04 - 16ms/step          \n",
      "Epoch 47/100\n",
      "step 341/341 [==============================] - loss: 0.0023 - ETA: 5s - 16ms/st - loss: 6.7688e-04 - ETA: 4s - 16ms/st - loss: 3.2967e-04 - ETA: 4s - 16ms/st - loss: 5.8942e-04 - ETA: 4s - 16ms/st - loss: 0.0011 - ETA: 4s - 16ms/step   - loss: 4.8822e-04 - ETA: 4s - 16ms/st - loss: 7.2718e-04 - ETA: 4s - 16ms/st - loss: 4.8143e-04 - ETA: 4s - 16ms/st - loss: 6.3571e-04 - ETA: 3s - 16ms/st - loss: 0.0018 - ETA: 3s - 16ms/step   - loss: 0.0027 - ETA: 3s - 16ms/st - loss: 0.0017 - ETA: 3s - 16ms/st - loss: 0.0012 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 3s - 16ms/st - loss: 9.2542e-04 - ETA: 3s - 16ms/st - loss: 6.5808e-04 - ETA: 2s - 16ms/st - loss: 6.8643e-04 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 0.0012 - ETA: 2s - 16ms/st - loss: 6.1647e-04 - ETA: 2s - 16ms/st - loss: 0.0015 - ETA: 2s - 16ms/step   - loss: 3.8233e-04 - ETA: 1s - 16ms/st - loss: 4.4244e-04 - ETA: 1s - 16ms/st - loss: 6.0899e-04 - ETA: 1s - 16ms/st - loss: 0.0016 - ETA: 1s - 16ms/step   - loss: 0.0012 - ETA: 1s - 16ms/st - loss: 5.1505e-04 - ETA: 1s - 16ms/st - loss: 3.7987e-04 - ETA: 0s - 16ms/st - loss: 0.0024 - ETA: 0s - 16ms/step   - loss: 7.3156e-04 - ETA: 0s - 16ms/st - loss: 8.7316e-04 - ETA: 0s - 16ms/st - loss: 3.4852e-04 - ETA: 0s - 16ms/st - loss: 8.1089e-04 - ETA: 0s - 16ms/st - loss: 0.0018 - ETA: 0s - 16ms/step   - loss: 7.9238e-04 - 16ms/step      \n",
      "Epoch 48/100\n",
      "step 341/341 [==============================] - loss: 5.4762e-04 - ETA: 5s - 16ms/st - loss: 0.0012 - ETA: 5s - 16ms/step   - loss: 0.0012 - ETA: 4s - 16ms/st - loss: 7.8795e-04 - ETA: 4s - 16ms/st - loss: 9.0216e-04 - ETA: 4s - 16ms/st - loss: 0.0018 - ETA: 4s - 16ms/step   - loss: 0.0012 - ETA: 4s - 16ms/st - loss: 0.0023 - ETA: 4s - 16ms/st - loss: 0.0012 - ETA: 3s - 16ms/st - loss: 2.7368e-04 - ETA: 3s - 16ms/st - loss: 5.7850e-04 - ETA: 3s - 16ms/st - loss: 4.9958e-04 - ETA: 3s - 16ms/st - loss: 0.0020 - ETA: 3s - 16ms/step   - loss: 0.0010 - ETA: 3s - 16ms/st - loss: 0.0013 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/st - loss: 0.0021 - ETA: 2s - 16ms/st - loss: 0.0020 - ETA: 2s - 16ms/st - loss: 6.4736e-04 - ETA: 2s - 16ms/st - loss: 0.0010 - ETA: 2s - 16ms/step   - loss: 0.0016 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 1s - 16ms/st - loss: 9.5968e-04 - ETA: 1s - 16ms/st - loss: 0.0012 - ETA: 1s - 16ms/step   - loss: 9.4975e-04 - ETA: 1s - 16ms/st - loss: 0.0013 - ETA: 1s - 16ms/step   - loss: 6.6145e-04 - ETA: 1s - 16ms/st - loss: 6.2856e-04 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/step   - loss: 0.0013 - ETA: 0s - 16ms/st - loss: 5.6085e-04 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/step   - loss: 7.0160e-04 - ETA: 0s - 16ms/st - loss: 6.5411e-04 - ETA: 0s - 16ms/st - loss: 0.0015 - 16ms/step              \n",
      "Epoch 49/100\n",
      "step 341/341 [==============================] - loss: 6.7735e-04 - ETA: 5s - 16ms/st - loss: 9.6575e-04 - ETA: 5s - 16ms/st - loss: 0.0028 - ETA: 5s - 16ms/step   - loss: 9.7056e-04 - ETA: 4s - 16ms/st - loss: 3.2793e-04 - ETA: 4s - 16ms/st - loss: 0.0021 - ETA: 4s - 16ms/step   - loss: 0.0025 - ETA: 4s - 16ms/st - loss: 0.0010 - ETA: 4s - 16ms/st - loss: 0.0016 - ETA: 4s - 16ms/st - loss: 4.0879e-04 - ETA: 3s - 16ms/st - loss: 0.0014 - ETA: 3s - 16ms/step   - loss: 0.0013 - ETA: 3s - 16ms/st - loss: 6.0156e-04 - ETA: 3s - 16ms/st - loss: 8.1762e-04 - ETA: 3s - 16ms/st - loss: 8.3180e-04 - ETA: 3s - 16ms/st - loss: 8.3292e-04 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 2.6070e-04 - ETA: 2s - 16ms/st - loss: 0.0049 - ETA: 2s - 16ms/step   - loss: 8.1571e-04 - ETA: 2s - 16ms/st - loss: 0.0023 - ETA: 2s - 16ms/step   - loss: 8.2602e-04 - ETA: 1s - 16ms/st - loss: 9.0643e-04 - ETA: 1s - 16ms/st - loss: 9.9219e-04 - ETA: 1s - 16ms/st - loss: 5.5469e-04 - ETA: 1s - 16ms/st - loss: 4.2217e-04 - ETA: 1s - 16ms/st - loss: 8.3111e-04 - ETA: 1s - 16ms/st - loss: 5.0157e-04 - ETA: 0s - 16ms/st - loss: 5.2011e-04 - ETA: 0s - 16ms/st - loss: 8.4966e-04 - ETA: 0s - 16ms/st - loss: 0.0016 - ETA: 0s - 16ms/step   - loss: 0.0020 - ETA: 0s - 16ms/st - loss: 5.0299e-04 - ETA: 0s - 16ms/st - loss: 0.0017 - ETA: 0s - 16ms/step   - loss: 6.0626e-04 - 16ms/step      \n",
      "Epoch 50/100\n",
      "step 341/341 [==============================] - loss: 7.6006e-04 - ETA: 5s - 16ms/st - loss: 0.0058 - ETA: 5s - 16ms/step   - loss: 3.0394e-04 - ETA: 4s - 16ms/st - loss: 8.7392e-04 - ETA: 4s - 16ms/st - loss: 0.0013 - ETA: 4s - 16ms/step   - loss: 0.0013 - ETA: 4s - 16ms/st - loss: 7.4063e-04 - ETA: 4s - 16ms/st - loss: 5.1001e-04 - ETA: 4s - 16ms/st - loss: 0.0011 - ETA: 3s - 16ms/step   - loss: 5.9024e-04 - ETA: 3s - 16ms/st - loss: 4.5362e-04 - ETA: 3s - 16ms/st - loss: 0.0012 - ETA: 3s - 16ms/step   - loss: 5.6478e-04 - ETA: 3s - 16ms/st - loss: 6.5652e-04 - ETA: 3s - 16ms/st - loss: 0.0032 - ETA: 2s - 16ms/step   - loss: 0.0020 - ETA: 2s - 16ms/st - loss: 9.2171e-04 - ETA: 2s - 16ms/st - loss: 6.8089e-04 - ETA: 2s - 16ms/st - loss: 9.6519e-04 - ETA: 2s - 16ms/st - loss: 9.0274e-04 - ETA: 2s - 16ms/st - loss: 4.5207e-04 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 1s - 16ms/step   - loss: 7.5175e-04 - ETA: 1s - 16ms/st - loss: 0.0012 - ETA: 1s - 16ms/step   - loss: 0.0027 - ETA: 1s - 16ms/st - loss: 0.0010 - ETA: 1s - 16ms/st - loss: 0.0061 - ETA: 1s - 16ms/st - loss: 6.5732e-04 - ETA: 0s - 16ms/st - loss: 6.8468e-04 - ETA: 0s - 16ms/st - loss: 0.0022 - ETA: 0s - 16ms/step   - loss: 4.8521e-04 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 0.0014 - ETA: 0s - 16ms/st - loss: 0.0025 - ETA: 0s - 16ms/st - loss: 5.7998e-04 - 16ms/step      \n",
      "Epoch 51/100\n",
      "step 341/341 [==============================] - loss: 0.0014 - ETA: 5s - 16ms/st - loss: 0.0015 - ETA: 5s - 16ms/st - loss: 7.0706e-04 - ETA: 4s - 16ms/st - loss: 0.0013 - ETA: 4s - 16ms/step   - loss: 6.4016e-04 - ETA: 4s - 16ms/st - loss: 7.1319e-04 - ETA: 4s - 16ms/st - loss: 6.1053e-04 - ETA: 4s - 16ms/st - loss: 8.9398e-04 - ETA: 4s - 16ms/st - loss: 4.9518e-04 - ETA: 4s - 16ms/st - loss: 6.2187e-04 - ETA: 3s - 16ms/st - loss: 0.0026 - ETA: 3s - 16ms/step   - loss: 4.8294e-04 - ETA: 3s - 16ms/st - loss: 3.6215e-04 - ETA: 3s - 16ms/st - loss: 2.9006e-04 - ETA: 3s - 16ms/st - loss: 0.0013 - ETA: 3s - 16ms/step   - loss: 0.0012 - ETA: 2s - 16ms/st - loss: 0.0014 - ETA: 2s - 16ms/st - loss: 3.2426e-04 - ETA: 2s - 16ms/st - loss: 4.5859e-04 - ETA: 2s - 16ms/st - loss: 7.0154e-04 - ETA: 2s - 16ms/st - loss: 8.3724e-04 - ETA: 2s - 16ms/st - loss: 0.0017 - ETA: 1s - 16ms/step   - loss: 0.0011 - ETA: 1s - 16ms/st - loss: 6.1960e-04 - ETA: 1s - 16ms/st - loss: 0.0013 - ETA: 1s - 16ms/step   - loss: 5.7862e-04 - ETA: 1s - 16ms/st - loss: 7.4711e-04 - ETA: 1s - 16ms/st - loss: 9.1878e-04 - ETA: 0s - 16ms/st - loss: 0.0019 - ETA: 0s - 16ms/step   - loss: 0.0019 - ETA: 0s - 16ms/st - loss: 3.4179e-04 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 6.0654e-04 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 6.0782e-04 - 16ms/step      \n",
      "save checkpoint at /home/aistudio/save_dir/50\n",
      "Epoch 52/100\n",
      "step 341/341 [==============================] - loss: 0.0029 - ETA: 5s - 16ms/st - loss: 0.0012 - ETA: 5s - 16ms/st - loss: 4.4409e-04 - ETA: 4s - 16ms/st - loss: 8.2164e-04 - ETA: 4s - 16ms/st - loss: 0.0010 - ETA: 4s - 16ms/step   - loss: 9.7387e-04 - ETA: 4s - 16ms/st - loss: 0.0011 - ETA: 4s - 16ms/step   - loss: 0.0012 - ETA: 4s - 16ms/st - loss: 8.0918e-04 - ETA: 3s - 16ms/st - loss: 8.1129e-04 - ETA: 3s - 16ms/st - loss: 0.0015 - ETA: 3s - 16ms/step   - loss: 0.0029 - ETA: 3s - 16ms/st - loss: 7.7819e-04 - ETA: 3s - 16ms/st - loss: 0.0012 - ETA: 3s - 16ms/step   - loss: 3.1397e-04 - ETA: 3s - 16ms/st - loss: 9.3710e-04 - ETA: 2s - 16ms/st - loss: 5.1061e-04 - ETA: 2s - 16ms/st - loss: 7.1176e-04 - ETA: 2s - 16ms/st - loss: 8.8613e-04 - ETA: 2s - 16ms/st - loss: 0.0029 - ETA: 2s - 16ms/step   - loss: 0.0012 - ETA: 2s - 16ms/st - loss: 0.0010 - ETA: 1s - 16ms/st - loss: 0.0014 - ETA: 1s - 16ms/st - loss: 8.7927e-04 - ETA: 1s - 16ms/st - loss: 0.0014 - ETA: 1s - 16ms/step   - loss: 3.9499e-04 - ETA: 1s - 16ms/st - loss: 0.0014 - ETA: 1s - 16ms/step   - loss: 0.0014 - ETA: 0s - 16ms/st - loss: 7.6885e-04 - ETA: 0s - 16ms/st - loss: 0.0014 - ETA: 0s - 16ms/step   - loss: 8.3484e-04 - ETA: 0s - 16ms/st - loss: 3.4243e-04 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/step   - loss: 8.3805e-04 - ETA: 0s - 16ms/st - loss: 0.0014 - 16ms/step              \n",
      "Epoch 53/100\n",
      "step 341/341 [==============================] - loss: 5.9394e-04 - ETA: 5s - 16ms/st - loss: 0.0033 - ETA: 5s - 16ms/step   - loss: 0.0022 - ETA: 4s - 16ms/st - loss: 3.0889e-04 - ETA: 4s - 16ms/st - loss: 9.9639e-04 - ETA: 4s - 16ms/st - loss: 5.8909e-04 - ETA: 4s - 16ms/st - loss: 0.0010 - ETA: 4s - 16ms/step   - loss: 2.7734e-04 - ETA: 4s - 16ms/st - loss: 5.4739e-04 - ETA: 3s - 16ms/st - loss: 0.0052 - ETA: 3s - 16ms/step   - loss: 0.0012 - ETA: 3s - 16ms/st - loss: 0.0020 - ETA: 3s - 16ms/st - loss: 0.0027 - ETA: 3s - 16ms/st - loss: 0.0022 - ETA: 3s - 16ms/st - loss: 4.6577e-04 - ETA: 3s - 16ms/st - loss: 9.8687e-04 - ETA: 2s - 16ms/st - loss: 3.9284e-04 - ETA: 2s - 16ms/st - loss: 8.3979e-04 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 8.9647e-04 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 6.7941e-04 - ETA: 1s - 16ms/st - loss: 0.0011 - ETA: 1s - 16ms/step   - loss: 7.7737e-04 - ETA: 1s - 16ms/st - loss: 0.0011 - ETA: 1s - 16ms/step   - loss: 4.4092e-04 - ETA: 1s - 16ms/st - loss: 5.8820e-04 - ETA: 1s - 16ms/st - loss: 0.0010 - ETA: 0s - 16ms/step   - loss: 4.4542e-04 - ETA: 0s - 16ms/st - loss: 8.0876e-04 - ETA: 0s - 16ms/st - loss: 0.0010 - ETA: 0s - 16ms/step   - loss: 7.2320e-04 - ETA: 0s - 16ms/st - loss: 6.4799e-04 - ETA: 0s - 16ms/st - loss: 8.1455e-04 - ETA: 0s - 16ms/st - loss: 0.0015 - 16ms/step              \n",
      "Epoch 54/100\n",
      "step 341/341 [==============================] - loss: 5.3064e-04 - ETA: 5s - 16ms/st - loss: 3.5097e-04 - ETA: 5s - 16ms/st - loss: 4.4079e-04 - ETA: 4s - 16ms/st - loss: 6.8863e-04 - ETA: 4s - 16ms/st - loss: 9.3476e-04 - ETA: 4s - 16ms/st - loss: 0.0014 - ETA: 4s - 16ms/step   - loss: 9.1189e-04 - ETA: 4s - 16ms/st - loss: 0.0011 - ETA: 4s - 16ms/step   - loss: 5.2461e-04 - ETA: 3s - 16ms/st - loss: 9.1193e-04 - ETA: 3s - 16ms/st - loss: 4.8575e-04 - ETA: 3s - 16ms/st - loss: 5.8677e-04 - ETA: 3s - 16ms/st - loss: 5.6309e-04 - ETA: 3s - 16ms/st - loss: 8.0379e-04 - ETA: 3s - 16ms/st - loss: 7.2646e-04 - ETA: 2s - 16ms/st - loss: 2.5633e-04 - ETA: 2s - 16ms/st - loss: 4.2046e-04 - ETA: 2s - 16ms/st - loss: 8.8700e-04 - ETA: 2s - 16ms/st - loss: 7.9513e-04 - ETA: 2s - 16ms/st - loss: 9.9782e-04 - ETA: 2s - 16ms/st - loss: 6.3025e-04 - ETA: 2s - 16ms/st - loss: 0.0027 - ETA: 1s - 16ms/step   - loss: 8.4013e-04 - ETA: 1s - 16ms/st - loss: 2.9219e-04 - ETA: 1s - 16ms/st - loss: 0.0011 - ETA: 1s - 16ms/step   - loss: 3.3342e-04 - ETA: 1s - 16ms/st - loss: 0.0017 - ETA: 1s - 16ms/step   - loss: 5.4971e-04 - ETA: 0s - 16ms/st - loss: 8.3728e-04 - ETA: 0s - 16ms/st - loss: 0.0030 - ETA: 0s - 16ms/step   - loss: 6.6823e-04 - ETA: 0s - 16ms/st - loss: 9.5970e-04 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/step   - loss: 0.0013 - ETA: 0s - 16ms/st - loss: 0.0023 - 16ms/step          \n",
      "Epoch 55/100\n",
      "step 341/341 [==============================] - loss: 9.6285e-04 - ETA: 5s - 16ms/st - loss: 9.8801e-04 - ETA: 5s - 16ms/st - loss: 0.0040 - ETA: 4s - 16ms/step   - loss: 0.0011 - ETA: 4s - 16ms/st - loss: 0.0010 - ETA: 4s - 16ms/st - loss: 0.0010 - ETA: 4s - 16ms/st - loss: 3.5453e-04 - ETA: 4s - 16ms/st - loss: 3.7296e-04 - ETA: 4s - 16ms/st - loss: 0.0011 - ETA: 4s - 16ms/step   - loss: 0.0014 - ETA: 3s - 16ms/st - loss: 5.9606e-04 - ETA: 3s - 16ms/st - loss: 7.0861e-04 - ETA: 3s - 16ms/st - loss: 0.0015 - ETA: 3s - 16ms/step   - loss: 0.0043 - ETA: 3s - 16ms/st - loss: 8.1313e-04 - ETA: 3s - 16ms/st - loss: 7.9456e-04 - ETA: 2s - 16ms/st - loss: 5.5158e-04 - ETA: 2s - 16ms/st - loss: 0.0013 - ETA: 2s - 16ms/step   - loss: 7.7858e-04 - ETA: 2s - 16ms/st - loss: 0.0019 - ETA: 2s - 16ms/step   - loss: 0.0016 - ETA: 2s - 16ms/st - loss: 3.2632e-04 - ETA: 1s - 16ms/st - loss: 0.0028 - ETA: 1s - 16ms/step   - loss: 2.4465e-04 - ETA: 1s - 16ms/st - loss: 3.2290e-04 - ETA: 1s - 16ms/st - loss: 5.6752e-04 - ETA: 1s - 16ms/st - loss: 0.0012 - ETA: 1s - 16ms/step   - loss: 2.5133e-04 - ETA: 0s - 16ms/st - loss: 0.0024 - ETA: 0s - 16ms/step   - loss: 8.3555e-04 - ETA: 0s - 16ms/st - loss: 6.2308e-04 - ETA: 0s - 16ms/st - loss: 5.6094e-04 - ETA: 0s - 16ms/st - loss: 6.7833e-04 - ETA: 0s - 16ms/st - loss: 7.7637e-04 - ETA: 0s - 16ms/st - loss: 0.0013 - 16ms/step              \n",
      "Epoch 56/100\n",
      "step 341/341 [==============================] - loss: 3.8759e-04 - ETA: 5s - 16ms/st - loss: 0.0015 - ETA: 5s - 16ms/step   - loss: 8.0457e-04 - ETA: 4s - 16ms/st - loss: 7.7365e-04 - ETA: 4s - 16ms/st - loss: 0.0022 - ETA: 4s - 16ms/step   - loss: 5.7266e-04 - ETA: 4s - 16ms/st - loss: 4.1673e-04 - ETA: 4s - 16ms/st - loss: 0.0011 - ETA: 4s - 16ms/step   - loss: 0.0054 - ETA: 3s - 16ms/st - loss: 0.0014 - ETA: 3s - 16ms/st - loss: 4.6492e-04 - ETA: 3s - 16ms/st - loss: 9.9107e-04 - ETA: 3s - 16ms/st - loss: 3.5111e-04 - ETA: 3s - 16ms/st - loss: 0.0014 - ETA: 3s - 16ms/step   - loss: 0.0022 - ETA: 3s - 16ms/st - loss: 0.0014 - ETA: 2s - 16ms/st - loss: 2.4423e-04 - ETA: 2s - 16ms/st - loss: 0.0016 - ETA: 2s - 16ms/step   - loss: 0.0015 - ETA: 2s - 16ms/st - loss: 3.3247e-04 - ETA: 2s - 16ms/st - loss: 9.7555e-04 - ETA: 2s - 16ms/st - loss: 6.6470e-04 - ETA: 1s - 16ms/st - loss: 3.2416e-04 - ETA: 1s - 16ms/st - loss: 5.5258e-04 - ETA: 1s - 16ms/st - loss: 3.7503e-04 - ETA: 1s - 16ms/st - loss: 9.6162e-04 - ETA: 1s - 16ms/st - loss: 0.0018 - ETA: 1s - 16ms/step   - loss: 8.0589e-04 - ETA: 0s - 16ms/st - loss: 2.8885e-04 - ETA: 0s - 16ms/st - loss: 4.2196e-04 - ETA: 0s - 16ms/st - loss: 4.1142e-04 - ETA: 0s - 16ms/st - loss: 0.0020 - ETA: 0s - 16ms/step   - loss: 7.9045e-04 - ETA: 0s - 16ms/st - loss: 3.5123e-04 - ETA: 0s - 16ms/st - loss: 3.4888e-04 - 16ms/step          \n",
      "Epoch 57/100\n",
      "step 341/341 [==============================] - loss: 5.5689e-04 - ETA: 5s - 16ms/st - loss: 9.9581e-04 - ETA: 5s - 16ms/st - loss: 0.0016 - ETA: 4s - 16ms/step   - loss: 9.2636e-04 - ETA: 4s - 16ms/st - loss: 6.3418e-04 - ETA: 4s - 16ms/st - loss: 0.0024 - ETA: 4s - 16ms/step   - loss: 0.0049 - ETA: 4s - 16ms/st - loss: 0.0015 - ETA: 4s - 16ms/st - loss: 0.0010 - ETA: 3s - 16ms/st - loss: 5.4101e-04 - ETA: 3s - 16ms/st - loss: 7.4215e-04 - ETA: 3s - 16ms/st - loss: 3.1891e-04 - ETA: 3s - 16ms/st - loss: 0.0018 - ETA: 3s - 16ms/step   - loss: 0.0025 - ETA: 3s - 16ms/st - loss: 1.7809e-04 - ETA: 2s - 16ms/st - loss: 6.8632e-04 - ETA: 2s - 16ms/st - loss: 7.5749e-04 - ETA: 2s - 16ms/st - loss: 0.0020 - ETA: 2s - 16ms/step   - loss: 0.0013 - ETA: 2s - 16ms/st - loss: 9.5604e-04 - ETA: 2s - 16ms/st - loss: 0.0013 - ETA: 2s - 16ms/step   - loss: 0.0014 - ETA: 1s - 16ms/st - loss: 0.0022 - ETA: 1s - 16ms/st - loss: 8.6517e-04 - ETA: 1s - 16ms/st - loss: 0.0013 - ETA: 1s - 16ms/step   - loss: 0.0015 - ETA: 1s - 16ms/st - loss: 0.0015 - ETA: 1s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/st - loss: 0.0026 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/st - loss: 0.0025 - ETA: 0s - 16ms/st - loss: 9.7391e-04 - ETA: 0s - 16ms/st - loss: 5.8651e-04 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 6.4459e-04 - 16ms/step      \n",
      "Epoch 58/100\n",
      "step 341/341 [==============================] - loss: 0.0018 - ETA: 5s - 16ms/st - loss: 6.0054e-04 - ETA: 4s - 16ms/st - loss: 0.0015 - ETA: 4s - 16ms/step   - loss: 7.8580e-04 - ETA: 4s - 16ms/st - loss: 6.0161e-04 - ETA: 4s - 16ms/st - loss: 0.0010 - ETA: 4s - 16ms/step   - loss: 7.1152e-04 - ETA: 4s - 16ms/st - loss: 0.0018 - ETA: 4s - 16ms/step   - loss: 3.2346e-04 - ETA: 3s - 16ms/st - loss: 8.0812e-04 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 3s - 16ms/step   - loss: 4.7383e-04 - ETA: 3s - 16ms/st - loss: 0.0012 - ETA: 3s - 16ms/step   - loss: 0.0024 - ETA: 3s - 16ms/st - loss: 0.0010 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/st - loss: 0.0020 - ETA: 2s - 16ms/st - loss: 0.0012 - ETA: 2s - 16ms/st - loss: 7.0879e-04 - ETA: 2s - 16ms/st - loss: 6.5555e-04 - ETA: 2s - 16ms/st - loss: 5.9942e-04 - ETA: 2s - 16ms/st - loss: 9.6456e-04 - ETA: 1s - 16ms/st - loss: 4.9029e-04 - ETA: 1s - 16ms/st - loss: 3.9869e-04 - ETA: 1s - 16ms/st - loss: 6.7864e-04 - ETA: 1s - 16ms/st - loss: 9.7986e-04 - ETA: 1s - 16ms/st - loss: 2.6991e-04 - ETA: 1s - 16ms/st - loss: 9.6540e-04 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/step   - loss: 5.9777e-04 - ETA: 0s - 16ms/st - loss: 7.3297e-04 - ETA: 0s - 16ms/st - loss: 0.0064 - ETA: 0s - 16ms/step   - loss: 6.7525e-04 - ETA: 0s - 16ms/st - loss: 0.0018 - ETA: 0s - 16ms/step   - loss: 0.0012 - 16ms/step          \n",
      "Epoch 59/100\n",
      "step 341/341 [==============================] - loss: 3.5982e-04 - ETA: 5s - 16ms/st - loss: 7.0436e-04 - ETA: 5s - 16ms/st - loss: 9.2377e-04 - ETA: 4s - 16ms/st - loss: 5.0419e-04 - ETA: 4s - 16ms/st - loss: 9.3479e-04 - ETA: 4s - 16ms/st - loss: 7.3544e-04 - ETA: 4s - 16ms/st - loss: 5.3894e-04 - ETA: 4s - 16ms/st - loss: 0.0022 - ETA: 4s - 16ms/step   - loss: 0.0011 - ETA: 3s - 16ms/st - loss: 8.5094e-04 - ETA: 3s - 16ms/st - loss: 7.8826e-04 - ETA: 3s - 16ms/st - loss: 4.0633e-04 - ETA: 3s - 16ms/st - loss: 8.5212e-04 - ETA: 3s - 16ms/st - loss: 0.0037 - ETA: 3s - 16ms/step   - loss: 0.0019 - ETA: 3s - 16ms/st - loss: 0.0013 - ETA: 2s - 16ms/st - loss: 9.6144e-04 - ETA: 2s - 16ms/st - loss: 0.0015 - ETA: 2s - 16ms/step   - loss: 0.0012 - ETA: 2s - 16ms/st - loss: 8.0913e-04 - ETA: 2s - 16ms/st - loss: 4.5611e-04 - ETA: 2s - 16ms/st - loss: 6.0295e-04 - ETA: 1s - 16ms/st - loss: 4.9846e-04 - ETA: 1s - 16ms/st - loss: 0.0015 - ETA: 1s - 16ms/step   - loss: 6.1966e-04 - ETA: 1s - 16ms/st - loss: 0.0027 - ETA: 1s - 16ms/step   - loss: 6.5044e-04 - ETA: 1s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/step   - loss: 0.0031 - ETA: 0s - 16ms/st - loss: 9.6808e-04 - ETA: 0s - 16ms/st - loss: 4.1087e-04 - ETA: 0s - 16ms/st - loss: 7.5187e-04 - ETA: 0s - 16ms/st - loss: 8.4341e-04 - ETA: 0s - 16ms/st - loss: 4.6501e-04 - ETA: 0s - 16ms/st - loss: 7.3286e-04 - 16ms/step          \n",
      "Epoch 60/100\n",
      "step 341/341 [==============================] - loss: 4.5978e-04 - ETA: 5s - 16ms/st - loss: 5.8473e-04 - ETA: 4s - 15ms/st - loss: 8.7408e-04 - ETA: 4s - 16ms/st - loss: 0.0014 - ETA: 4s - 16ms/step   - loss: 5.9333e-04 - ETA: 4s - 16ms/st - loss: 6.3126e-04 - ETA: 4s - 16ms/st - loss: 3.9759e-04 - ETA: 4s - 16ms/st - loss: 0.0010 - ETA: 4s - 16ms/step   - loss: 7.0016e-04 - ETA: 3s - 16ms/st - loss: 0.0012 - ETA: 3s - 16ms/step   - loss: 0.0011 - ETA: 3s - 16ms/st - loss: 0.0016 - ETA: 3s - 16ms/st - loss: 5.8368e-04 - ETA: 3s - 16ms/st - loss: 0.0015 - ETA: 3s - 16ms/step   - loss: 0.0012 - ETA: 3s - 16ms/st - loss: 5.5018e-04 - ETA: 2s - 16ms/st - loss: 0.0019 - ETA: 2s - 16ms/step   - loss: 2.7458e-04 - ETA: 2s - 16ms/st - loss: 0.0012 - ETA: 2s - 16ms/step   - loss: 5.9368e-04 - ETA: 2s - 16ms/st - loss: 4.7738e-04 - ETA: 2s - 16ms/st - loss: 5.4624e-04 - ETA: 1s - 16ms/st - loss: 8.8636e-04 - ETA: 1s - 16ms/st - loss: 0.0010 - ETA: 1s - 16ms/step   - loss: 0.0015 - ETA: 1s - 16ms/st - loss: 2.9628e-04 - ETA: 1s - 16ms/st - loss: 0.0010 - ETA: 1s - 16ms/step   - loss: 0.0015 - ETA: 0s - 16ms/st - loss: 3.5170e-04 - ETA: 0s - 16ms/st - loss: 0.0028 - ETA: 0s - 16ms/step   - loss: 0.0013 - ETA: 0s - 16ms/st - loss: 7.1860e-04 - ETA: 0s - 16ms/st - loss: 9.6074e-04 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/step   - loss: 0.0025 - 16ms/step          \n",
      "Epoch 61/100\n",
      "step 341/341 [==============================] - loss: 5.5083e-04 - ETA: 5s - 16ms/st - loss: 6.1855e-04 - ETA: 5s - 16ms/st - loss: 5.4956e-04 - ETA: 4s - 16ms/st - loss: 4.4678e-04 - ETA: 4s - 16ms/st - loss: 6.6635e-04 - ETA: 4s - 16ms/st - loss: 3.5489e-04 - ETA: 4s - 16ms/st - loss: 6.1637e-04 - ETA: 4s - 16ms/st - loss: 0.0025 - ETA: 4s - 16ms/step   - loss: 9.6145e-04 - ETA: 3s - 16ms/st - loss: 4.2748e-04 - ETA: 3s - 16ms/st - loss: 0.0013 - ETA: 3s - 16ms/step   - loss: 7.0138e-04 - ETA: 3s - 16ms/st - loss: 0.0014 - ETA: 3s - 16ms/step   - loss: 7.9476e-04 - ETA: 3s - 16ms/st - loss: 5.0089e-04 - ETA: 2s - 16ms/st - loss: 2.3286e-04 - ETA: 2s - 16ms/st - loss: 6.7922e-04 - ETA: 2s - 16ms/st - loss: 6.6510e-04 - ETA: 2s - 16ms/st - loss: 0.0017 - ETA: 2s - 16ms/step   - loss: 7.7154e-04 - ETA: 2s - 16ms/st - loss: 0.0015 - ETA: 2s - 16ms/step   - loss: 8.3376e-04 - ETA: 1s - 16ms/st - loss: 7.4746e-04 - ETA: 1s - 16ms/st - loss: 9.7697e-04 - ETA: 1s - 16ms/st - loss: 7.6700e-04 - ETA: 1s - 16ms/st - loss: 0.0011 - ETA: 1s - 16ms/step   - loss: 0.0011 - ETA: 1s - 16ms/st - loss: 4.1870e-04 - ETA: 0s - 16ms/st - loss: 9.1876e-04 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 3.6358e-04 - ETA: 0s - 16ms/st - loss: 5.5142e-04 - ETA: 0s - 16ms/st - loss: 4.7364e-04 - ETA: 0s - 16ms/st - loss: 6.4482e-04 - ETA: 0s - 16ms/st - loss: 0.0013 - 16ms/step              \n",
      "save checkpoint at /home/aistudio/save_dir/60\n",
      "Epoch 62/100\n",
      "step 341/341 [==============================] - loss: 8.7728e-04 - ETA: 5s - 16ms/st - loss: 2.7194e-04 - ETA: 5s - 16ms/st - loss: 0.0014 - ETA: 4s - 16ms/step   - loss: 0.0014 - ETA: 4s - 16ms/st - loss: 5.8919e-04 - ETA: 4s - 16ms/st - loss: 0.0010 - ETA: 4s - 16ms/step   - loss: 9.1070e-04 - ETA: 4s - 16ms/st - loss: 9.3144e-04 - ETA: 4s - 16ms/st - loss: 0.0010 - ETA: 3s - 16ms/step   - loss: 0.0015 - ETA: 3s - 16ms/st - loss: 5.5155e-04 - ETA: 3s - 16ms/st - loss: 0.0016 - ETA: 3s - 16ms/step   - loss: 6.7739e-04 - ETA: 3s - 16ms/st - loss: 0.0010 - ETA: 3s - 16ms/step   - loss: 2.4003e-04 - ETA: 3s - 16ms/st - loss: 7.5766e-04 - ETA: 2s - 16ms/st - loss: 6.2405e-04 - ETA: 2s - 16ms/st - loss: 4.5668e-04 - ETA: 2s - 16ms/st - loss: 6.1722e-04 - ETA: 2s - 16ms/st - loss: 0.0012 - ETA: 2s - 16ms/step   - loss: 4.4085e-04 - ETA: 2s - 16ms/st - loss: 5.4663e-04 - ETA: 1s - 16ms/st - loss: 0.0010 - ETA: 1s - 16ms/step   - loss: 5.9979e-04 - ETA: 1s - 16ms/st - loss: 9.5887e-04 - ETA: 1s - 16ms/st - loss: 5.1583e-04 - ETA: 1s - 16ms/st - loss: 0.0011 - ETA: 1s - 16ms/step   - loss: 0.0046 - ETA: 0s - 16ms/st - loss: 9.1682e-04 - ETA: 0s - 16ms/st - loss: 1.6067e-04 - ETA: 0s - 16ms/st - loss: 0.0044 - ETA: 0s - 16ms/step   - loss: 4.0273e-04 - ETA: 0s - 16ms/st - loss: 3.9692e-04 - ETA: 0s - 16ms/st - loss: 9.9685e-04 - ETA: 0s - 16ms/st - loss: 5.6684e-04 - 16ms/step          \n",
      "Epoch 63/100\n",
      "step 341/341 [==============================] - loss: 0.0031 - ETA: 5s - 16ms/st - loss: 7.1080e-04 - ETA: 5s - 16ms/st - loss: 7.1613e-04 - ETA: 4s - 16ms/st - loss: 9.3705e-04 - ETA: 4s - 16ms/st - loss: 0.0015 - ETA: 4s - 16ms/step   - loss: 0.0011 - ETA: 4s - 16ms/st - loss: 7.3917e-04 - ETA: 4s - 16ms/st - loss: 8.9087e-04 - ETA: 4s - 16ms/st - loss: 8.8908e-04 - ETA: 3s - 16ms/st - loss: 9.3771e-04 - ETA: 3s - 16ms/st - loss: 5.9142e-04 - ETA: 3s - 16ms/st - loss: 5.6066e-04 - ETA: 3s - 16ms/st - loss: 0.0016 - ETA: 3s - 16ms/step   - loss: 0.0013 - ETA: 3s - 16ms/st - loss: 9.2987e-04 - ETA: 2s - 16ms/st - loss: 7.1858e-04 - ETA: 2s - 16ms/st - loss: 5.9887e-04 - ETA: 2s - 16ms/st - loss: 6.0174e-04 - ETA: 2s - 16ms/st - loss: 4.4020e-04 - ETA: 2s - 16ms/st - loss: 0.0025 - ETA: 2s - 16ms/step   - loss: 9.3663e-04 - ETA: 2s - 16ms/st - loss: 0.0024 - ETA: 1s - 16ms/step   - loss: 0.0016 - ETA: 1s - 16ms/st - loss: 0.0014 - ETA: 1s - 16ms/st - loss: 0.0028 - ETA: 1s - 16ms/st - loss: 5.3975e-04 - ETA: 1s - 16ms/st - loss: 0.0014 - ETA: 1s - 16ms/step   - loss: 0.0017 - ETA: 0s - 16ms/st - loss: 0.0019 - ETA: 0s - 16ms/st - loss: 6.7677e-04 - ETA: 0s - 16ms/st - loss: 0.0033 - ETA: 0s - 16ms/step   - loss: 5.6452e-04 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/step   - loss: 9.8516e-04 - ETA: 0s - 16ms/st - loss: 5.9865e-04 - 16ms/step          \n",
      "Epoch 64/100\n",
      "step 341/341 [==============================] - loss: 5.3570e-04 - ETA: 5s - 16ms/st - loss: 0.0010 - ETA: 5s - 16ms/step   - loss: 0.0016 - ETA: 4s - 16ms/st - loss: 4.2184e-04 - ETA: 4s - 16ms/st - loss: 0.0012 - ETA: 4s - 16ms/step   - loss: 0.0012 - ETA: 4s - 16ms/st - loss: 0.0026 - ETA: 4s - 16ms/st - loss: 0.0010 - ETA: 4s - 16ms/st - loss: 7.3826e-04 - ETA: 3s - 16ms/st - loss: 2.9996e-04 - ETA: 3s - 16ms/st - loss: 0.0013 - ETA: 3s - 16ms/step   - loss: 0.0011 - ETA: 3s - 16ms/st - loss: 0.0014 - ETA: 3s - 16ms/st - loss: 6.2204e-04 - ETA: 3s - 16ms/st - loss: 7.4259e-04 - ETA: 3s - 16ms/st - loss: 4.3090e-04 - ETA: 2s - 16ms/st - loss: 9.6449e-04 - ETA: 2s - 16ms/st - loss: 0.0038 - ETA: 2s - 16ms/step   - loss: 4.6513e-04 - ETA: 2s - 16ms/st - loss: 0.0012 - ETA: 2s - 16ms/step   - loss: 0.0037 - ETA: 2s - 16ms/st - loss: 5.5581e-04 - ETA: 1s - 16ms/st - loss: 0.0012 - ETA: 1s - 16ms/step   - loss: 6.8358e-04 - ETA: 1s - 16ms/st - loss: 9.7458e-04 - ETA: 1s - 16ms/st - loss: 0.0021 - ETA: 1s - 16ms/step   - loss: 6.1646e-04 - ETA: 1s - 16ms/st - loss: 0.0015 - ETA: 0s - 16ms/step   - loss: 5.3228e-04 - ETA: 0s - 16ms/st - loss: 8.8074e-04 - ETA: 0s - 16ms/st - loss: 2.3939e-04 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 6.0773e-04 - ETA: 0s - 16ms/st - loss: 3.6492e-04 - ETA: 0s - 16ms/st - loss: 2.9997e-04 - 16ms/step          \n",
      "Epoch 65/100\n",
      "step 341/341 [==============================] - loss: 0.0012 - ETA: 5s - 16ms/st - loss: 0.0011 - ETA: 5s - 16ms/st - loss: 0.0011 - ETA: 5s - 16ms/st - loss: 6.6891e-04 - ETA: 4s - 16ms/st - loss: 6.4888e-04 - ETA: 4s - 16ms/st - loss: 0.0011 - ETA: 4s - 16ms/step   - loss: 0.0017 - ETA: 4s - 16ms/st - loss: 9.9732e-04 - ETA: 4s - 16ms/st - loss: 4.1344e-04 - ETA: 4s - 16ms/st - loss: 8.5581e-04 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 3s - 16ms/step   - loss: 4.1806e-04 - ETA: 3s - 16ms/st - loss: 0.0027 - ETA: 3s - 16ms/step   - loss: 0.0017 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 3s - 16ms/st - loss: 7.3541e-04 - ETA: 2s - 16ms/st - loss: 8.4315e-04 - ETA: 2s - 16ms/st - loss: 7.3519e-04 - ETA: 2s - 16ms/st - loss: 0.0015 - ETA: 2s - 16ms/step   - loss: 0.0018 - ETA: 2s - 16ms/st - loss: 2.1235e-04 - ETA: 2s - 16ms/st - loss: 9.0879e-04 - ETA: 1s - 16ms/st - loss: 0.0022 - ETA: 1s - 16ms/step   - loss: 7.4152e-04 - ETA: 1s - 16ms/st - loss: 5.4819e-04 - ETA: 1s - 16ms/st - loss: 0.0011 - ETA: 1s - 16ms/step   - loss: 0.0011 - ETA: 1s - 16ms/st - loss: 8.1966e-04 - ETA: 0s - 16ms/st - loss: 5.2426e-04 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 0.0021 - ETA: 0s - 16ms/st - loss: 4.3746e-04 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/step   - loss: 4.0828e-04 - ETA: 0s - 16ms/st - loss: 0.0016 - 16ms/step              \n",
      "Epoch 66/100\n",
      "step 341/341 [==============================] - loss: 0.0036 - ETA: 5s - 16ms/st - loss: 4.1503e-04 - ETA: 5s - 16ms/st - loss: 3.2187e-04 - ETA: 4s - 16ms/st - loss: 0.0017 - ETA: 4s - 16ms/step   - loss: 0.0010 - ETA: 4s - 16ms/st - loss: 0.0019 - ETA: 4s - 16ms/st - loss: 0.0011 - ETA: 4s - 16ms/st - loss: 0.0011 - ETA: 4s - 16ms/st - loss: 0.0012 - ETA: 3s - 16ms/st - loss: 6.8241e-04 - ETA: 3s - 16ms/st - loss: 0.0012 - ETA: 3s - 16ms/step   - loss: 0.0012 - ETA: 3s - 16ms/st - loss: 8.6834e-04 - ETA: 3s - 16ms/st - loss: 6.2406e-04 - ETA: 3s - 16ms/st - loss: 0.0034 - ETA: 3s - 16ms/step   - loss: 0.0021 - ETA: 2s - 16ms/st - loss: 4.2795e-04 - ETA: 2s - 16ms/st - loss: 6.5856e-04 - ETA: 2s - 16ms/st - loss: 7.0562e-04 - ETA: 2s - 16ms/st - loss: 7.4881e-04 - ETA: 2s - 16ms/st - loss: 8.9865e-04 - ETA: 2s - 16ms/st - loss: 0.0010 - ETA: 1s - 16ms/step   - loss: 7.1078e-04 - ETA: 1s - 16ms/st - loss: 0.0017 - ETA: 1s - 16ms/step   - loss: 0.0028 - ETA: 1s - 16ms/st - loss: 0.0010 - ETA: 1s - 16ms/st - loss: 0.0011 - ETA: 1s - 16ms/st - loss: 7.9284e-04 - ETA: 0s - 16ms/st - loss: 7.2573e-04 - ETA: 0s - 16ms/st - loss: 6.6377e-04 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/step   - loss: 0.0020 - ETA: 0s - 16ms/st - loss: 7.3416e-04 - ETA: 0s - 16ms/st - loss: 4.5728e-04 - ETA: 0s - 16ms/st - loss: 9.6639e-04 - 16ms/step          \n",
      "Epoch 67/100\n",
      "step 341/341 [==============================] - loss: 0.0013 - ETA: 5s - 16ms/st - loss: 5.4322e-04 - ETA: 5s - 16ms/st - loss: 7.5185e-04 - ETA: 4s - 16ms/st - loss: 7.3908e-04 - ETA: 4s - 16ms/st - loss: 9.5301e-04 - ETA: 4s - 16ms/st - loss: 9.7161e-04 - ETA: 4s - 16ms/st - loss: 7.3329e-04 - ETA: 4s - 16ms/st - loss: 0.0029 - ETA: 4s - 16ms/step   - loss: 0.0044 - ETA: 3s - 16ms/st - loss: 9.7318e-04 - ETA: 3s - 16ms/st - loss: 2.0703e-04 - ETA: 3s - 16ms/st - loss: 6.0003e-04 - ETA: 3s - 16ms/st - loss: 2.2234e-04 - ETA: 3s - 16ms/st - loss: 0.0022 - ETA: 3s - 16ms/step   - loss: 5.9061e-04 - ETA: 3s - 16ms/st - loss: 0.0025 - ETA: 2s - 16ms/step   - loss: 0.0014 - ETA: 2s - 16ms/st - loss: 0.0014 - ETA: 2s - 16ms/st - loss: 0.0010 - ETA: 2s - 16ms/st - loss: 0.0012 - ETA: 2s - 16ms/st - loss: 6.9086e-04 - ETA: 2s - 16ms/st - loss: 0.0052 - ETA: 1s - 16ms/step   - loss: 2.2956e-04 - ETA: 1s - 16ms/st - loss: 8.4051e-04 - ETA: 1s - 16ms/st - loss: 0.0011 - ETA: 1s - 16ms/step   - loss: 0.0011 - ETA: 1s - 16ms/st - loss: 2.7248e-04 - ETA: 1s - 16ms/st - loss: 0.0013 - ETA: 0s - 16ms/step   - loss: 7.6157e-04 - ETA: 0s - 16ms/st - loss: 0.0016 - ETA: 0s - 16ms/step   - loss: 0.0010 - ETA: 0s - 16ms/st - loss: 0.0036 - ETA: 0s - 16ms/st - loss: 4.3893e-04 - ETA: 0s - 16ms/st - loss: 3.4277e-04 - ETA: 0s - 16ms/st - loss: 8.1706e-04 - 16ms/step          \n",
      "Epoch 68/100\n",
      "step 341/341 [==============================] - loss: 4.2902e-04 - ETA: 5s - 16ms/st - loss: 5.4192e-04 - ETA: 5s - 16ms/st - loss: 5.4299e-04 - ETA: 4s - 16ms/st - loss: 0.0010 - ETA: 4s - 16ms/step   - loss: 0.0011 - ETA: 4s - 16ms/st - loss: 8.6154e-04 - ETA: 4s - 16ms/st - loss: 8.0948e-04 - ETA: 4s - 16ms/st - loss: 6.2420e-04 - ETA: 4s - 16ms/st - loss: 2.1177e-04 - ETA: 3s - 16ms/st - loss: 2.3254e-04 - ETA: 3s - 16ms/st - loss: 0.0023 - ETA: 3s - 16ms/step   - loss: 5.5401e-04 - ETA: 3s - 16ms/st - loss: 0.0019 - ETA: 3s - 16ms/step   - loss: 0.0016 - ETA: 3s - 16ms/st - loss: 0.0029 - ETA: 3s - 16ms/st - loss: 0.0015 - ETA: 2s - 16ms/st - loss: 0.0020 - ETA: 2s - 16ms/st - loss: 0.0029 - ETA: 2s - 16ms/st - loss: 8.0967e-04 - ETA: 2s - 16ms/st - loss: 0.0013 - ETA: 2s - 16ms/step   - loss: 0.0022 - ETA: 2s - 16ms/st - loss: 0.0077 - ETA: 1s - 16ms/st - loss: 0.0013 - ETA: 1s - 16ms/st - loss: 7.9586e-04 - ETA: 1s - 16ms/st - loss: 6.5897e-04 - ETA: 1s - 16ms/st - loss: 8.3249e-04 - ETA: 1s - 16ms/st - loss: 3.7106e-04 - ETA: 1s - 16ms/st - loss: 8.1063e-04 - ETA: 0s - 16ms/st - loss: 0.0032 - ETA: 0s - 16ms/step   - loss: 8.0114e-04 - ETA: 0s - 16ms/st - loss: 6.1613e-04 - ETA: 0s - 16ms/st - loss: 3.0408e-04 - ETA: 0s - 16ms/st - loss: 9.9637e-04 - ETA: 0s - 16ms/st - loss: 5.2095e-04 - ETA: 0s - 16ms/st - loss: 0.0021 - 16ms/step              \n",
      "Epoch 69/100\n",
      "step 341/341 [==============================] - loss: 0.0011 - ETA: 5s - 16ms/st - loss: 0.0041 - ETA: 5s - 16ms/st - loss: 0.0019 - ETA: 4s - 16ms/st - loss: 6.3099e-04 - ETA: 4s - 16ms/st - loss: 5.3267e-04 - ETA: 4s - 16ms/st - loss: 8.6865e-04 - ETA: 4s - 16ms/st - loss: 4.7776e-04 - ETA: 4s - 16ms/st - loss: 4.0373e-04 - ETA: 4s - 16ms/st - loss: 0.0012 - ETA: 3s - 16ms/step   - loss: 0.0010 - ETA: 3s - 16ms/st - loss: 1.8323e-04 - ETA: 3s - 16ms/st - loss: 6.8401e-04 - ETA: 3s - 16ms/st - loss: 6.4066e-04 - ETA: 3s - 16ms/st - loss: 2.9832e-04 - ETA: 3s - 16ms/st - loss: 7.2504e-04 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 4.8595e-04 - ETA: 2s - 16ms/st - loss: 5.5669e-04 - ETA: 2s - 16ms/st - loss: 3.8128e-04 - ETA: 2s - 16ms/st - loss: 9.2678e-04 - ETA: 2s - 16ms/st - loss: 9.7558e-04 - ETA: 2s - 16ms/st - loss: 4.9030e-04 - ETA: 1s - 16ms/st - loss: 0.0010 - ETA: 1s - 16ms/step   - loss: 3.7343e-04 - ETA: 1s - 16ms/st - loss: 3.6288e-04 - ETA: 1s - 16ms/st - loss: 5.1046e-04 - ETA: 1s - 16ms/st - loss: 0.0011 - ETA: 1s - 16ms/step   - loss: 4.7847e-04 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/step   - loss: 4.0807e-04 - ETA: 0s - 16ms/st - loss: 5.7600e-04 - ETA: 0s - 16ms/st - loss: 6.1931e-04 - ETA: 0s - 16ms/st - loss: 6.9451e-04 - ETA: 0s - 16ms/st - loss: 6.3421e-04 - ETA: 0s - 16ms/st - loss: 5.5097e-04 - 16ms/step          \n",
      "Epoch 70/100\n",
      "step 341/341 [==============================] - loss: 0.0035 - ETA: 5s - 16ms/st - loss: 0.0011 - ETA: 5s - 16ms/st - loss: 0.0016 - ETA: 4s - 16ms/st - loss: 0.0012 - ETA: 4s - 16ms/st - loss: 7.4469e-04 - ETA: 4s - 16ms/st - loss: 0.0026 - ETA: 4s - 16ms/step   - loss: 0.0018 - ETA: 4s - 16ms/st - loss: 0.0026 - ETA: 4s - 16ms/st - loss: 0.0016 - ETA: 3s - 16ms/st - loss: 0.0012 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 3s - 16ms/st - loss: 4.3325e-04 - ETA: 3s - 16ms/st - loss: 0.0014 - ETA: 3s - 16ms/step   - loss: 0.0013 - ETA: 3s - 16ms/st - loss: 3.4768e-04 - ETA: 3s - 16ms/st - loss: 0.0012 - ETA: 2s - 16ms/step   - loss: 0.0029 - ETA: 2s - 16ms/st - loss: 6.8645e-04 - ETA: 2s - 16ms/st - loss: 0.0032 - ETA: 2s - 16ms/step   - loss: 0.0011 - ETA: 2s - 16ms/st - loss: 0.0012 - ETA: 2s - 16ms/st - loss: 0.0016 - ETA: 1s - 16ms/st - loss: 6.0403e-04 - ETA: 1s - 16ms/st - loss: 8.0168e-04 - ETA: 1s - 16ms/st - loss: 7.2641e-04 - ETA: 1s - 16ms/st - loss: 3.9844e-04 - ETA: 1s - 16ms/st - loss: 5.4722e-04 - ETA: 1s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 9.1340e-04 - ETA: 0s - 16ms/st - loss: 7.4816e-04 - ETA: 0s - 16ms/st - loss: 3.9451e-04 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 8.4694e-04 - ETA: 0s - 16ms/st - loss: 0.0014 - ETA: 0s - 16ms/step   - loss: 8.6932e-04 - 16ms/step      \n",
      "Epoch 71/100\n",
      "step 341/341 [==============================] - loss: 0.0013 - ETA: 5s - 16ms/st - loss: 0.0023 - ETA: 5s - 16ms/st - loss: 0.0011 - ETA: 4s - 16ms/st - loss: 3.1030e-04 - ETA: 4s - 16ms/st - loss: 0.0019 - ETA: 4s - 16ms/step   - loss: 0.0012 - ETA: 4s - 16ms/st - loss: 4.9734e-04 - ETA: 4s - 16ms/st - loss: 7.7196e-04 - ETA: 4s - 16ms/st - loss: 4.5333e-04 - ETA: 3s - 16ms/st - loss: 0.0014 - ETA: 3s - 16ms/step   - loss: 0.0013 - ETA: 3s - 16ms/st - loss: 0.0015 - ETA: 3s - 16ms/st - loss: 5.7114e-04 - ETA: 3s - 16ms/st - loss: 0.0015 - ETA: 3s - 16ms/step   - loss: 0.0014 - ETA: 3s - 16ms/st - loss: 8.3733e-04 - ETA: 2s - 16ms/st - loss: 6.1850e-04 - ETA: 2s - 16ms/st - loss: 6.5799e-04 - ETA: 2s - 16ms/st - loss: 7.4846e-04 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 6.7723e-04 - ETA: 2s - 16ms/st - loss: 2.8487e-04 - ETA: 1s - 16ms/st - loss: 3.4711e-04 - ETA: 1s - 16ms/st - loss: 0.0018 - ETA: 1s - 16ms/step   - loss: 0.0022 - ETA: 1s - 16ms/st - loss: 8.4864e-04 - ETA: 1s - 16ms/st - loss: 8.9909e-04 - ETA: 1s - 16ms/st - loss: 4.6106e-04 - ETA: 0s - 16ms/st - loss: 3.2073e-04 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 3.9670e-04 - ETA: 0s - 16ms/st - loss: 0.0021 - ETA: 0s - 16ms/step   - loss: 0.0017 - ETA: 0s - 16ms/st - loss: 6.8931e-04 - ETA: 0s - 16ms/st - loss: 0.0034 - 16ms/step              \n",
      "save checkpoint at /home/aistudio/save_dir/70\n",
      "Epoch 72/100\n",
      "step 341/341 [==============================] - loss: 0.0032 - ETA: 5s - 16ms/st - loss: 0.0016 - ETA: 5s - 16ms/st - loss: 0.0013 - ETA: 4s - 16ms/st - loss: 6.6893e-04 - ETA: 4s - 16ms/st - loss: 3.6758e-04 - ETA: 4s - 16ms/st - loss: 2.8722e-04 - ETA: 4s - 16ms/st - loss: 0.0014 - ETA: 4s - 16ms/step   - loss: 0.0029 - ETA: 4s - 16ms/st - loss: 0.0018 - ETA: 3s - 16ms/st - loss: 0.0020 - ETA: 3s - 16ms/st - loss: 4.6043e-04 - ETA: 3s - 16ms/st - loss: 7.0984e-04 - ETA: 3s - 16ms/st - loss: 5.2809e-04 - ETA: 3s - 16ms/st - loss: 6.4611e-04 - ETA: 3s - 16ms/st - loss: 9.1162e-04 - ETA: 3s - 16ms/st - loss: 9.0213e-04 - ETA: 2s - 16ms/st - loss: 0.0016 - ETA: 2s - 16ms/step   - loss: 0.0018 - ETA: 2s - 16ms/st - loss: 5.2265e-04 - ETA: 2s - 16ms/st - loss: 0.0013 - ETA: 2s - 16ms/step   - loss: 5.4113e-04 - ETA: 2s - 16ms/st - loss: 5.6456e-04 - ETA: 1s - 16ms/st - loss: 7.6156e-04 - ETA: 1s - 16ms/st - loss: 5.9082e-04 - ETA: 1s - 16ms/st - loss: 6.9772e-04 - ETA: 1s - 16ms/st - loss: 0.0016 - ETA: 1s - 16ms/step   - loss: 8.7831e-04 - ETA: 1s - 16ms/st - loss: 0.0015 - ETA: 0s - 16ms/step   - loss: 0.0018 - ETA: 0s - 16ms/st - loss: 5.3458e-04 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 0.0012 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/st - loss: 7.3558e-04 - ETA: 0s - 16ms/st - loss: 5.4885e-04 - 16ms/step          \n",
      "Epoch 73/100\n",
      "step 341/341 [==============================] - loss: 9.4048e-04 - ETA: 5s - 15ms/st - loss: 0.0012 - ETA: 4s - 16ms/step   - loss: 6.9908e-04 - ETA: 4s - 16ms/st - loss: 5.8673e-04 - ETA: 4s - 16ms/st - loss: 4.4110e-04 - ETA: 4s - 16ms/st - loss: 6.6235e-04 - ETA: 4s - 16ms/st - loss: 7.5440e-04 - ETA: 4s - 16ms/st - loss: 0.0020 - ETA: 4s - 16ms/step   - loss: 0.0013 - ETA: 3s - 16ms/st - loss: 0.0014 - ETA: 3s - 16ms/st - loss: 6.5347e-04 - ETA: 3s - 16ms/st - loss: 0.0018 - ETA: 3s - 16ms/step   - loss: 0.0011 - ETA: 3s - 16ms/st - loss: 4.2917e-04 - ETA: 3s - 16ms/st - loss: 0.0010 - ETA: 3s - 16ms/step   - loss: 0.0019 - ETA: 2s - 16ms/st - loss: 5.9764e-04 - ETA: 2s - 16ms/st - loss: 0.0018 - ETA: 2s - 16ms/step   - loss: 0.0014 - ETA: 2s - 16ms/st - loss: 7.5201e-04 - ETA: 2s - 16ms/st - loss: 9.1758e-04 - ETA: 2s - 16ms/st - loss: 7.0567e-04 - ETA: 1s - 16ms/st - loss: 0.0071 - ETA: 1s - 16ms/step   - loss: 0.0053 - ETA: 1s - 16ms/st - loss: 0.0013 - ETA: 1s - 16ms/st - loss: 0.0014 - ETA: 1s - 16ms/st - loss: 3.0328e-04 - ETA: 1s - 16ms/st - loss: 3.3221e-04 - ETA: 0s - 16ms/st - loss: 0.0019 - ETA: 0s - 16ms/step   - loss: 4.8004e-04 - ETA: 0s - 16ms/st - loss: 0.0019 - ETA: 0s - 16ms/step   - loss: 7.0048e-04 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 0.0010 - ETA: 0s - 16ms/st - loss: 8.6070e-04 - 16ms/step      \n",
      "Epoch 74/100\n",
      "step 341/341 [==============================] - loss: 0.0011 - ETA: 5s - 16ms/st - loss: 0.0030 - ETA: 5s - 16ms/st - loss: 0.0011 - ETA: 4s - 16ms/st - loss: 0.0032 - ETA: 4s - 16ms/st - loss: 7.9208e-04 - ETA: 4s - 16ms/st - loss: 0.0012 - ETA: 4s - 16ms/step   - loss: 0.0023 - ETA: 4s - 16ms/st - loss: 5.3911e-04 - ETA: 4s - 16ms/st - loss: 0.0021 - ETA: 3s - 16ms/step   - loss: 0.0031 - ETA: 3s - 16ms/st - loss: 0.0012 - ETA: 3s - 16ms/st - loss: 0.0024 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 3s - 16ms/st - loss: 7.1933e-04 - ETA: 3s - 16ms/st - loss: 0.0015 - ETA: 3s - 16ms/step   - loss: 6.1591e-04 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 0.0012 - ETA: 2s - 16ms/st - loss: 6.8159e-04 - ETA: 2s - 16ms/st - loss: 0.0019 - ETA: 2s - 16ms/step   - loss: 9.0088e-04 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 1s - 16ms/step   - loss: 0.0012 - ETA: 1s - 16ms/st - loss: 0.0013 - ETA: 1s - 16ms/st - loss: 9.2829e-04 - ETA: 1s - 16ms/st - loss: 0.0010 - ETA: 1s - 16ms/step   - loss: 0.0022 - ETA: 1s - 16ms/st - loss: 6.3284e-04 - ETA: 0s - 16ms/st - loss: 9.9158e-04 - ETA: 0s - 16ms/st - loss: 8.3388e-04 - ETA: 0s - 16ms/st - loss: 6.8989e-04 - ETA: 0s - 16ms/st - loss: 9.1903e-04 - ETA: 0s - 16ms/st - loss: 5.8263e-04 - ETA: 0s - 16ms/st - loss: 6.3226e-04 - ETA: 0s - 16ms/st - loss: 3.6974e-04 - 16ms/step          \n",
      "Epoch 75/100\n",
      "step 341/341 [==============================] - loss: 0.0012 - ETA: 5s - 16ms/st - loss: 4.1025e-04 - ETA: 5s - 16ms/st - loss: 4.5392e-04 - ETA: 4s - 16ms/st - loss: 3.7948e-04 - ETA: 4s - 16ms/st - loss: 0.0029 - ETA: 4s - 16ms/step   - loss: 0.0018 - ETA: 4s - 16ms/st - loss: 4.2034e-04 - ETA: 4s - 16ms/st - loss: 0.0024 - ETA: 4s - 16ms/step   - loss: 5.8361e-04 - ETA: 3s - 16ms/st - loss: 4.9451e-04 - ETA: 3s - 16ms/st - loss: 2.7601e-04 - ETA: 3s - 16ms/st - loss: 0.0019 - ETA: 3s - 16ms/step   - loss: 2.7754e-04 - ETA: 3s - 16ms/st - loss: 9.2656e-04 - ETA: 3s - 16ms/st - loss: 3.8567e-04 - ETA: 3s - 16ms/st - loss: 5.0566e-04 - ETA: 2s - 16ms/st - loss: 6.2733e-04 - ETA: 2s - 16ms/st - loss: 6.5283e-04 - ETA: 2s - 16ms/st - loss: 5.2172e-04 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 0.0013 - ETA: 2s - 16ms/st - loss: 0.0012 - ETA: 1s - 16ms/st - loss: 0.0022 - ETA: 1s - 16ms/st - loss: 8.1862e-04 - ETA: 1s - 16ms/st - loss: 8.2662e-04 - ETA: 1s - 16ms/st - loss: 3.4314e-04 - ETA: 1s - 16ms/st - loss: 0.0010 - ETA: 1s - 16ms/step   - loss: 0.0017 - ETA: 0s - 16ms/st - loss: 0.0025 - ETA: 0s - 16ms/st - loss: 4.0315e-04 - ETA: 0s - 16ms/st - loss: 0.0013 - ETA: 0s - 16ms/step   - loss: 7.4440e-04 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 0.0039 - ETA: 0s - 16ms/st - loss: 0.0018 - 16ms/step          \n",
      "Epoch 76/100\n",
      "step 341/341 [==============================] - loss: 0.0012 - ETA: 5s - 16ms/st - loss: 0.0019 - ETA: 5s - 16ms/st - loss: 0.0014 - ETA: 4s - 16ms/st - loss: 0.0010 - ETA: 4s - 16ms/st - loss: 5.7819e-04 - ETA: 4s - 16ms/st - loss: 9.2306e-04 - ETA: 4s - 16ms/st - loss: 9.5112e-04 - ETA: 4s - 16ms/st - loss: 0.0023 - ETA: 4s - 16ms/step   - loss: 3.8761e-04 - ETA: 3s - 16ms/st - loss: 9.0661e-04 - ETA: 3s - 16ms/st - loss: 4.4519e-04 - ETA: 3s - 16ms/st - loss: 8.4755e-04 - ETA: 3s - 16ms/st - loss: 4.1457e-04 - ETA: 3s - 16ms/st - loss: 9.1640e-04 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 3s - 16ms/step   - loss: 9.0146e-04 - ETA: 2s - 16ms/st - loss: 6.1886e-04 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 0.0015 - ETA: 2s - 16ms/st - loss: 8.7759e-04 - ETA: 2s - 16ms/st - loss: 2.9973e-04 - ETA: 2s - 16ms/st - loss: 6.3824e-04 - ETA: 1s - 16ms/st - loss: 0.0021 - ETA: 1s - 16ms/step   - loss: 9.9510e-04 - ETA: 1s - 16ms/st - loss: 5.3267e-04 - ETA: 1s - 16ms/st - loss: 0.0020 - ETA: 1s - 16ms/step   - loss: 7.2633e-04 - ETA: 1s - 16ms/st - loss: 8.0367e-04 - ETA: 0s - 16ms/st - loss: 8.4378e-04 - ETA: 0s - 16ms/st - loss: 8.8562e-04 - ETA: 0s - 16ms/st - loss: 0.0014 - ETA: 0s - 16ms/step   - loss: 0.0046 - ETA: 0s - 16ms/st - loss: 8.7146e-04 - ETA: 0s - 16ms/st - loss: 9.1477e-04 - ETA: 0s - 16ms/st - loss: 0.0019 - 16ms/step              \n",
      "Epoch 77/100\n",
      "step 341/341 [==============================] - loss: 8.8009e-04 - ETA: 5s - 16ms/st - loss: 9.1758e-04 - ETA: 5s - 16ms/st - loss: 5.5398e-04 - ETA: 5s - 16ms/st - loss: 0.0010 - ETA: 4s - 16ms/step   - loss: 7.2112e-04 - ETA: 4s - 16ms/st - loss: 0.0013 - ETA: 4s - 16ms/step   - loss: 0.0021 - ETA: 4s - 16ms/st - loss: 0.0017 - ETA: 4s - 16ms/st - loss: 0.0017 - ETA: 4s - 16ms/st - loss: 8.8443e-04 - ETA: 3s - 16ms/st - loss: 3.5986e-04 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 3s - 16ms/step   - loss: 6.0335e-04 - ETA: 3s - 16ms/st - loss: 4.2945e-04 - ETA: 3s - 16ms/st - loss: 0.0010 - ETA: 3s - 16ms/step   - loss: 0.0017 - ETA: 2s - 16ms/st - loss: 9.4559e-04 - ETA: 2s - 16ms/st - loss: 0.0061 - ETA: 2s - 16ms/step   - loss: 9.4179e-04 - ETA: 2s - 16ms/st - loss: 0.0021 - ETA: 2s - 16ms/step   - loss: 7.2090e-04 - ETA: 2s - 16ms/st - loss: 4.0797e-04 - ETA: 1s - 16ms/st - loss: 0.0023 - ETA: 1s - 16ms/step   - loss: 9.3699e-04 - ETA: 1s - 16ms/st - loss: 0.0013 - ETA: 1s - 16ms/step   - loss: 8.0502e-04 - ETA: 1s - 16ms/st - loss: 7.9187e-04 - ETA: 1s - 16ms/st - loss: 6.2266e-04 - ETA: 0s - 16ms/st - loss: 3.5724e-04 - ETA: 0s - 16ms/st - loss: 5.8246e-04 - ETA: 0s - 16ms/st - loss: 1.9947e-04 - ETA: 0s - 16ms/st - loss: 0.0039 - ETA: 0s - 16ms/step   - loss: 0.0012 - ETA: 0s - 16ms/st - loss: 4.4568e-04 - ETA: 0s - 16ms/st - loss: 0.0027 - 16ms/step              \n",
      "Epoch 78/100\n",
      "step 341/341 [==============================] - loss: 0.0011 - ETA: 5s - 16ms/st - loss: 0.0016 - ETA: 5s - 16ms/st - loss: 0.0045 - ETA: 4s - 16ms/st - loss: 9.7417e-04 - ETA: 4s - 16ms/st - loss: 0.0023 - ETA: 4s - 16ms/step   - loss: 8.8093e-04 - ETA: 4s - 16ms/st - loss: 0.0013 - ETA: 4s - 16ms/step   - loss: 0.0027 - ETA: 4s - 16ms/st - loss: 0.0010 - ETA: 3s - 16ms/st - loss: 3.5119e-04 - ETA: 3s - 16ms/st - loss: 0.0012 - ETA: 3s - 16ms/step   - loss: 3.9842e-04 - ETA: 3s - 16ms/st - loss: 4.0249e-04 - ETA: 3s - 16ms/st - loss: 9.8304e-04 - ETA: 3s - 16ms/st - loss: 5.9052e-04 - ETA: 2s - 16ms/st - loss: 7.9784e-04 - ETA: 2s - 16ms/st - loss: 3.2394e-04 - ETA: 2s - 16ms/st - loss: 5.0945e-04 - ETA: 2s - 16ms/st - loss: 6.2907e-04 - ETA: 2s - 16ms/st - loss: 0.0012 - ETA: 2s - 16ms/step   - loss: 7.0491e-04 - ETA: 2s - 16ms/st - loss: 0.0013 - ETA: 1s - 16ms/step   - loss: 0.0011 - ETA: 1s - 16ms/st - loss: 5.5399e-04 - ETA: 1s - 16ms/st - loss: 9.3082e-04 - ETA: 1s - 16ms/st - loss: 5.2748e-04 - ETA: 1s - 16ms/st - loss: 5.2085e-04 - ETA: 1s - 16ms/st - loss: 3.2864e-04 - ETA: 0s - 16ms/st - loss: 7.9529e-04 - ETA: 0s - 16ms/st - loss: 5.1166e-04 - ETA: 0s - 16ms/st - loss: 0.0013 - ETA: 0s - 16ms/step   - loss: 5.2945e-04 - ETA: 0s - 16ms/st - loss: 0.0018 - ETA: 0s - 16ms/step   - loss: 0.0011 - ETA: 0s - 16ms/st - loss: 0.0021 - 16ms/step          \n",
      "Epoch 79/100\n",
      "step 341/341 [==============================] - loss: 0.0020 - ETA: 5s - 16ms/st - loss: 3.5632e-04 - ETA: 4s - 15ms/st - loss: 5.8990e-04 - ETA: 4s - 16ms/st - loss: 5.5421e-04 - ETA: 4s - 16ms/st - loss: 0.0010 - ETA: 4s - 16ms/step   - loss: 9.6051e-04 - ETA: 4s - 16ms/st - loss: 0.0011 - ETA: 4s - 16ms/step   - loss: 0.0011 - ETA: 4s - 16ms/st - loss: 1.8414e-04 - ETA: 3s - 16ms/st - loss: 0.0013 - ETA: 3s - 16ms/step   - loss: 0.0034 - ETA: 3s - 16ms/st - loss: 0.0019 - ETA: 3s - 16ms/st - loss: 9.9818e-04 - ETA: 3s - 16ms/st - loss: 6.7438e-04 - ETA: 3s - 16ms/st - loss: 0.0018 - ETA: 2s - 16ms/step   - loss: 5.5064e-04 - ETA: 2s - 16ms/st - loss: 3.5128e-04 - ETA: 2s - 16ms/st - loss: 0.0024 - ETA: 2s - 16ms/step   - loss: 3.2528e-04 - ETA: 2s - 16ms/st - loss: 5.4359e-04 - ETA: 2s - 16ms/st - loss: 8.2815e-04 - ETA: 2s - 16ms/st - loss: 9.2147e-04 - ETA: 1s - 16ms/st - loss: 8.9641e-04 - ETA: 1s - 16ms/st - loss: 4.6580e-04 - ETA: 1s - 16ms/st - loss: 9.0389e-04 - ETA: 1s - 16ms/st - loss: 0.0017 - ETA: 1s - 16ms/step   - loss: 0.0012 - ETA: 1s - 16ms/st - loss: 7.6294e-04 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 7.3968e-04 - ETA: 0s - 16ms/st - loss: 4.1416e-04 - ETA: 0s - 16ms/st - loss: 4.7292e-04 - ETA: 0s - 16ms/st - loss: 0.0020 - ETA: 0s - 16ms/step   - loss: 5.8878e-04 - ETA: 0s - 16ms/st - loss: 6.8709e-04 - 16ms/step          \n",
      "Epoch 80/100\n",
      "step 341/341 [==============================] - loss: 0.0011 - ETA: 5s - 15ms/st - loss: 0.0013 - ETA: 4s - 15ms/st - loss: 0.0016 - ETA: 4s - 15ms/st - loss: 6.1566e-04 - ETA: 4s - 15ms/st - loss: 7.5138e-04 - ETA: 4s - 15ms/st - loss: 7.6755e-04 - ETA: 4s - 15ms/st - loss: 7.3670e-04 - ETA: 4s - 15ms/st - loss: 0.0011 - ETA: 4s - 15ms/step   - loss: 6.9416e-04 - ETA: 3s - 15ms/st - loss: 0.0015 - ETA: 3s - 15ms/step   - loss: 4.6028e-04 - ETA: 3s - 15ms/st - loss: 6.9947e-04 - ETA: 3s - 15ms/st - loss: 0.0041 - ETA: 3s - 15ms/step   - loss: 0.0011 - ETA: 3s - 15ms/st - loss: 0.0032 - ETA: 2s - 15ms/st - loss: 8.9034e-04 - ETA: 2s - 15ms/st - loss: 0.0026 - ETA: 2s - 15ms/step   - loss: 4.2301e-04 - ETA: 2s - 15ms/st - loss: 6.3941e-04 - ETA: 2s - 16ms/st - loss: 8.1255e-04 - ETA: 2s - 16ms/st - loss: 0.0024 - ETA: 2s - 16ms/step   - loss: 0.0027 - ETA: 1s - 16ms/st - loss: 9.5561e-04 - ETA: 1s - 16ms/st - loss: 0.0011 - ETA: 1s - 16ms/step   - loss: 3.3845e-04 - ETA: 1s - 16ms/st - loss: 0.0019 - ETA: 1s - 16ms/step   - loss: 6.6601e-04 - ETA: 1s - 16ms/st - loss: 4.5372e-04 - ETA: 0s - 16ms/st - loss: 5.3860e-04 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 0.0011 - ETA: 0s - 16ms/st - loss: 8.7529e-04 - ETA: 0s - 16ms/st - loss: 1.8610e-04 - ETA: 0s - 16ms/st - loss: 0.0083 - ETA: 0s - 16ms/step   - loss: 7.7314e-04 - 16ms/step      \n",
      "Epoch 81/100\n",
      "step 341/341 [==============================] - loss: 0.0010 - ETA: 5s - 16ms/st - loss: 6.1891e-04 - ETA: 5s - 16ms/st - loss: 0.0050 - ETA: 4s - 16ms/step   - loss: 0.0015 - ETA: 4s - 16ms/st - loss: 4.9494e-04 - ETA: 4s - 16ms/st - loss: 5.9993e-04 - ETA: 4s - 16ms/st - loss: 3.9895e-04 - ETA: 4s - 16ms/st - loss: 0.0039 - ETA: 4s - 16ms/step   - loss: 0.0012 - ETA: 3s - 16ms/st - loss: 0.0010 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 3s - 16ms/st - loss: 0.0020 - ETA: 3s - 16ms/st - loss: 6.3991e-04 - ETA: 3s - 16ms/st - loss: 0.0017 - ETA: 3s - 16ms/step   - loss: 7.5406e-04 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 0.0011 - ETA: 2s - 16ms/st - loss: 0.0014 - ETA: 2s - 16ms/st - loss: 4.2270e-04 - ETA: 2s - 16ms/st - loss: 2.6408e-04 - ETA: 2s - 16ms/st - loss: 2.5353e-04 - ETA: 2s - 16ms/st - loss: 6.8915e-04 - ETA: 1s - 16ms/st - loss: 4.6346e-04 - ETA: 1s - 16ms/st - loss: 6.8162e-04 - ETA: 1s - 16ms/st - loss: 6.7837e-04 - ETA: 1s - 16ms/st - loss: 0.0017 - ETA: 1s - 16ms/step   - loss: 2.8854e-04 - ETA: 1s - 16ms/st - loss: 7.6100e-04 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 9.6410e-04 - ETA: 0s - 16ms/st - loss: 3.9328e-04 - ETA: 0s - 16ms/st - loss: 0.0016 - ETA: 0s - 16ms/step   - loss: 3.1586e-04 - ETA: 0s - 16ms/st - loss: 0.0015 - ETA: 0s - 16ms/step   - loss: 4.7620e-04 - 16ms/step      \n",
      "save checkpoint at /home/aistudio/save_dir/80\n",
      "Epoch 82/100\n",
      "step 341/341 [==============================] - loss: 7.9093e-04 - ETA: 5s - 16ms/st - loss: 3.6008e-04 - ETA: 5s - 16ms/st - loss: 0.0046 - ETA: 4s - 16ms/step   - loss: 9.5971e-04 - ETA: 4s - 16ms/st - loss: 0.0011 - ETA: 4s - 16ms/step   - loss: 0.0011 - ETA: 4s - 16ms/st - loss: 6.9258e-04 - ETA: 4s - 16ms/st - loss: 9.1896e-04 - ETA: 4s - 16ms/st - loss: 3.3239e-04 - ETA: 4s - 16ms/st - loss: 6.1165e-04 - ETA: 3s - 16ms/st - loss: 8.0771e-04 - ETA: 3s - 16ms/st - loss: 9.3853e-04 - ETA: 3s - 16ms/st - loss: 7.5234e-04 - ETA: 3s - 16ms/st - loss: 5.9555e-04 - ETA: 3s - 16ms/st - loss: 6.5364e-04 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 0.0074 - ETA: 2s - 16ms/st - loss: 4.0399e-04 - ETA: 2s - 16ms/st - loss: 0.0012 - ETA: 2s - 16ms/step   - loss: 0.0016 - ETA: 2s - 16ms/st - loss: 8.0665e-04 - ETA: 2s - 16ms/st - loss: 3.4595e-04 - ETA: 1s - 16ms/st - loss: 0.0013 - ETA: 1s - 16ms/step   - loss: 8.1433e-04 - ETA: 1s - 16ms/st - loss: 6.3904e-04 - ETA: 1s - 16ms/st - loss: 2.3749e-04 - ETA: 1s - 16ms/st - loss: 0.0014 - ETA: 1s - 16ms/step   - loss: 0.0010 - ETA: 0s - 16ms/st - loss: 0.0010 - ETA: 0s - 16ms/st - loss: 4.7277e-04 - ETA: 0s - 16ms/st - loss: 0.0014 - ETA: 0s - 16ms/step   - loss: 5.8294e-04 - ETA: 0s - 16ms/st - loss: 4.8306e-04 - ETA: 0s - 16ms/st - loss: 0.0022 - ETA: 0s - 16ms/step   - loss: 8.3967e-04 - 16ms/step      \n",
      "Epoch 83/100\n",
      "step 341/341 [==============================] - loss: 5.9108e-04 - ETA: 5s - 16ms/st - loss: 0.0013 - ETA: 5s - 16ms/step   - loss: 0.0016 - ETA: 4s - 16ms/st - loss: 0.0022 - ETA: 4s - 16ms/st - loss: 6.6414e-04 - ETA: 4s - 16ms/st - loss: 8.6649e-04 - ETA: 4s - 16ms/st - loss: 0.0024 - ETA: 4s - 16ms/step   - loss: 0.0014 - ETA: 4s - 16ms/st - loss: 8.4820e-04 - ETA: 3s - 16ms/st - loss: 2.7214e-04 - ETA: 3s - 16ms/st - loss: 3.7551e-04 - ETA: 3s - 16ms/st - loss: 5.0720e-04 - ETA: 3s - 16ms/st - loss: 9.7184e-04 - ETA: 3s - 16ms/st - loss: 8.5622e-04 - ETA: 3s - 16ms/st - loss: 8.1487e-04 - ETA: 3s - 16ms/st - loss: 0.0030 - ETA: 2s - 16ms/step   - loss: 4.8206e-04 - ETA: 2s - 16ms/st - loss: 0.0010 - ETA: 2s - 16ms/step   - loss: 6.1337e-04 - ETA: 2s - 16ms/st - loss: 0.0010 - ETA: 2s - 16ms/step   - loss: 5.1034e-04 - ETA: 2s - 16ms/st - loss: 7.6677e-04 - ETA: 1s - 16ms/st - loss: 4.9596e-04 - ETA: 1s - 16ms/st - loss: 7.6269e-04 - ETA: 1s - 16ms/st - loss: 6.7260e-04 - ETA: 1s - 16ms/st - loss: 6.9254e-04 - ETA: 1s - 16ms/st - loss: 3.9020e-04 - ETA: 1s - 16ms/st - loss: 7.7905e-04 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 4.3467e-04 - ETA: 0s - 16ms/st - loss: 9.8862e-04 - ETA: 0s - 16ms/st - loss: 6.7095e-04 - ETA: 0s - 16ms/st - loss: 3.1118e-04 - ETA: 0s - 16ms/st - loss: 8.1702e-04 - ETA: 0s - 16ms/st - loss: 0.0014 - 16ms/step              \n",
      "Epoch 84/100\n",
      "step 341/341 [==============================] - loss: 0.0014 - ETA: 5s - 16ms/st - loss: 0.0020 - ETA: 5s - 16ms/st - loss: 0.0012 - ETA: 4s - 16ms/st - loss: 0.0016 - ETA: 4s - 16ms/st - loss: 0.0012 - ETA: 4s - 16ms/st - loss: 2.5803e-04 - ETA: 4s - 16ms/st - loss: 4.0071e-04 - ETA: 4s - 16ms/st - loss: 8.2001e-04 - ETA: 4s - 16ms/st - loss: 0.0013 - ETA: 3s - 16ms/step   - loss: 6.5157e-04 - ETA: 3s - 16ms/st - loss: 3.0111e-04 - ETA: 3s - 16ms/st - loss: 4.8581e-04 - ETA: 3s - 16ms/st - loss: 0.0014 - ETA: 3s - 16ms/step   - loss: 3.6270e-04 - ETA: 3s - 16ms/st - loss: 0.0012 - ETA: 3s - 16ms/step   - loss: 0.0015 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/st - loss: 9.2712e-04 - ETA: 2s - 16ms/st - loss: 0.0017 - ETA: 2s - 16ms/step   - loss: 6.7621e-04 - ETA: 2s - 16ms/st - loss: 0.0029 - ETA: 2s - 16ms/step   - loss: 6.9226e-04 - ETA: 1s - 16ms/st - loss: 0.0016 - ETA: 1s - 16ms/step   - loss: 0.0010 - ETA: 1s - 16ms/st - loss: 0.0012 - ETA: 1s - 16ms/st - loss: 0.0013 - ETA: 1s - 16ms/st - loss: 0.0017 - ETA: 1s - 16ms/st - loss: 7.9975e-04 - ETA: 0s - 16ms/st - loss: 7.7519e-04 - ETA: 0s - 16ms/st - loss: 6.8005e-04 - ETA: 0s - 16ms/st - loss: 7.6519e-04 - ETA: 0s - 16ms/st - loss: 9.7254e-04 - ETA: 0s - 16ms/st - loss: 9.5737e-04 - ETA: 0s - 16ms/st - loss: 0.0023 - ETA: 0s - 16ms/step   - loss: 0.0013 - 16ms/step          \n",
      "Epoch 85/100\n",
      "step 341/341 [==============================] - loss: 8.3480e-04 - ETA: 5s - 16ms/st - loss: 5.5707e-04 - ETA: 4s - 16ms/st - loss: 2.4182e-04 - ETA: 4s - 16ms/st - loss: 0.0020 - ETA: 4s - 16ms/step   - loss: 7.2752e-04 - ETA: 4s - 16ms/st - loss: 0.0012 - ETA: 4s - 16ms/step   - loss: 5.1091e-04 - ETA: 4s - 16ms/st - loss: 7.4402e-04 - ETA: 4s - 16ms/st - loss: 8.8287e-04 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 3s - 16ms/step   - loss: 0.0021 - ETA: 3s - 16ms/st - loss: 0.0013 - ETA: 3s - 16ms/st - loss: 5.6311e-04 - ETA: 3s - 16ms/st - loss: 7.7299e-04 - ETA: 3s - 16ms/st - loss: 4.8658e-04 - ETA: 2s - 16ms/st - loss: 9.9935e-04 - ETA: 2s - 16ms/st - loss: 0.0015 - ETA: 2s - 16ms/step   - loss: 6.1890e-04 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 7.7123e-04 - ETA: 2s - 16ms/st - loss: 7.0121e-04 - ETA: 2s - 16ms/st - loss: 5.1015e-04 - ETA: 1s - 16ms/st - loss: 4.7198e-04 - ETA: 1s - 16ms/st - loss: 9.1458e-04 - ETA: 1s - 16ms/st - loss: 3.3206e-04 - ETA: 1s - 16ms/st - loss: 0.0021 - ETA: 1s - 16ms/step   - loss: 0.0022 - ETA: 1s - 16ms/st - loss: 9.1267e-04 - ETA: 0s - 16ms/st - loss: 5.5731e-04 - ETA: 0s - 16ms/st - loss: 7.1525e-04 - ETA: 0s - 16ms/st - loss: 0.0015 - ETA: 0s - 16ms/step   - loss: 0.0011 - ETA: 0s - 16ms/st - loss: 3.5310e-04 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/step   - loss: 0.0029 - 16ms/step          \n",
      "Epoch 86/100\n",
      "step 341/341 [==============================] - loss: 0.0019 - ETA: 5s - 15ms/st - loss: 0.0021 - ETA: 4s - 15ms/st - loss: 0.0017 - ETA: 4s - 15ms/st - loss: 8.4059e-04 - ETA: 4s - 15ms/st - loss: 6.7684e-04 - ETA: 4s - 16ms/st - loss: 0.0023 - ETA: 4s - 16ms/step   - loss: 5.3606e-04 - ETA: 4s - 16ms/st - loss: 0.0017 - ETA: 4s - 16ms/step   - loss: 0.0012 - ETA: 3s - 16ms/st - loss: 5.2414e-04 - ETA: 3s - 16ms/st - loss: 7.3961e-04 - ETA: 3s - 16ms/st - loss: 0.0030 - ETA: 3s - 16ms/step   - loss: 7.9092e-04 - ETA: 3s - 16ms/st - loss: 4.7664e-04 - ETA: 3s - 16ms/st - loss: 7.4437e-04 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 6.2613e-04 - ETA: 2s - 16ms/st - loss: 0.0026 - ETA: 2s - 16ms/step   - loss: 0.0011 - ETA: 2s - 16ms/st - loss: 8.7581e-04 - ETA: 2s - 16ms/st - loss: 8.6032e-04 - ETA: 2s - 16ms/st - loss: 0.0018 - ETA: 1s - 16ms/step   - loss: 0.0016 - ETA: 1s - 16ms/st - loss: 0.0011 - ETA: 1s - 16ms/st - loss: 0.0012 - ETA: 1s - 16ms/st - loss: 4.5291e-04 - ETA: 1s - 16ms/st - loss: 7.9096e-04 - ETA: 1s - 16ms/st - loss: 8.9183e-04 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/step   - loss: 5.8720e-04 - ETA: 0s - 16ms/st - loss: 9.2926e-04 - ETA: 0s - 16ms/st - loss: 0.0013 - ETA: 0s - 16ms/step   - loss: 0.0011 - ETA: 0s - 16ms/st - loss: 7.8883e-04 - ETA: 0s - 16ms/st - loss: 0.0013 - 16ms/step              \n",
      "Epoch 87/100\n",
      "step 341/341 [==============================] - loss: 0.0011 - ETA: 5s - 16ms/st - loss: 0.0024 - ETA: 5s - 16ms/st - loss: 0.0019 - ETA: 4s - 16ms/st - loss: 3.8079e-04 - ETA: 4s - 16ms/st - loss: 0.0014 - ETA: 4s - 16ms/step   - loss: 7.1564e-04 - ETA: 4s - 16ms/st - loss: 3.6537e-04 - ETA: 4s - 16ms/st - loss: 0.0014 - ETA: 4s - 16ms/step   - loss: 5.7452e-04 - ETA: 3s - 16ms/st - loss: 0.0012 - ETA: 3s - 16ms/step   - loss: 2.3643e-04 - ETA: 3s - 16ms/st - loss: 4.8491e-04 - ETA: 3s - 16ms/st - loss: 0.0040 - ETA: 3s - 16ms/step   - loss: 7.5769e-04 - ETA: 3s - 16ms/st - loss: 9.6608e-04 - ETA: 3s - 16ms/st - loss: 7.3015e-04 - ETA: 2s - 16ms/st - loss: 5.5382e-04 - ETA: 2s - 16ms/st - loss: 6.7860e-04 - ETA: 2s - 16ms/st - loss: 7.7949e-04 - ETA: 2s - 16ms/st - loss: 7.2887e-04 - ETA: 2s - 16ms/st - loss: 0.0010 - ETA: 2s - 16ms/step   - loss: 5.3518e-04 - ETA: 1s - 16ms/st - loss: 7.6189e-04 - ETA: 1s - 16ms/st - loss: 0.0010 - ETA: 1s - 16ms/step   - loss: 0.0014 - ETA: 1s - 16ms/st - loss: 4.1824e-04 - ETA: 1s - 16ms/st - loss: 7.6380e-04 - ETA: 1s - 16ms/st - loss: 0.0018 - ETA: 0s - 16ms/step   - loss: 0.0011 - ETA: 0s - 16ms/st - loss: 7.1178e-04 - ETA: 0s - 16ms/st - loss: 4.7437e-04 - ETA: 0s - 16ms/st - loss: 5.0521e-04 - ETA: 0s - 16ms/st - loss: 0.0013 - ETA: 0s - 16ms/step   - loss: 0.0011 - ETA: 0s - 16ms/st - loss: 0.0026 - 16ms/step          \n",
      "Epoch 88/100\n",
      "step 341/341 [==============================] - loss: 6.7676e-04 - ETA: 5s - 16ms/st - loss: 0.0015 - ETA: 5s - 16ms/step   - loss: 6.5566e-04 - ETA: 4s - 16ms/st - loss: 4.8923e-04 - ETA: 4s - 16ms/st - loss: 6.2668e-04 - ETA: 4s - 16ms/st - loss: 0.0012 - ETA: 4s - 16ms/step   - loss: 5.0983e-04 - ETA: 4s - 16ms/st - loss: 0.0015 - ETA: 4s - 16ms/step   - loss: 0.0019 - ETA: 3s - 16ms/st - loss: 5.3009e-04 - ETA: 3s - 16ms/st - loss: 7.8574e-04 - ETA: 3s - 16ms/st - loss: 1.8664e-04 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 3s - 16ms/step   - loss: 5.3665e-04 - ETA: 3s - 16ms/st - loss: 4.1621e-04 - ETA: 3s - 16ms/st - loss: 0.0015 - ETA: 2s - 16ms/step   - loss: 5.7085e-04 - ETA: 2s - 16ms/st - loss: 4.0443e-04 - ETA: 2s - 16ms/st - loss: 4.8470e-04 - ETA: 2s - 16ms/st - loss: 8.3971e-04 - ETA: 2s - 16ms/st - loss: 8.6699e-04 - ETA: 2s - 16ms/st - loss: 4.6312e-04 - ETA: 1s - 16ms/st - loss: 0.0021 - ETA: 1s - 16ms/step   - loss: 0.0019 - ETA: 1s - 16ms/st - loss: 0.0020 - ETA: 1s - 16ms/st - loss: 8.6959e-04 - ETA: 1s - 16ms/st - loss: 7.5839e-04 - ETA: 1s - 16ms/st - loss: 3.9140e-04 - ETA: 0s - 16ms/st - loss: 3.2037e-04 - ETA: 0s - 16ms/st - loss: 4.5127e-04 - ETA: 0s - 16ms/st - loss: 2.1873e-04 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 0.0029 - ETA: 0s - 16ms/st - loss: 0.0062 - ETA: 0s - 16ms/st - loss: 0.0011 - 16ms/step          \n",
      "Epoch 89/100\n",
      "step 341/341 [==============================] - loss: 5.7649e-04 - ETA: 5s - 16ms/st - loss: 0.0012 - ETA: 5s - 16ms/step   - loss: 2.5186e-04 - ETA: 4s - 16ms/st - loss: 0.0014 - ETA: 4s - 16ms/step   - loss: 8.2166e-04 - ETA: 4s - 16ms/st - loss: 4.5062e-04 - ETA: 4s - 16ms/st - loss: 0.0012 - ETA: 4s - 16ms/step   - loss: 0.0103 - ETA: 4s - 16ms/st - loss: 7.9861e-04 - ETA: 3s - 16ms/st - loss: 7.8723e-04 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 3s - 16ms/step   - loss: 1.7028e-04 - ETA: 3s - 16ms/st - loss: 0.0020 - ETA: 3s - 16ms/step   - loss: 5.5136e-04 - ETA: 3s - 16ms/st - loss: 7.3194e-04 - ETA: 3s - 16ms/st - loss: 2.2028e-04 - ETA: 2s - 16ms/st - loss: 0.0027 - ETA: 2s - 16ms/step   - loss: 0.0013 - ETA: 2s - 16ms/st - loss: 9.1319e-04 - ETA: 2s - 16ms/st - loss: 3.9256e-04 - ETA: 2s - 16ms/st - loss: 3.9322e-04 - ETA: 2s - 16ms/st - loss: 4.7376e-04 - ETA: 1s - 16ms/st - loss: 4.6404e-04 - ETA: 1s - 16ms/st - loss: 7.8358e-04 - ETA: 1s - 16ms/st - loss: 4.0986e-04 - ETA: 1s - 16ms/st - loss: 0.0013 - ETA: 1s - 16ms/step   - loss: 0.0012 - ETA: 1s - 16ms/st - loss: 6.2950e-04 - ETA: 0s - 16ms/st - loss: 8.1599e-04 - ETA: 0s - 16ms/st - loss: 0.0010 - ETA: 0s - 16ms/step   - loss: 4.4955e-04 - ETA: 0s - 16ms/st - loss: 6.6518e-04 - ETA: 0s - 16ms/st - loss: 0.0017 - ETA: 0s - 16ms/step   - loss: 0.0018 - ETA: 0s - 16ms/st - loss: 0.0012 - 16ms/step          \n",
      "Epoch 90/100\n",
      "step 341/341 [==============================] - loss: 0.0033 - ETA: 5s - 16ms/st - loss: 0.0019 - ETA: 5s - 16ms/st - loss: 4.2562e-04 - ETA: 4s - 16ms/st - loss: 3.4125e-04 - ETA: 4s - 16ms/st - loss: 0.0022 - ETA: 4s - 16ms/step   - loss: 0.0016 - ETA: 4s - 16ms/st - loss: 8.6715e-04 - ETA: 4s - 16ms/st - loss: 0.0017 - ETA: 4s - 16ms/step   - loss: 0.0013 - ETA: 3s - 16ms/st - loss: 0.0024 - ETA: 3s - 16ms/st - loss: 0.0014 - ETA: 3s - 16ms/st - loss: 9.9601e-04 - ETA: 3s - 16ms/st - loss: 5.6067e-04 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 3s - 16ms/step   - loss: 0.0011 - ETA: 3s - 16ms/st - loss: 8.4168e-04 - ETA: 2s - 16ms/st - loss: 3.9562e-04 - ETA: 2s - 16ms/st - loss: 9.5373e-04 - ETA: 2s - 16ms/st - loss: 4.2133e-04 - ETA: 2s - 16ms/st - loss: 0.0010 - ETA: 2s - 16ms/step   - loss: 7.2637e-04 - ETA: 2s - 16ms/st - loss: 2.6533e-04 - ETA: 1s - 16ms/st - loss: 3.4886e-04 - ETA: 1s - 16ms/st - loss: 0.0019 - ETA: 1s - 16ms/step   - loss: 0.0014 - ETA: 1s - 16ms/st - loss: 7.0391e-04 - ETA: 1s - 16ms/st - loss: 8.3364e-04 - ETA: 1s - 16ms/st - loss: 0.0014 - ETA: 0s - 16ms/step   - loss: 6.2978e-04 - ETA: 0s - 16ms/st - loss: 0.0023 - ETA: 0s - 16ms/step   - loss: 3.3784e-04 - ETA: 0s - 16ms/st - loss: 7.9630e-04 - ETA: 0s - 16ms/st - loss: 5.8021e-04 - ETA: 0s - 16ms/st - loss: 5.3156e-04 - ETA: 0s - 16ms/st - loss: 0.0025 - 16ms/step              \n",
      "Epoch 91/100\n",
      "step 341/341 [==============================] - loss: 9.6856e-04 - ETA: 5s - 16ms/st - loss: 0.0011 - ETA: 5s - 16ms/step   - loss: 5.7544e-04 - ETA: 4s - 16ms/st - loss: 0.0021 - ETA: 4s - 16ms/step   - loss: 0.0011 - ETA: 4s - 16ms/st - loss: 5.0450e-04 - ETA: 4s - 16ms/st - loss: 4.6576e-04 - ETA: 4s - 16ms/st - loss: 3.3922e-04 - ETA: 4s - 16ms/st - loss: 6.6005e-04 - ETA: 3s - 16ms/st - loss: 6.6666e-04 - ETA: 3s - 16ms/st - loss: 8.3620e-04 - ETA: 3s - 16ms/st - loss: 8.6003e-04 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 3s - 16ms/step   - loss: 4.5867e-04 - ETA: 3s - 16ms/st - loss: 5.8014e-04 - ETA: 3s - 16ms/st - loss: 7.6184e-04 - ETA: 2s - 16ms/st - loss: 8.9554e-04 - ETA: 2s - 16ms/st - loss: 1.8040e-04 - ETA: 2s - 16ms/st - loss: 2.7580e-04 - ETA: 2s - 16ms/st - loss: 0.0023 - ETA: 2s - 16ms/step   - loss: 8.2206e-04 - ETA: 2s - 16ms/st - loss: 0.0012 - ETA: 1s - 16ms/step   - loss: 0.0028 - ETA: 1s - 16ms/st - loss: 0.0039 - ETA: 1s - 16ms/st - loss: 0.0016 - ETA: 1s - 16ms/st - loss: 0.0072 - ETA: 1s - 16ms/st - loss: 0.0019 - ETA: 1s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/st - loss: 0.0019 - ETA: 0s - 16ms/st - loss: 0.0014 - ETA: 0s - 16ms/st - loss: 5.6293e-04 - ETA: 0s - 16ms/st - loss: 0.0015 - ETA: 0s - 16ms/step   - loss: 0.0010 - ETA: 0s - 16ms/st - loss: 0.0023 - ETA: 0s - 16ms/st - loss: 7.6110e-04 - 16ms/step      \n",
      "save checkpoint at /home/aistudio/save_dir/90\n",
      "Epoch 92/100\n",
      "step 341/341 [==============================] - loss: 3.3061e-04 - ETA: 5s - 16ms/st - loss: 3.1398e-04 - ETA: 5s - 16ms/st - loss: 0.0033 - ETA: 4s - 16ms/step   - loss: 7.7283e-04 - ETA: 4s - 16ms/st - loss: 0.0018 - ETA: 4s - 16ms/step   - loss: 9.8475e-04 - ETA: 4s - 16ms/st - loss: 8.5514e-04 - ETA: 4s - 16ms/st - loss: 7.7312e-04 - ETA: 4s - 16ms/st - loss: 6.6145e-04 - ETA: 3s - 16ms/st - loss: 0.0027 - ETA: 3s - 16ms/step   - loss: 2.6084e-04 - ETA: 3s - 16ms/st - loss: 0.0022 - ETA: 3s - 16ms/step   - loss: 9.1600e-04 - ETA: 3s - 16ms/st - loss: 6.4601e-04 - ETA: 3s - 16ms/st - loss: 8.5095e-04 - ETA: 3s - 16ms/st - loss: 0.0018 - ETA: 2s - 16ms/step   - loss: 6.3099e-04 - ETA: 2s - 16ms/st - loss: 0.0013 - ETA: 2s - 16ms/step   - loss: 0.0017 - ETA: 2s - 16ms/st - loss: 4.1261e-04 - ETA: 2s - 16ms/st - loss: 3.3493e-04 - ETA: 2s - 16ms/st - loss: 0.0013 - ETA: 1s - 16ms/step   - loss: 6.4855e-04 - ETA: 1s - 16ms/st - loss: 0.0013 - ETA: 1s - 16ms/step   - loss: 8.9530e-04 - ETA: 1s - 16ms/st - loss: 6.7071e-04 - ETA: 1s - 16ms/st - loss: 0.0034 - ETA: 1s - 16ms/step   - loss: 0.0010 - ETA: 0s - 16ms/st - loss: 5.5007e-04 - ETA: 0s - 16ms/st - loss: 0.0020 - ETA: 0s - 16ms/step   - loss: 7.0083e-04 - ETA: 0s - 16ms/st - loss: 0.0010 - ETA: 0s - 16ms/step   - loss: 7.1976e-04 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 0.0022 - 16ms/step          \n",
      "Epoch 93/100\n",
      "step 341/341 [==============================] - loss: 0.0012 - ETA: 5s - 16ms/st - loss: 0.0017 - ETA: 5s - 16ms/st - loss: 7.1446e-04 - ETA: 4s - 16ms/st - loss: 5.1865e-04 - ETA: 4s - 16ms/st - loss: 0.0018 - ETA: 4s - 16ms/step   - loss: 4.1849e-04 - ETA: 4s - 16ms/st - loss: 0.0021 - ETA: 4s - 16ms/step   - loss: 0.0019 - ETA: 4s - 16ms/st - loss: 5.0271e-04 - ETA: 3s - 16ms/st - loss: 6.5950e-04 - ETA: 3s - 16ms/st - loss: 0.0011 - ETA: 3s - 16ms/step   - loss: 5.1551e-04 - ETA: 3s - 16ms/st - loss: 6.0069e-04 - ETA: 3s - 16ms/st - loss: 0.0021 - ETA: 3s - 16ms/step   - loss: 4.4614e-04 - ETA: 3s - 16ms/st - loss: 7.1697e-04 - ETA: 2s - 16ms/st - loss: 0.0010 - ETA: 2s - 16ms/step   - loss: 6.1613e-04 - ETA: 2s - 16ms/st - loss: 5.8611e-04 - ETA: 2s - 16ms/st - loss: 0.0013 - ETA: 2s - 16ms/step   - loss: 0.0016 - ETA: 2s - 16ms/st - loss: 0.0027 - ETA: 1s - 16ms/st - loss: 0.0010 - ETA: 1s - 16ms/st - loss: 0.0010 - ETA: 1s - 16ms/st - loss: 0.0024 - ETA: 1s - 16ms/st - loss: 0.0011 - ETA: 1s - 16ms/st - loss: 2.6326e-04 - ETA: 1s - 16ms/st - loss: 3.0990e-04 - ETA: 0s - 16ms/st - loss: 5.1796e-04 - ETA: 0s - 16ms/st - loss: 3.3549e-04 - ETA: 0s - 16ms/st - loss: 0.0019 - ETA: 0s - 16ms/step   - loss: 0.0032 - ETA: 0s - 16ms/st - loss: 0.0018 - ETA: 0s - 16ms/st - loss: 4.3363e-04 - ETA: 0s - 16ms/st - loss: 8.3371e-04 - 16ms/step          \n",
      "Epoch 94/100\n",
      "step 341/341 [==============================] - loss: 0.0022 - ETA: 5s - 16ms/st - loss: 0.0011 - ETA: 5s - 16ms/st - loss: 0.0016 - ETA: 5s - 16ms/st - loss: 9.8957e-04 - ETA: 4s - 16ms/st - loss: 0.0014 - ETA: 4s - 16ms/step   - loss: 0.0014 - ETA: 4s - 16ms/st - loss: 0.0012 - ETA: 4s - 16ms/st - loss: 5.5876e-04 - ETA: 4s - 16ms/st - loss: 1.9081e-04 - ETA: 4s - 16ms/st - loss: 6.3778e-04 - ETA: 3s - 16ms/st - loss: 6.1702e-04 - ETA: 3s - 16ms/st - loss: 4.6071e-04 - ETA: 3s - 16ms/st - loss: 5.8200e-04 - ETA: 3s - 16ms/st - loss: 3.7146e-04 - ETA: 3s - 16ms/st - loss: 5.6215e-04 - ETA: 3s - 16ms/st - loss: 4.8593e-04 - ETA: 2s - 16ms/st - loss: 0.0013 - ETA: 2s - 16ms/step   - loss: 0.0010 - ETA: 2s - 16ms/st - loss: 8.4631e-04 - ETA: 2s - 16ms/st - loss: 6.3597e-04 - ETA: 2s - 16ms/st - loss: 8.1509e-04 - ETA: 2s - 16ms/st - loss: 0.0018 - ETA: 1s - 16ms/step   - loss: 0.0012 - ETA: 1s - 16ms/st - loss: 6.8865e-04 - ETA: 1s - 16ms/st - loss: 8.3647e-04 - ETA: 1s - 16ms/st - loss: 3.8382e-04 - ETA: 1s - 16ms/st - loss: 5.3852e-04 - ETA: 1s - 16ms/st - loss: 3.8060e-04 - ETA: 0s - 16ms/st - loss: 7.9960e-04 - ETA: 0s - 16ms/st - loss: 0.0013 - ETA: 0s - 16ms/step   - loss: 8.9764e-04 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 7.3526e-04 - ETA: 0s - 16ms/st - loss: 6.2424e-04 - ETA: 0s - 16ms/st - loss: 8.9137e-04 - 16ms/step          \n",
      "Epoch 95/100\n",
      "step 341/341 [==============================] - loss: 9.0173e-04 - ETA: 5s - 16ms/st - loss: 6.7719e-04 - ETA: 5s - 17ms/st - loss: 0.0012 - ETA: 5s - 17ms/step   - loss: 0.0024 - ETA: 4s - 16ms/st - loss: 0.0024 - ETA: 4s - 16ms/st - loss: 7.6640e-04 - ETA: 4s - 17ms/st - loss: 0.0011 - ETA: 4s - 17ms/step   - loss: 8.5506e-04 - ETA: 4s - 16ms/st - loss: 0.0012 - ETA: 4s - 16ms/step   - loss: 7.8536e-04 - ETA: 3s - 16ms/st - loss: 0.0012 - ETA: 3s - 16ms/step   - loss: 0.0011 - ETA: 3s - 16ms/st - loss: 4.3630e-04 - ETA: 3s - 16ms/st - loss: 0.0020 - ETA: 3s - 16ms/step   - loss: 0.0012 - ETA: 3s - 16ms/st - loss: 0.0013 - ETA: 2s - 16ms/st - loss: 0.0044 - ETA: 2s - 16ms/st - loss: 7.7266e-04 - ETA: 2s - 16ms/st - loss: 3.3327e-04 - ETA: 2s - 16ms/st - loss: 5.0449e-04 - ETA: 2s - 16ms/st - loss: 4.4788e-04 - ETA: 2s - 16ms/st - loss: 3.6470e-04 - ETA: 1s - 16ms/st - loss: 6.0163e-04 - ETA: 1s - 16ms/st - loss: 0.0013 - ETA: 1s - 16ms/step   - loss: 0.0016 - ETA: 1s - 16ms/st - loss: 2.2921e-04 - ETA: 1s - 16ms/st - loss: 9.8539e-04 - ETA: 1s - 16ms/st - loss: 9.4939e-04 - ETA: 0s - 16ms/st - loss: 4.4992e-04 - ETA: 0s - 16ms/st - loss: 9.3413e-04 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/step   - loss: 6.9103e-04 - ETA: 0s - 16ms/st - loss: 7.1444e-04 - ETA: 0s - 16ms/st - loss: 7.0624e-04 - ETA: 0s - 16ms/st - loss: 8.3849e-04 - 16ms/step          \n",
      "Epoch 96/100\n",
      "step 341/341 [==============================] - loss: 0.0012 - ETA: 5s - 16ms/st - loss: 5.6154e-04 - ETA: 5s - 16ms/st - loss: 3.9925e-04 - ETA: 4s - 16ms/st - loss: 5.0409e-04 - ETA: 4s - 16ms/st - loss: 9.4368e-04 - ETA: 4s - 16ms/st - loss: 8.7150e-04 - ETA: 4s - 16ms/st - loss: 4.2821e-04 - ETA: 4s - 16ms/st - loss: 0.0016 - ETA: 4s - 16ms/step   - loss: 0.0012 - ETA: 3s - 16ms/st - loss: 6.9598e-04 - ETA: 3s - 16ms/st - loss: 7.0041e-04 - ETA: 3s - 16ms/st - loss: 0.0013 - ETA: 3s - 16ms/step   - loss: 7.7076e-04 - ETA: 3s - 16ms/st - loss: 4.2175e-04 - ETA: 3s - 16ms/st - loss: 0.0027 - ETA: 3s - 16ms/step   - loss: 4.1834e-04 - ETA: 2s - 16ms/st - loss: 5.7945e-04 - ETA: 2s - 16ms/st - loss: 8.8123e-04 - ETA: 2s - 16ms/st - loss: 0.0017 - ETA: 2s - 16ms/step   - loss: 6.6683e-04 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 7.5893e-04 - ETA: 1s - 16ms/st - loss: 4.2922e-04 - ETA: 1s - 16ms/st - loss: 8.1816e-04 - ETA: 1s - 16ms/st - loss: 8.0177e-04 - ETA: 1s - 16ms/st - loss: 0.0012 - ETA: 1s - 16ms/step   - loss: 8.1739e-04 - ETA: 1s - 16ms/st - loss: 7.6973e-04 - ETA: 0s - 16ms/st - loss: 7.2751e-04 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 7.0251e-04 - ETA: 0s - 16ms/st - loss: 3.8510e-04 - ETA: 0s - 16ms/st - loss: 8.0312e-04 - ETA: 0s - 16ms/st - loss: 0.0019 - ETA: 0s - 16ms/step   - loss: 8.9543e-04 - 16ms/step      \n",
      "Epoch 97/100\n",
      "step 341/341 [==============================] - loss: 6.5468e-04 - ETA: 5s - 16ms/st - loss: 7.7636e-04 - ETA: 5s - 16ms/st - loss: 0.0010 - ETA: 4s - 16ms/step   - loss: 8.9086e-04 - ETA: 4s - 16ms/st - loss: 9.1322e-04 - ETA: 4s - 16ms/st - loss: 0.0012 - ETA: 4s - 16ms/step   - loss: 0.0022 - ETA: 4s - 16ms/st - loss: 9.4754e-04 - ETA: 4s - 16ms/st - loss: 6.7848e-04 - ETA: 3s - 16ms/st - loss: 5.5356e-04 - ETA: 3s - 16ms/st - loss: 0.0016 - ETA: 3s - 16ms/step   - loss: 5.5768e-04 - ETA: 3s - 16ms/st - loss: 9.4914e-04 - ETA: 3s - 16ms/st - loss: 0.0022 - ETA: 3s - 16ms/step   - loss: 9.9309e-04 - ETA: 3s - 16ms/st - loss: 0.0020 - ETA: 2s - 16ms/step   - loss: 8.5547e-04 - ETA: 2s - 16ms/st - loss: 0.0020 - ETA: 2s - 16ms/step   - loss: 8.3693e-04 - ETA: 2s - 16ms/st - loss: 2.4572e-04 - ETA: 2s - 16ms/st - loss: 6.8526e-04 - ETA: 2s - 16ms/st - loss: 5.1881e-04 - ETA: 1s - 16ms/st - loss: 0.0011 - ETA: 1s - 16ms/step   - loss: 0.0014 - ETA: 1s - 16ms/st - loss: 0.0011 - ETA: 1s - 16ms/st - loss: 8.5267e-04 - ETA: 1s - 16ms/st - loss: 0.0013 - ETA: 1s - 16ms/step   - loss: 9.3175e-04 - ETA: 0s - 16ms/st - loss: 9.7626e-04 - ETA: 0s - 16ms/st - loss: 7.1310e-04 - ETA: 0s - 16ms/st - loss: 9.6765e-04 - ETA: 0s - 16ms/st - loss: 2.1769e-04 - ETA: 0s - 16ms/st - loss: 0.0018 - ETA: 0s - 16ms/step   - loss: 7.4593e-04 - ETA: 0s - 16ms/st - loss: 8.9995e-04 - 16ms/step          \n",
      "Epoch 98/100\n",
      "step 341/341 [==============================] - loss: 6.8268e-04 - ETA: 5s - 16ms/st - loss: 0.0022 - ETA: 5s - 16ms/step   - loss: 0.0012 - ETA: 4s - 16ms/st - loss: 6.2320e-04 - ETA: 4s - 16ms/st - loss: 0.0050 - ETA: 4s - 16ms/step   - loss: 0.0016 - ETA: 4s - 16ms/st - loss: 0.0016 - ETA: 4s - 16ms/st - loss: 0.0019 - ETA: 4s - 16ms/st - loss: 3.6607e-04 - ETA: 3s - 16ms/st - loss: 5.1989e-04 - ETA: 3s - 16ms/st - loss: 5.4809e-04 - ETA: 3s - 16ms/st - loss: 2.5490e-04 - ETA: 3s - 16ms/st - loss: 5.4763e-04 - ETA: 3s - 16ms/st - loss: 0.0020 - ETA: 3s - 16ms/step   - loss: 2.5044e-04 - ETA: 3s - 16ms/st - loss: 0.0010 - ETA: 2s - 16ms/step   - loss: 0.0015 - ETA: 2s - 16ms/st - loss: 0.0032 - ETA: 2s - 16ms/st - loss: 6.2850e-04 - ETA: 2s - 16ms/st - loss: 2.3578e-04 - ETA: 2s - 16ms/st - loss: 0.0013 - ETA: 2s - 16ms/step   - loss: 7.5612e-04 - ETA: 1s - 16ms/st - loss: 0.0011 - ETA: 1s - 16ms/step   - loss: 9.2083e-04 - ETA: 1s - 16ms/st - loss: 7.0110e-04 - ETA: 1s - 16ms/st - loss: 0.0014 - ETA: 1s - 16ms/step   - loss: 3.6192e-04 - ETA: 1s - 16ms/st - loss: 0.0014 - ETA: 0s - 16ms/step   - loss: 0.0010 - ETA: 0s - 16ms/st - loss: 4.8116e-04 - ETA: 0s - 16ms/st - loss: 0.0011 - ETA: 0s - 16ms/step   - loss: 0.0020 - ETA: 0s - 16ms/st - loss: 5.0511e-04 - ETA: 0s - 16ms/st - loss: 2.8705e-04 - ETA: 0s - 16ms/st - loss: 3.9981e-04 - 16ms/step          \n",
      "Epoch 99/100\n",
      "step 341/341 [==============================] - loss: 7.9104e-04 - ETA: 5s - 16ms/st - loss: 5.8943e-04 - ETA: 5s - 16ms/st - loss: 0.0010 - ETA: 4s - 16ms/step   - loss: 7.9156e-04 - ETA: 4s - 16ms/st - loss: 4.4286e-04 - ETA: 4s - 16ms/st - loss: 0.0012 - ETA: 4s - 16ms/step   - loss: 0.0014 - ETA: 4s - 16ms/st - loss: 0.0027 - ETA: 4s - 16ms/st - loss: 0.0013 - ETA: 3s - 16ms/st - loss: 0.0016 - ETA: 3s - 16ms/st - loss: 0.0020 - ETA: 3s - 16ms/st - loss: 6.2604e-04 - ETA: 3s - 16ms/st - loss: 3.4860e-04 - ETA: 3s - 16ms/st - loss: 5.9571e-04 - ETA: 3s - 16ms/st - loss: 0.0014 - ETA: 3s - 16ms/step   - loss: 2.6629e-04 - ETA: 2s - 16ms/st - loss: 8.2824e-04 - ETA: 2s - 16ms/st - loss: 9.2111e-04 - ETA: 2s - 16ms/st - loss: 0.0011 - ETA: 2s - 16ms/step   - loss: 2.2374e-04 - ETA: 2s - 16ms/st - loss: 6.9261e-04 - ETA: 2s - 16ms/st - loss: 8.4356e-04 - ETA: 1s - 16ms/st - loss: 0.0015 - ETA: 1s - 16ms/step   - loss: 0.0016 - ETA: 1s - 16ms/st - loss: 6.9174e-04 - ETA: 1s - 16ms/st - loss: 8.5111e-04 - ETA: 1s - 16ms/st - loss: 0.0014 - ETA: 1s - 16ms/step   - loss: 3.9578e-04 - ETA: 0s - 16ms/st - loss: 8.9946e-04 - ETA: 0s - 16ms/st - loss: 0.0023 - ETA: 0s - 16ms/step   - loss: 0.0066 - ETA: 0s - 16ms/st - loss: 0.0085 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/st - loss: 0.0015 - ETA: 0s - 16ms/st - loss: 8.7494e-04 - 16ms/step      \n",
      "Epoch 100/100\n",
      "step 341/341 [==============================] - loss: 4.8439e-04 - ETA: 5s - 16ms/st - loss: 8.4066e-04 - ETA: 5s - 16ms/st - loss: 4.3934e-04 - ETA: 5s - 16ms/st - loss: 6.6343e-04 - ETA: 4s - 16ms/st - loss: 7.6952e-04 - ETA: 4s - 16ms/st - loss: 6.0442e-04 - ETA: 4s - 16ms/st - loss: 7.5094e-04 - ETA: 4s - 16ms/st - loss: 3.4519e-04 - ETA: 4s - 16ms/st - loss: 0.0012 - ETA: 4s - 16ms/step   - loss: 0.0021 - ETA: 3s - 16ms/st - loss: 2.7411e-04 - ETA: 3s - 16ms/st - loss: 6.5491e-04 - ETA: 3s - 16ms/st - loss: 6.4991e-04 - ETA: 3s - 16ms/st - loss: 9.2334e-04 - ETA: 3s - 16ms/st - loss: 4.8762e-04 - ETA: 3s - 16ms/st - loss: 0.0016 - ETA: 2s - 16ms/step   - loss: 0.0019 - ETA: 2s - 16ms/st - loss: 5.2655e-04 - ETA: 2s - 16ms/st - loss: 0.0015 - ETA: 2s - 16ms/step   - loss: 0.0012 - ETA: 2s - 16ms/st - loss: 0.0013 - ETA: 2s - 16ms/st - loss: 0.0015 - ETA: 1s - 16ms/st - loss: 5.6303e-04 - ETA: 1s - 16ms/st - loss: 0.0012 - ETA: 1s - 16ms/step   - loss: 0.0014 - ETA: 1s - 16ms/st - loss: 7.1364e-04 - ETA: 1s - 16ms/st - loss: 0.0014 - ETA: 1s - 16ms/step   - loss: 3.7440e-04 - ETA: 0s - 16ms/st - loss: 0.0012 - ETA: 0s - 16ms/step   - loss: 0.0015 - ETA: 0s - 16ms/st - loss: 9.0845e-04 - ETA: 0s - 16ms/st - loss: 6.6027e-04 - ETA: 0s - 16ms/st - loss: 0.0024 - ETA: 0s - 16ms/step   - loss: 4.3070e-04 - ETA: 0s - 16ms/st - loss: 9.5170e-04 - 16ms/step          \n",
      "save checkpoint at /home/aistudio/save_dir/final\n"
     ]
    }
   ],
   "source": [
    "USE_GPU = True #The device should not be 'gpu', since PaddlePaddle is not compiled with CUDA\n",
    "TRAIN_EPOCH = 100\n",
    "LOG_FREQ = 10\n",
    "SAVE_DIR = os.path.join(os.getcwd(),\"save_dir\")\n",
    "SAVE_FREQ = 10\n",
    "\n",
    "NAME_FLAG=\"1\"\n",
    "\n",
    "if USE_GPU:\n",
    "    paddle.set_device(\"gpu\")\n",
    "else:\n",
    "    paddle.set_device(\"cpu\")\n",
    "\n",
    "model.fit(train_dataset, \n",
    "    batch_size=32,\n",
    "    drop_last=True,\n",
    "    epochs=TRAIN_EPOCH,\n",
    "    log_freq=LOG_FREQ,\n",
    "    save_dir=SAVE_DIR,\n",
    "    save_freq=SAVE_FREQ,\n",
    "    verbose=1\n",
    "    )\n",
    "\n",
    "\n",
    "model.save('./model_save/'+NAME_FLAG)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "## 预测\n",
    "\n",
    "使用训练完毕的模型，对测试集中的日期对应的价格数进行预测。\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Predict begin...\n",
      "step 30/30 [==============================] - ETA: 1s - 38ms/ste - ETA: 0s - 24ms/ste - ETA: 0s - 20ms/ste - ETA: 0s - 17ms/ste - ETA: 0s - 16ms/ste - ETA: 0s - 15ms/ste - ETA: 0s - 15ms/ste - ETA: 0s - 14ms/ste - ETA: 0s - 14ms/ste - ETA: 0s - 13ms/ste - ETA: 0s - 13ms/ste - ETA: 0s - 13ms/ste - ETA: 0s - 13ms/ste - ETA: 0s - 13ms/ste - 13ms/step          \n",
      "Predict samples: 30\n"
     ]
    }
   ],
   "source": [
    "preds = model.predict(\n",
    "        test_data=test_dataset\n",
    "        )"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "## 数据后处理\n",
    "\n",
    "将归一化的数据转换为原始数据，画出真实值对应的曲线和预测值对应的曲线。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "true_cases = scaler.inverse_transform(\n",
    "    np.expand_dims(y_test.flatten(), axis=0)\n",
    ").flatten()\n",
    "\n",
    "predicted_cases = scaler.inverse_transform(\n",
    "  np.expand_dims(np.array(preds).flatten(), axis=0)\n",
    ").flatten()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "10981 10951\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "DatetimeIndex(['1978-12-29', '1979-01-01', '1979-01-02', '1979-01-03',\n",
       "               '1979-01-04', '1979-01-05', '1979-01-08', '1979-01-09',\n",
       "               '1979-01-10', '1979-01-11',\n",
       "               ...\n",
       "               '2020-12-07', '2020-12-08', '2020-12-09', '2020-12-10',\n",
       "               '2020-12-11', '2020-12-14', '2020-12-15', '2020-12-16',\n",
       "               '2020-12-17', '2020-12-18'],\n",
       "              dtype='datetime64[ns]', name='days', length=10951, freq=None)"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print (type(daily_cases))\n",
    "daily_cases[1:3]\n",
    "print (len(daily_cases), len(train_data))\n",
    "daily_cases.index[:len(train_data)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plt.plot(\n",
    "#   daily_cases.index[:len(train_data)], \n",
    "#   scaler.inverse_transform(train_data).flatten(),\n",
    "#   label='Historical Daily Cases'\n",
    "# )\n",
    "plt.figure(figsize=(10,4))\n",
    "\n",
    "plt.plot(\n",
    "  daily_cases.index[len(train_data):len(train_data) + len(true_cases)], \n",
    "  true_cases,\n",
    "  label='Real Daily Cases'\n",
    ")\n",
    "\n",
    "plt.plot(\n",
    "  daily_cases.index[len(train_data):len(train_data) + len(true_cases)], \n",
    "  predicted_cases, \n",
    "  label='Predicted Daily Cases'\n",
    ")\n",
    "\n",
    "plt.legend();"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "## 进一步优化模型\n",
    "\n",
    "从上图中，我们可以看到模型大体上预测到了涨幅的升降情况，在具体数值上则出现了一些误差。读者可以发挥创造力，进一步提升模型的精度与功能，例如：\n",
    "\n",
    "1. 预测未来n天\n",
    "\n",
    "我们现在是用已知的9天价格数，预测第10天的价格数，我们可以将第10天的预测结果与前8天的真实价格数拼接，预测第11天的价格数，以此类推即可预测未来n天的价格数。\n",
    "\n",
    "2. 优化模型网络\n",
    "\n",
    "本文采用的是TCN模型，如果不考虑模型的速度性能，可以尝试LSTM, GRU, transformer等模型，进一步提升模型的拟合能力。\n",
    "\n",
    "3. 优化模型超参数\n",
    "\n",
    "本文没有对超参设置进行探索，读者可以探索设置更加合理的学习率，训练轮次，TCN通道数等。\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "# PaddleNLP 更多项目\n",
    " - [seq2vec是什么? 瞧瞧怎么用它做情感分析](https://aistudio.baidu.com/aistudio/projectdetail/1283423)\n",
    " - [如何通过预训练模型Fine-tune下游任务](https://aistudio.baidu.com/aistudio/projectdetail/1294333)\n",
    " - [使用BiGRU-CRF模型完成快递单信息抽取](https://aistudio.baidu.com/aistudio/projectdetail/1317771)\n",
    " - [使用预训练模型ERNIE优化快递单信息抽取](https://aistudio.baidu.com/aistudio/projectdetail/1329361)\n",
    " - [使用Seq2Seq模型完成自动对联](https://aistudio.baidu.com/aistudio/projectdetail/1321118)\n",
    " - [使用预训练模型ERNIE-GEN实现智能写诗](https://aistudio.baidu.com/aistudio/projectdetail/1339888)\n",
    " - [使用预训练模型完成阅读理解](https://aistudio.baidu.com/aistudio/projectdetail/1339612)\n",
    " - [自定义数据集实现文本多分类任务](https://aistudio.baidu.com/aistudio/projectdetail/1468469)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": false
   },
   "source": [
    "# 加入交流群，一起学习吧\n",
    "\n",
    "现在就加入PaddleNLP的QQ技术交流群，一起交流NLP技术吧！\n",
    "\n",
    "<img src=\"https://ai-studio-static-online.cdn.bcebos.com/d953727af0c24a7c806ab529495f0904f22f809961be420b8c88cdf59b837394\" width=\"200\" height=\"250\" >"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "PaddlePaddle 2.0.0b0 (Python 3.5)",
   "language": "python",
   "name": "py35-paddle1.2.0"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}