age-gender-transfer-learning.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'1.5.0'"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import tensorflow as tf\n",
    "import tensorflow.contrib.slim as slim\n",
    "import numpy as np\n",
    "import os\n",
    "from random import shuffle\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from scipy.misc import imread,imresize\n",
    "import matplotlib.pyplot as plt\n",
    "import conv_blocks as op\n",
    "import multiprocessing\n",
    "import mobilenet_v2\n",
    "import mobilenet\n",
    "tf.__version__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "batch_size = 16\n",
    "epoch = 1000\n",
    "learning_rate = 1e-3\n",
    "DIR = 'crop/'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def _int64_feature(value):\n",
    "    return tf.train.Feature(int64_list=tf.train.Int64List(value=[value]))\n",
    "\n",
    "def _bytes_feature(value):\n",
    "    return tf.train.Feature(bytes_list=tf.train.BytesList(value=[value]))\n",
    "\n",
    "def load_image(addr):\n",
    "    \n",
    "    return img.astype(np.float32)\n",
    "\n",
    "def convert_to(dataset, labels_age, labels_gender, name, i):\n",
    "    if not os.path.exists(name):\n",
    "        os.mkdir(name)\n",
    "    filename = os.path.join(name,'file-%d.tfrecords'%(i))\n",
    "    print('writing %s, cpu %d'%(filename,i))\n",
    "    with tf.python_io.TFRecordWriter(filename) as writer:\n",
    "        for index in range(dataset.shape[0]):\n",
    "            img = imresize(imread(dataset[index]),(224,224)).reshape((224,224,1)).astype(np.float32)\n",
    "            example = tf.train.Example(features=tf.train.Features(feature={\n",
    "                'gender': _int64_feature(int(labels_gender[index])),\n",
    "                'age': _int64_feature(int(labels_age[index])),\n",
    "                'image':_bytes_feature(tf.compat.as_bytes(img.tostring()))}))\n",
    "            writer.write(example.SerializeToString())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "img_lists = os.listdir(DIR)\n",
    "shuffle(img_lists)\n",
    "img_labels, age_labels, images = [], [], []\n",
    "for i in img_lists:\n",
    "    splitted = i.split('--')\n",
    "    if int(splitted[1]) > 100:\n",
    "        continue\n",
    "    img_labels.append(splitted[0])\n",
    "    age_labels.append(int(splitted[1]))\n",
    "    images.append(DIR+i)\n",
    "img_Y = LabelEncoder().fit_transform(img_labels)\n",
    "img_lists = images\n",
    "img_lists = np.array(img_lists)\n",
    "age_labels = np.array(age_labels)\n",
    "img_Y = np.array(img_Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "cpu_cores = 4\n",
    "train_idx = np.linspace(0, img_lists.shape[0], cpu_cores + 1, dtype=np.int)\n",
    "pool = multiprocessing.Pool(processes=cpu_cores)\n",
    "\n",
    "for p in range(cpu_cores):\n",
    "    pool.apply_async(convert_to,\n",
    "                (img_lists[train_idx[p]:train_idx[p + 1] - 1], \n",
    "                 age_labels[train_idx[p]:train_idx[p + 1] - 1],\n",
    "                 img_Y[train_idx[p]:train_idx[p + 1] - 1],\n",
    "                 'train', p,))\n",
    "    \n",
    "pool.close()\n",
    "pool.join()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def read_and_decode(filename_queue):\n",
    "    reader = tf.TFRecordReader()\n",
    "    _, serialized_example = reader.read(filename_queue)\n",
    "    features = tf.parse_single_example(\n",
    "        serialized_example,\n",
    "        features={'gender': tf.FixedLenFeature([], tf.int64),\n",
    "                'age': tf.FixedLenFeature([], tf.int64),\n",
    "                 'image': tf.FixedLenFeature([], tf.string)})\n",
    "    image = tf.decode_raw(features['image'], tf.float32)\n",
    "    image = tf.reshape(image, [224,224,1])\n",
    "    gender = tf.cast(features['gender'], tf.int32)\n",
    "    age = tf.cast(features['age'], tf.int32)\n",
    "    return image, gender, age"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[<tf.Variable 'age/expanded_conv/expand/weights:0' shape=(1, 1, 160, 960) dtype=float32_ref>, <tf.Variable 'age/expanded_conv/expand/biases:0' shape=(960,) dtype=float32_ref>, <tf.Variable 'age/expanded_conv/depthwise/depthwise_weights:0' shape=(3, 3, 960, 1) dtype=float32_ref>, <tf.Variable 'age/expanded_conv/depthwise/biases:0' shape=(960,) dtype=float32_ref>, <tf.Variable 'age/expanded_conv/project/weights:0' shape=(1, 1, 960, 160) dtype=float32_ref>, <tf.Variable 'age/expanded_conv/project/biases:0' shape=(160,) dtype=float32_ref>, <tf.Variable 'age/expanded_conv_1/expand/weights:0' shape=(1, 1, 160, 960) dtype=float32_ref>, <tf.Variable 'age/expanded_conv_1/expand/biases:0' shape=(960,) dtype=float32_ref>, <tf.Variable 'age/expanded_conv_1/depthwise/depthwise_weights:0' shape=(3, 3, 960, 1) dtype=float32_ref>, <tf.Variable 'age/expanded_conv_1/depthwise/biases:0' shape=(960,) dtype=float32_ref>, <tf.Variable 'age/expanded_conv_1/project/weights:0' shape=(1, 1, 960, 320) dtype=float32_ref>, <tf.Variable 'age/expanded_conv_1/project/biases:0' shape=(320,) dtype=float32_ref>, <tf.Variable 'age/expanded_conv_2/expand/weights:0' shape=(1, 1, 320, 1920) dtype=float32_ref>, <tf.Variable 'age/expanded_conv_2/expand/biases:0' shape=(1920,) dtype=float32_ref>, <tf.Variable 'age/expanded_conv_2/depthwise/depthwise_weights:0' shape=(1, 1, 1920, 1) dtype=float32_ref>, <tf.Variable 'age/expanded_conv_2/depthwise/biases:0' shape=(1920,) dtype=float32_ref>, <tf.Variable 'age/expanded_conv_2/project/weights:0' shape=(1, 1, 1920, 1280) dtype=float32_ref>, <tf.Variable 'age/expanded_conv_2/project/biases:0' shape=(1280,) dtype=float32_ref>, <tf.Variable 'age/Conv2d_1c_1x1/weights:0' shape=(1, 1, 1280, 101) dtype=float32_ref>, <tf.Variable 'age/Conv2d_1c_1x1/biases:0' shape=(101,) dtype=float32_ref>, <tf.Variable 'gender/expanded_conv/expand/weights:0' shape=(1, 1, 160, 960) dtype=float32_ref>, <tf.Variable 'gender/expanded_conv/expand/biases:0' shape=(960,) dtype=float32_ref>, <tf.Variable 'gender/expanded_conv/depthwise/depthwise_weights:0' shape=(3, 3, 960, 1) dtype=float32_ref>, <tf.Variable 'gender/expanded_conv/depthwise/biases:0' shape=(960,) dtype=float32_ref>, <tf.Variable 'gender/expanded_conv/project/weights:0' shape=(1, 1, 960, 160) dtype=float32_ref>, <tf.Variable 'gender/expanded_conv/project/biases:0' shape=(160,) dtype=float32_ref>, <tf.Variable 'gender/expanded_conv_1/expand/weights:0' shape=(1, 1, 160, 960) dtype=float32_ref>, <tf.Variable 'gender/expanded_conv_1/expand/biases:0' shape=(960,) dtype=float32_ref>, <tf.Variable 'gender/expanded_conv_1/depthwise/depthwise_weights:0' shape=(3, 3, 960, 1) dtype=float32_ref>, <tf.Variable 'gender/expanded_conv_1/depthwise/biases:0' shape=(960,) dtype=float32_ref>, <tf.Variable 'gender/expanded_conv_1/project/weights:0' shape=(1, 1, 960, 320) dtype=float32_ref>, <tf.Variable 'gender/expanded_conv_1/project/biases:0' shape=(320,) dtype=float32_ref>, <tf.Variable 'gender/expanded_conv_2/expand/weights:0' shape=(1, 1, 320, 1920) dtype=float32_ref>, <tf.Variable 'gender/expanded_conv_2/expand/biases:0' shape=(1920,) dtype=float32_ref>, <tf.Variable 'gender/expanded_conv_2/depthwise/depthwise_weights:0' shape=(1, 1, 1920, 1) dtype=float32_ref>, <tf.Variable 'gender/expanded_conv_2/depthwise/biases:0' shape=(1920,) dtype=float32_ref>, <tf.Variable 'gender/expanded_conv_2/project/weights:0' shape=(1, 1, 1920, 1280) dtype=float32_ref>, <tf.Variable 'gender/expanded_conv_2/project/biases:0' shape=(1280,) dtype=float32_ref>, <tf.Variable 'gender/Conv2d_1c_1x1/weights:0' shape=(1, 1, 1280, 2) dtype=float32_ref>, <tf.Variable 'gender/Conv2d_1c_1x1/biases:0' shape=(2,) dtype=float32_ref>]\n",
      "INFO:tensorflow:Restoring parameters from new/emotion-checkpoint-mobilenet.ckpt\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'new/combine-checkpoint-mobilenet.ckpt'"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tf.reset_default_graph()\n",
    "sess = tf.InteractiveSession()\n",
    "\n",
    "filename_queue = tf.train.string_input_producer(['train/'+i for i in os.listdir('train')], num_epochs=epoch)\n",
    "img, gender, age = read_and_decode(filename_queue)\n",
    "X, Y_gender, Y_age = tf.train.shuffle_batch([img, gender, age], batch_size=batch_size, num_threads=6,\n",
    "                                        capacity=10000,\n",
    "                                        min_after_dequeue=1000, allow_smaller_final_batch=False)\n",
    "\n",
    "images = tf.image.grayscale_to_rgb(X)\n",
    "images = images / 128. - 1\n",
    "with tf.contrib.slim.arg_scope(mobilenet_v2.training_scope(is_training=True)):\n",
    "    logits, endpoints = mobilenet_v2.mobilenet(images)\n",
    "logits = tf.nn.relu6(logits)\n",
    "emotion_logits = slim.fully_connected(logits, 7, activation_fn=None,\n",
    "                                      weights_initializer=tf.truncated_normal_initializer(stddev=0.01),\n",
    "                                      weights_regularizer=slim.l2_regularizer(1e-5),\n",
    "                                      scope='emo/emotion_1', reuse=False)\n",
    "\n",
    "with tf.variable_scope(\"age\"):\n",
    "    age1=op.expanded_conv(endpoints['layer_16'],160,stride=1)\n",
    "    age2=op.expanded_conv(age1,320,stride=1)\n",
    "    age3=mobilenet.global_pool(op.expanded_conv(age2,1280,stride=1,kernel_size=[1, 1]))\n",
    "    age_logits=slim.conv2d(age3,101, [1, 1],activation_fn=None,normalizer_fn=None,\n",
    "                biases_initializer=tf.zeros_initializer(),scope='Conv2d_1c_1x1')\n",
    "    age_logits = tf.squeeze(age_logits, [1, 2])\n",
    "    \n",
    "with tf.variable_scope(\"gender\"):\n",
    "    gender1=op.expanded_conv(endpoints['layer_16'],160,stride=1)\n",
    "    gender2=op.expanded_conv(gender1,320,stride=1)\n",
    "    gender3=mobilenet.global_pool(op.expanded_conv(gender2,1280,stride=1,kernel_size=[1, 1]))\n",
    "    gender_logits=slim.conv2d(gender3,2, [1, 1],activation_fn=None,normalizer_fn=None,\n",
    "                biases_initializer=tf.zeros_initializer(),scope='Conv2d_1c_1x1')\n",
    "    gender_logits = tf.squeeze(gender_logits, [1, 2])\n",
    "\n",
    "gender_accuracy = tf.reduce_mean(tf.cast(tf.nn.in_top_k(gender_logits, Y_gender, 1), tf.float32))\n",
    "age_ = tf.cast(tf.constant([i for i in range(0, 101)]), tf.float32)\n",
    "age = tf.reduce_sum(tf.multiply(tf.nn.softmax(age_logits), age_), axis=1)\n",
    "abs_loss = tf.reduce_mean(tf.abs(tf.cast(Y_age,tf.float32)-age))\n",
    "\n",
    "global_step = tf.Variable(0, name=\"global_step\", trainable=False)\n",
    "lr = tf.train.exponential_decay(learning_rate, global_step=global_step, \n",
    "                                decay_steps=5000, decay_rate=0.9,staircase=True)\n",
    "\n",
    "age_cross_entropy = tf.nn.sparse_softmax_cross_entropy_with_logits(labels=Y_age, logits=age_logits)\n",
    "age_cross_entropy_mean = tf.reduce_mean(age_cross_entropy)\n",
    "gender_cross_entropy = tf.nn.sparse_softmax_cross_entropy_with_logits(labels=Y_gender, logits=gender_logits)\n",
    "gender_cross_entropy_mean = tf.reduce_mean(gender_cross_entropy)\n",
    "cost = tf.add_n([gender_cross_entropy_mean, abs_loss])\n",
    "list_vars = tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES,'age') + tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES,'gender')\n",
    "\n",
    "print(list_vars)\n",
    "tf.summary.scalar(\"gender_cross_entropy\", gender_cross_entropy_mean)\n",
    "tf.summary.scalar(\"total_loss\", cost)\n",
    "tf.summary.scalar(\"train_abs_age_error\", abs_loss)\n",
    "tf.summary.scalar(\"gender_accuracy\", gender_accuracy)\n",
    "tf.summary.scalar(\"learning_rate\", lr)\n",
    "optimizer = tf.train.AdamOptimizer(lr).minimize(cost, var_list=list_vars,global_step=global_step)\n",
    "\n",
    "init_op = tf.group(tf.global_variables_initializer(),tf.local_variables_initializer())\n",
    "sess.run(init_op)\n",
    "\n",
    "var_lists = tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES, scope = 'MobilenetV2') + tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES, scope = 'emo')\n",
    "saver = tf.train.Saver(var_list = var_lists)\n",
    "saver.restore(sess, \"checkpoints/emotion-checkpoint-mobilenet.ckpt\")\n",
    "saver = tf.train.Saver(tf.global_variables())\n",
    "saver.save(sess, \"checkpoints/combine-checkpoint-mobilenet.ckpt\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "merged = tf.summary.merge_all()\n",
    "writer = tf.summary.FileWriter('./logs', sess.graph)\n",
    "coord = tf.train.Coordinator()\n",
    "threads = tf.train.start_queue_runners(sess=sess, coord=coord)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "step 0, loss 22.192686, gender accuracy 0.437500, age diff 21.499508\n",
      "step 200, loss 14.507321, gender accuracy 0.875000, age diff 14.128583\n",
      "step 400, loss 8.341151, gender accuracy 0.812500, age diff 7.891661\n",
      "step 600, loss 14.059522, gender accuracy 0.875000, age diff 13.778696\n",
      "step 800, loss 9.018284, gender accuracy 0.875000, age diff 8.561370\n",
      "step 1000, loss 14.856874, gender accuracy 0.812500, age diff 14.456733\n",
      "step 1200, loss 12.615247, gender accuracy 0.750000, age diff 12.164854\n",
      "step 1400, loss 8.732638, gender accuracy 0.937500, age diff 8.429897\n",
      "step 1600, loss 12.108310, gender accuracy 0.875000, age diff 11.759733\n",
      "step 1800, loss 14.477577, gender accuracy 0.937500, age diff 14.239824\n",
      "step 2000, loss 11.499493, gender accuracy 0.937500, age diff 11.329531\n",
      "step 2200, loss 8.749978, gender accuracy 0.812500, age diff 8.338078\n",
      "step 2400, loss 12.379673, gender accuracy 0.937500, age diff 11.986933\n",
      "step 2600, loss 12.786985, gender accuracy 0.750000, age diff 12.356604\n",
      "step 2800, loss 11.265399, gender accuracy 0.937500, age diff 11.103647\n",
      "step 3000, loss 8.239431, gender accuracy 0.812500, age diff 7.693986\n",
      "step 3200, loss 9.106414, gender accuracy 0.937500, age diff 8.931524\n",
      "step 3400, loss 10.425470, gender accuracy 0.875000, age diff 10.113008\n",
      "step 3600, loss 12.131413, gender accuracy 0.812500, age diff 11.643917\n",
      "step 3800, loss 7.637062, gender accuracy 0.687500, age diff 6.859552\n",
      "step 4000, loss 7.238130, gender accuracy 1.000000, age diff 7.126163\n",
      "step 4200, loss 11.845860, gender accuracy 0.937500, age diff 11.637257\n",
      "step 4400, loss 15.164520, gender accuracy 0.687500, age diff 14.575222\n",
      "step 4600, loss 12.656749, gender accuracy 1.000000, age diff 12.461889\n",
      "step 4800, loss 9.634697, gender accuracy 0.937500, age diff 9.444800\n",
      "step 5000, loss 10.634505, gender accuracy 1.000000, age diff 10.464025\n",
      "step 5200, loss 8.695565, gender accuracy 0.812500, age diff 8.325580\n",
      "step 5400, loss 10.113972, gender accuracy 1.000000, age diff 10.009913\n",
      "step 5600, loss 12.358335, gender accuracy 0.750000, age diff 11.775385\n",
      "step 5800, loss 10.692649, gender accuracy 0.812500, age diff 10.331584\n",
      "step 6000, loss 12.772675, gender accuracy 0.937500, age diff 12.592178\n",
      "step 6200, loss 8.931356, gender accuracy 0.875000, age diff 8.675436\n",
      "step 6400, loss 8.560945, gender accuracy 0.812500, age diff 8.214705\n",
      "step 6600, loss 6.677983, gender accuracy 0.812500, age diff 6.381080\n",
      "step 6800, loss 10.329557, gender accuracy 0.875000, age diff 10.071276\n",
      "step 7000, loss 10.099072, gender accuracy 0.937500, age diff 9.961805\n",
      "step 7200, loss 8.876174, gender accuracy 0.937500, age diff 8.658430\n",
      "step 7400, loss 8.143209, gender accuracy 0.812500, age diff 7.610870\n",
      "step 7600, loss 10.691816, gender accuracy 0.875000, age diff 10.426470\n",
      "step 7800, loss 12.652567, gender accuracy 0.937500, age diff 12.420307\n",
      "step 8000, loss 7.642555, gender accuracy 1.000000, age diff 7.541938\n",
      "step 8200, loss 11.606746, gender accuracy 1.000000, age diff 11.455927\n",
      "step 8400, loss 11.102815, gender accuracy 1.000000, age diff 11.020293\n",
      "step 8600, loss 5.633094, gender accuracy 0.687500, age diff 5.032263\n",
      "step 8800, loss 11.843752, gender accuracy 1.000000, age diff 11.758423\n",
      "step 9000, loss 11.378027, gender accuracy 0.875000, age diff 11.195855\n",
      "step 9200, loss 11.116655, gender accuracy 0.937500, age diff 10.931168\n",
      "step 9400, loss 11.623341, gender accuracy 1.000000, age diff 11.490917\n",
      "step 9600, loss 8.706734, gender accuracy 0.937500, age diff 8.486672\n",
      "step 9800, loss 17.629368, gender accuracy 0.875000, age diff 17.170122\n",
      "step 10000, loss 9.915776, gender accuracy 0.937500, age diff 9.770147\n",
      "step 10200, loss 9.340197, gender accuracy 0.812500, age diff 8.989094\n",
      "step 10400, loss 10.706006, gender accuracy 0.875000, age diff 10.535940\n",
      "step 10600, loss 8.998595, gender accuracy 0.937500, age diff 8.756364\n",
      "step 10800, loss 12.840219, gender accuracy 1.000000, age diff 12.806702\n",
      "step 11000, loss 11.918503, gender accuracy 0.812500, age diff 11.358045\n",
      "step 11200, loss 7.885969, gender accuracy 0.750000, age diff 7.426413\n",
      "step 11400, loss 9.919608, gender accuracy 0.937500, age diff 9.670818\n",
      "step 11600, loss 6.766475, gender accuracy 0.937500, age diff 6.580400\n",
      "step 11800, loss 10.144558, gender accuracy 0.937500, age diff 9.937340\n",
      "step 12000, loss 9.027014, gender accuracy 0.875000, age diff 8.801760\n",
      "step 12200, loss 10.825631, gender accuracy 0.875000, age diff 10.518124\n",
      "step 12400, loss 7.728178, gender accuracy 0.937500, age diff 7.584677\n",
      "step 12600, loss 11.745664, gender accuracy 0.875000, age diff 11.462975\n",
      "step 12800, loss 7.017634, gender accuracy 0.875000, age diff 6.559800\n",
      "step 13000, loss 8.952596, gender accuracy 0.937500, age diff 8.849583\n",
      "step 13200, loss 11.448152, gender accuracy 0.937500, age diff 11.307461\n",
      "step 13400, loss 12.494826, gender accuracy 1.000000, age diff 12.433371\n",
      "step 13600, loss 10.666669, gender accuracy 0.937500, age diff 10.388496\n",
      "step 13800, loss 6.958322, gender accuracy 0.875000, age diff 6.672266\n",
      "step 14000, loss 9.965609, gender accuracy 1.000000, age diff 9.886786\n",
      "step 14200, loss 10.328918, gender accuracy 0.875000, age diff 9.796244\n",
      "step 14400, loss 7.279850, gender accuracy 0.937500, age diff 7.035466\n",
      "step 14600, loss 10.472555, gender accuracy 0.937500, age diff 10.296368\n",
      "step 14800, loss 6.638916, gender accuracy 0.937500, age diff 6.530468\n",
      "step 15000, loss 10.692768, gender accuracy 1.000000, age diff 10.573516\n",
      "step 15200, loss 8.888339, gender accuracy 1.000000, age diff 8.832748\n",
      "step 15400, loss 11.090458, gender accuracy 1.000000, age diff 10.952550\n",
      "step 15600, loss 6.305442, gender accuracy 0.937500, age diff 6.191317\n",
      "step 15800, loss 8.165283, gender accuracy 0.937500, age diff 7.903529\n",
      "step 16000, loss 7.254377, gender accuracy 0.937500, age diff 7.010426\n",
      "step 16200, loss 10.607522, gender accuracy 0.875000, age diff 10.442170\n",
      "step 16400, loss 9.296673, gender accuracy 0.875000, age diff 9.124743\n",
      "step 16600, loss 8.237012, gender accuracy 0.875000, age diff 7.811885\n",
      "step 16800, loss 7.901515, gender accuracy 0.937500, age diff 7.685165\n",
      "step 17000, loss 9.647214, gender accuracy 0.937500, age diff 9.531346\n",
      "step 17200, loss 11.482663, gender accuracy 1.000000, age diff 11.411819\n",
      "step 17400, loss 12.561852, gender accuracy 0.875000, age diff 12.284046\n",
      "step 17600, loss 10.138721, gender accuracy 0.875000, age diff 9.821953\n",
      "step 17800, loss 5.949656, gender accuracy 0.937500, age diff 5.785635\n",
      "step 18000, loss 7.531598, gender accuracy 0.937500, age diff 7.391979\n",
      "step 18200, loss 6.201928, gender accuracy 0.937500, age diff 6.014148\n",
      "step 18400, loss 9.598062, gender accuracy 0.937500, age diff 9.187671\n",
      "step 18600, loss 9.409477, gender accuracy 0.937500, age diff 9.312788\n",
      "step 18800, loss 11.633529, gender accuracy 0.875000, age diff 11.412721\n",
      "step 19000, loss 10.420402, gender accuracy 1.000000, age diff 10.387541\n",
      "step 19200, loss 7.998573, gender accuracy 0.937500, age diff 7.774528\n",
      "step 19400, loss 9.589331, gender accuracy 1.000000, age diff 9.545249\n",
      "step 19600, loss 9.086046, gender accuracy 0.937500, age diff 8.759861\n",
      "step 19800, loss 7.422109, gender accuracy 0.812500, age diff 7.053295\n",
      "step 20000, loss 9.472198, gender accuracy 0.937500, age diff 9.246598\n",
      "step 20200, loss 5.743455, gender accuracy 0.875000, age diff 5.569788\n",
      "step 20400, loss 9.628748, gender accuracy 0.937500, age diff 9.328176\n",
      "step 20600, loss 9.979369, gender accuracy 0.812500, age diff 9.503251\n",
      "step 20800, loss 10.184956, gender accuracy 0.812500, age diff 9.900046\n",
      "step 21000, loss 8.085057, gender accuracy 0.812500, age diff 7.581524\n",
      "step 21200, loss 8.457066, gender accuracy 0.875000, age diff 8.257521\n",
      "step 21400, loss 7.142791, gender accuracy 0.937500, age diff 6.965573\n",
      "step 21600, loss 8.571746, gender accuracy 0.875000, age diff 8.343554\n",
      "step 21800, loss 9.334108, gender accuracy 1.000000, age diff 9.267293\n",
      "step 22000, loss 7.069529, gender accuracy 0.812500, age diff 6.696679\n",
      "step 22200, loss 12.957796, gender accuracy 1.000000, age diff 12.895372\n",
      "step 22400, loss 6.387525, gender accuracy 1.000000, age diff 6.278584\n",
      "step 22600, loss 11.761610, gender accuracy 0.875000, age diff 11.576927\n",
      "step 22800, loss 6.371956, gender accuracy 1.000000, age diff 6.338483\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "step 23000, loss 10.232597, gender accuracy 1.000000, age diff 10.158941\n",
      "step 23200, loss 9.113963, gender accuracy 0.937500, age diff 8.900523\n",
      "step 23400, loss 9.919561, gender accuracy 0.875000, age diff 9.503361\n",
      "step 23600, loss 9.743428, gender accuracy 0.937500, age diff 9.477063\n",
      "step 23800, loss 10.177210, gender accuracy 0.875000, age diff 9.766184\n",
      "step 24000, loss 7.720944, gender accuracy 0.875000, age diff 7.451595\n",
      "step 24200, loss 11.162285, gender accuracy 0.937500, age diff 10.889971\n",
      "step 24400, loss 7.765008, gender accuracy 0.812500, age diff 7.475551\n",
      "step 24600, loss 10.725598, gender accuracy 1.000000, age diff 10.645592\n",
      "step 24800, loss 8.391574, gender accuracy 0.875000, age diff 8.136404\n",
      "step 25000, loss 6.444004, gender accuracy 0.875000, age diff 6.272488\n",
      "step 25200, loss 6.999681, gender accuracy 0.937500, age diff 6.907752\n",
      "step 25400, loss 10.738687, gender accuracy 1.000000, age diff 10.705489\n",
      "step 25600, loss 9.786022, gender accuracy 0.875000, age diff 9.493807\n",
      "step 25800, loss 12.453335, gender accuracy 1.000000, age diff 12.372948\n",
      "step 26000, loss 9.082864, gender accuracy 1.000000, age diff 8.986232\n",
      "step 26200, loss 6.820661, gender accuracy 0.937500, age diff 6.600734\n",
      "step 26400, loss 8.689387, gender accuracy 0.937500, age diff 8.571494\n",
      "step 26600, loss 11.371673, gender accuracy 0.875000, age diff 11.156063\n",
      "step 26800, loss 10.120234, gender accuracy 0.812500, age diff 9.508089\n",
      "step 27000, loss 15.371859, gender accuracy 0.875000, age diff 15.181623\n",
      "step 27200, loss 7.363397, gender accuracy 0.812500, age diff 7.008426\n",
      "step 27400, loss 9.571310, gender accuracy 0.937500, age diff 9.441635\n",
      "step 27600, loss 6.365754, gender accuracy 0.875000, age diff 6.047544\n",
      "step 27800, loss 9.001048, gender accuracy 0.937500, age diff 8.843369\n",
      "step 28000, loss 9.735421, gender accuracy 0.875000, age diff 9.490185\n",
      "step 28200, loss 10.357264, gender accuracy 0.812500, age diff 10.030910\n",
      "step 28400, loss 7.958627, gender accuracy 0.937500, age diff 7.775043\n",
      "step 28600, loss 8.920217, gender accuracy 0.937500, age diff 8.843696\n",
      "step 28800, loss 8.426517, gender accuracy 0.750000, age diff 7.916965\n",
      "step 29000, loss 10.798929, gender accuracy 0.937500, age diff 10.493858\n",
      "step 29200, loss 8.117023, gender accuracy 0.875000, age diff 7.876748\n",
      "step 29400, loss 11.614983, gender accuracy 0.750000, age diff 11.207440\n",
      "step 29600, loss 7.386829, gender accuracy 1.000000, age diff 7.375783\n",
      "step 29800, loss 7.162596, gender accuracy 0.875000, age diff 6.877424\n",
      "step 30000, loss 10.088637, gender accuracy 0.875000, age diff 9.897511\n",
      "step 30200, loss 12.705653, gender accuracy 0.875000, age diff 12.541536\n",
      "step 30400, loss 8.816688, gender accuracy 0.875000, age diff 8.662755\n",
      "step 30600, loss 6.071247, gender accuracy 0.812500, age diff 5.787233\n",
      "step 30800, loss 8.205566, gender accuracy 1.000000, age diff 8.168352\n",
      "step 31000, loss 9.436595, gender accuracy 1.000000, age diff 9.307051\n",
      "step 31200, loss 9.296901, gender accuracy 0.875000, age diff 9.178407\n",
      "step 31400, loss 7.606805, gender accuracy 0.812500, age diff 7.241457\n",
      "step 31600, loss 12.870743, gender accuracy 1.000000, age diff 12.827225\n",
      "step 31800, loss 8.885884, gender accuracy 0.875000, age diff 8.730122\n",
      "step 32000, loss 8.314382, gender accuracy 0.875000, age diff 8.165947\n",
      "step 32200, loss 8.228582, gender accuracy 1.000000, age diff 8.201294\n",
      "step 32400, loss 9.935289, gender accuracy 0.875000, age diff 9.642957\n",
      "step 32600, loss 9.394534, gender accuracy 1.000000, age diff 9.230305\n",
      "step 32800, loss 9.264453, gender accuracy 0.937500, age diff 9.140673\n",
      "step 33000, loss 12.554236, gender accuracy 0.875000, age diff 12.294332\n",
      "step 33200, loss 7.908188, gender accuracy 0.937500, age diff 7.571620\n",
      "step 33400, loss 9.683484, gender accuracy 1.000000, age diff 9.637035\n",
      "step 33600, loss 11.394361, gender accuracy 0.875000, age diff 11.131975\n",
      "step 33800, loss 8.600754, gender accuracy 0.937500, age diff 8.438721\n",
      "step 34000, loss 7.343167, gender accuracy 1.000000, age diff 7.248756\n",
      "step 34200, loss 11.134276, gender accuracy 1.000000, age diff 11.081413\n",
      "step 34400, loss 8.412242, gender accuracy 1.000000, age diff 8.301497\n",
      "step 34600, loss 8.325890, gender accuracy 0.937500, age diff 8.204714\n",
      "step 34800, loss 11.065581, gender accuracy 1.000000, age diff 11.005175\n",
      "step 35000, loss 6.812912, gender accuracy 0.875000, age diff 6.524718\n",
      "step 35200, loss 8.384560, gender accuracy 0.937500, age diff 8.242764\n",
      "step 35400, loss 6.776310, gender accuracy 1.000000, age diff 6.713306\n",
      "step 35600, loss 8.566159, gender accuracy 1.000000, age diff 8.493992\n",
      "step 35800, loss 8.741134, gender accuracy 1.000000, age diff 8.670790\n",
      "step 36000, loss 8.946682, gender accuracy 1.000000, age diff 8.895151\n",
      "step 36200, loss 10.031071, gender accuracy 0.937500, age diff 9.939004\n",
      "step 36400, loss 8.516103, gender accuracy 1.000000, age diff 8.475624\n",
      "step 36600, loss 10.348659, gender accuracy 1.000000, age diff 10.286919\n",
      "step 36800, loss 7.761091, gender accuracy 0.875000, age diff 7.439442\n",
      "step 37000, loss 7.420151, gender accuracy 0.750000, age diff 6.963294\n",
      "step 37200, loss 9.309318, gender accuracy 1.000000, age diff 9.199577\n",
      "step 37400, loss 9.049835, gender accuracy 0.875000, age diff 8.729015\n",
      "step 37600, loss 8.179249, gender accuracy 0.875000, age diff 7.905039\n",
      "step 37800, loss 8.854062, gender accuracy 0.937500, age diff 8.759964\n",
      "step 38000, loss 10.430850, gender accuracy 0.937500, age diff 10.258900\n",
      "step 38200, loss 10.573936, gender accuracy 0.812500, age diff 10.091194\n",
      "step 38400, loss 16.396103, gender accuracy 0.875000, age diff 16.186478\n",
      "step 38600, loss 10.776916, gender accuracy 0.812500, age diff 10.394094\n",
      "step 38800, loss 7.742566, gender accuracy 1.000000, age diff 7.657590\n",
      "step 39000, loss 9.205831, gender accuracy 0.937500, age diff 8.627777\n",
      "step 39200, loss 7.309316, gender accuracy 0.937500, age diff 7.159792\n",
      "step 39400, loss 8.796338, gender accuracy 0.875000, age diff 8.454374\n",
      "step 39600, loss 8.824011, gender accuracy 0.812500, age diff 8.401544\n",
      "step 39800, loss 8.458420, gender accuracy 1.000000, age diff 8.376956\n",
      "step 40000, loss 7.899652, gender accuracy 1.000000, age diff 7.799088\n",
      "step 40200, loss 10.014953, gender accuracy 0.875000, age diff 9.618781\n",
      "step 40400, loss 6.866280, gender accuracy 1.000000, age diff 6.791124\n",
      "step 40600, loss 9.679733, gender accuracy 0.937500, age diff 9.469592\n",
      "step 40800, loss 10.690363, gender accuracy 0.937500, age diff 10.575425\n",
      "step 41000, loss 9.204327, gender accuracy 0.750000, age diff 8.656849\n",
      "step 41200, loss 6.158283, gender accuracy 0.875000, age diff 5.903772\n",
      "step 41400, loss 7.966691, gender accuracy 0.875000, age diff 7.750874\n",
      "step 41600, loss 7.490274, gender accuracy 0.875000, age diff 7.089842\n",
      "step 41800, loss 9.459007, gender accuracy 1.000000, age diff 9.374935\n",
      "step 42000, loss 10.888577, gender accuracy 0.937500, age diff 10.660568\n",
      "step 42200, loss 6.640122, gender accuracy 1.000000, age diff 6.601013\n",
      "step 42400, loss 10.690172, gender accuracy 0.937500, age diff 10.297030\n",
      "step 42600, loss 9.495924, gender accuracy 0.875000, age diff 9.351486\n",
      "step 42800, loss 11.403465, gender accuracy 0.937500, age diff 11.276414\n",
      "step 43000, loss 9.289396, gender accuracy 0.937500, age diff 9.114232\n",
      "step 43200, loss 13.499248, gender accuracy 0.937500, age diff 13.336872\n",
      "step 43400, loss 8.145902, gender accuracy 1.000000, age diff 8.083175\n",
      "step 43600, loss 8.569244, gender accuracy 0.937500, age diff 8.376690\n",
      "step 43800, loss 8.414320, gender accuracy 0.937500, age diff 8.162876\n",
      "step 44000, loss 13.129267, gender accuracy 1.000000, age diff 13.091146\n",
      "step 44200, loss 8.033551, gender accuracy 0.937500, age diff 7.853171\n",
      "step 44400, loss 8.500471, gender accuracy 0.937500, age diff 8.183201\n",
      "step 44600, loss 7.652180, gender accuracy 1.000000, age diff 7.605954\n",
      "step 44800, loss 5.793468, gender accuracy 1.000000, age diff 5.711270\n",
      "step 45000, loss 8.381581, gender accuracy 0.937500, age diff 8.221815\n",
      "step 45200, loss 8.167121, gender accuracy 0.937500, age diff 8.056684\n",
      "step 45400, loss 9.211373, gender accuracy 0.937500, age diff 8.978178\n",
      "step 45600, loss 7.396837, gender accuracy 0.875000, age diff 6.934801\n",
      "step 45800, loss 6.019891, gender accuracy 0.937500, age diff 5.909371\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "step 46000, loss 9.691566, gender accuracy 0.937500, age diff 9.623816\n",
      "step 46200, loss 7.827639, gender accuracy 0.812500, age diff 7.437222\n",
      "step 46400, loss 9.499778, gender accuracy 0.937500, age diff 9.353777\n",
      "step 46600, loss 9.717016, gender accuracy 1.000000, age diff 9.633421\n",
      "step 46800, loss 9.742319, gender accuracy 0.875000, age diff 9.521626\n",
      "step 47000, loss 10.434070, gender accuracy 1.000000, age diff 10.379274\n",
      "step 47200, loss 8.578937, gender accuracy 0.937500, age diff 8.474881\n",
      "step 47400, loss 8.644316, gender accuracy 1.000000, age diff 8.633080\n",
      "step 47600, loss 11.010694, gender accuracy 1.000000, age diff 10.869202\n",
      "step 47800, loss 10.045769, gender accuracy 0.937500, age diff 9.943533\n",
      "step 48000, loss 9.499102, gender accuracy 1.000000, age diff 9.407628\n",
      "step 48200, loss 8.974433, gender accuracy 0.875000, age diff 8.696869\n",
      "step 48400, loss 6.980006, gender accuracy 0.937500, age diff 6.847468\n",
      "step 48600, loss 8.420555, gender accuracy 0.937500, age diff 8.209229\n",
      "step 48800, loss 9.370664, gender accuracy 0.937500, age diff 9.311350\n",
      "step 49000, loss 8.323671, gender accuracy 1.000000, age diff 8.257151\n",
      "step 49200, loss 7.826358, gender accuracy 0.937500, age diff 7.765036\n",
      "step 49400, loss 6.823288, gender accuracy 0.937500, age diff 6.510955\n",
      "step 49600, loss 13.643204, gender accuracy 0.937500, age diff 13.493626\n",
      "step 49800, loss 7.841042, gender accuracy 0.812500, age diff 7.209445\n",
      "step 50000, loss 11.726859, gender accuracy 0.937500, age diff 11.599003\n",
      "step 50200, loss 8.010859, gender accuracy 0.937500, age diff 7.885092\n",
      "step 50400, loss 7.813880, gender accuracy 1.000000, age diff 7.723607\n",
      "step 50600, loss 8.261930, gender accuracy 0.875000, age diff 8.029804\n",
      "step 50800, loss 7.669840, gender accuracy 1.000000, age diff 7.593266\n",
      "step 51000, loss 5.386976, gender accuracy 0.937500, age diff 5.271478\n",
      "step 51200, loss 10.466228, gender accuracy 1.000000, age diff 10.412231\n",
      "step 51400, loss 10.556049, gender accuracy 0.875000, age diff 10.296566\n",
      "step 51600, loss 8.749153, gender accuracy 1.000000, age diff 8.652899\n",
      "step 51800, loss 9.738459, gender accuracy 0.875000, age diff 9.543026\n",
      "step 52000, loss 8.132329, gender accuracy 0.937500, age diff 8.025893\n",
      "step 52200, loss 9.205869, gender accuracy 0.937500, age diff 9.105015\n",
      "step 52400, loss 6.180870, gender accuracy 1.000000, age diff 6.156742\n",
      "step 52600, loss 7.342809, gender accuracy 1.000000, age diff 7.317899\n",
      "step 52800, loss 8.247925, gender accuracy 1.000000, age diff 8.172865\n",
      "step 53000, loss 9.838828, gender accuracy 1.000000, age diff 9.782872\n",
      "step 53200, loss 8.156157, gender accuracy 1.000000, age diff 8.104257\n",
      "step 53400, loss 10.658769, gender accuracy 1.000000, age diff 10.548035\n",
      "step 53600, loss 8.127513, gender accuracy 0.937500, age diff 7.993989\n",
      "step 53800, loss 9.573679, gender accuracy 0.937500, age diff 9.439030\n",
      "step 54000, loss 6.575261, gender accuracy 1.000000, age diff 6.510856\n",
      "step 54200, loss 8.024885, gender accuracy 1.000000, age diff 7.951020\n",
      "step 54400, loss 8.411953, gender accuracy 1.000000, age diff 8.385582\n",
      "step 54600, loss 7.584858, gender accuracy 1.000000, age diff 7.552723\n",
      "step 54800, loss 12.812325, gender accuracy 0.937500, age diff 12.731687\n",
      "step 55000, loss 8.072612, gender accuracy 0.875000, age diff 7.836900\n",
      "step 55200, loss 7.093176, gender accuracy 1.000000, age diff 7.080154\n",
      "step 55400, loss 11.435358, gender accuracy 1.000000, age diff 11.389160\n",
      "step 55600, loss 6.441956, gender accuracy 0.875000, age diff 6.170304\n",
      "step 55800, loss 13.027705, gender accuracy 0.875000, age diff 12.685867\n",
      "step 56000, loss 8.546417, gender accuracy 0.750000, age diff 8.149405\n",
      "step 56200, loss 10.214815, gender accuracy 0.812500, age diff 9.781864\n",
      "step 56400, loss 8.807690, gender accuracy 0.937500, age diff 8.687105\n",
      "step 56600, loss 7.362513, gender accuracy 0.937500, age diff 7.223131\n",
      "step 56800, loss 6.958272, gender accuracy 0.812500, age diff 6.669971\n",
      "step 57000, loss 10.128019, gender accuracy 0.875000, age diff 9.916342\n",
      "step 57200, loss 5.328611, gender accuracy 0.875000, age diff 5.040311\n",
      "step 57400, loss 8.609104, gender accuracy 0.937500, age diff 8.463815\n",
      "step 57600, loss 8.536201, gender accuracy 1.000000, age diff 8.467738\n",
      "step 57800, loss 10.908389, gender accuracy 0.937500, age diff 10.736704\n",
      "step 58000, loss 8.039773, gender accuracy 0.937500, age diff 7.859101\n",
      "step 58200, loss 7.431969, gender accuracy 0.937500, age diff 7.238010\n",
      "step 58400, loss 5.361001, gender accuracy 0.937500, age diff 5.169782\n",
      "step 58600, loss 13.260759, gender accuracy 1.000000, age diff 13.231308\n",
      "step 58800, loss 5.700260, gender accuracy 0.937500, age diff 5.380891\n",
      "step 59000, loss 15.048615, gender accuracy 0.875000, age diff 14.740885\n",
      "step 59200, loss 10.916557, gender accuracy 1.000000, age diff 10.829841\n",
      "step 59400, loss 7.601211, gender accuracy 1.000000, age diff 7.543455\n",
      "step 59600, loss 12.215972, gender accuracy 1.000000, age diff 12.143442\n",
      "step 59800, loss 9.034924, gender accuracy 1.000000, age diff 8.985945\n",
      "step 60000, loss 10.229966, gender accuracy 0.937500, age diff 10.062256\n",
      "step 60200, loss 7.126793, gender accuracy 0.937500, age diff 7.014352\n",
      "step 60400, loss 11.152195, gender accuracy 0.937500, age diff 11.081781\n",
      "step 60600, loss 8.776347, gender accuracy 1.000000, age diff 8.727847\n",
      "step 60800, loss 8.520176, gender accuracy 0.937500, age diff 8.437386\n",
      "step 61000, loss 9.931352, gender accuracy 1.000000, age diff 9.904614\n",
      "step 61200, loss 10.142290, gender accuracy 0.937500, age diff 10.008932\n",
      "step 61400, loss 9.098572, gender accuracy 0.812500, age diff 8.775547\n",
      "step 61600, loss 8.723823, gender accuracy 0.875000, age diff 8.457167\n",
      "step 61800, loss 8.724539, gender accuracy 0.875000, age diff 8.459232\n",
      "step 62000, loss 9.203468, gender accuracy 0.937500, age diff 9.018183\n",
      "step 62200, loss 7.734260, gender accuracy 1.000000, age diff 7.716048\n",
      "step 62400, loss 6.739122, gender accuracy 0.937500, age diff 6.630586\n",
      "step 62600, loss 7.327944, gender accuracy 0.875000, age diff 7.147848\n",
      "step 62800, loss 10.921033, gender accuracy 0.937500, age diff 10.837863\n",
      "step 63000, loss 11.194761, gender accuracy 1.000000, age diff 11.151985\n",
      "step 63200, loss 10.086213, gender accuracy 1.000000, age diff 10.034636\n",
      "step 63400, loss 9.712614, gender accuracy 0.875000, age diff 9.451817\n",
      "step 63600, loss 8.610101, gender accuracy 1.000000, age diff 8.593685\n",
      "step 63800, loss 11.279027, gender accuracy 0.937500, age diff 11.174274\n",
      "step 64000, loss 7.095858, gender accuracy 0.937500, age diff 6.806205\n",
      "step 64200, loss 6.475690, gender accuracy 1.000000, age diff 6.410770\n",
      "step 64400, loss 6.849268, gender accuracy 0.812500, age diff 6.464913\n",
      "step 64600, loss 7.443002, gender accuracy 0.937500, age diff 7.148734\n",
      "step 64800, loss 8.318248, gender accuracy 1.000000, age diff 8.288411\n",
      "step 65000, loss 6.911494, gender accuracy 1.000000, age diff 6.863624\n",
      "step 65200, loss 7.397148, gender accuracy 0.875000, age diff 7.154789\n",
      "step 65400, loss 8.830845, gender accuracy 1.000000, age diff 8.791678\n",
      "step 65600, loss 9.977069, gender accuracy 1.000000, age diff 9.946098\n",
      "step 65800, loss 9.423855, gender accuracy 1.000000, age diff 9.407276\n",
      "step 66000, loss 12.908448, gender accuracy 1.000000, age diff 12.890803\n",
      "step 66200, loss 10.618552, gender accuracy 0.937500, age diff 10.499517\n",
      "step 66400, loss 6.723740, gender accuracy 1.000000, age diff 6.687421\n",
      "step 66600, loss 9.201924, gender accuracy 0.937500, age diff 9.109308\n",
      "step 66800, loss 10.188029, gender accuracy 0.875000, age diff 9.907298\n",
      "step 67000, loss 8.821872, gender accuracy 0.875000, age diff 8.486937\n",
      "step 67200, loss 8.617531, gender accuracy 0.937500, age diff 8.536829\n",
      "step 67400, loss 9.680121, gender accuracy 0.875000, age diff 9.490726\n",
      "step 67600, loss 9.122805, gender accuracy 1.000000, age diff 9.077054\n",
      "step 67800, loss 8.472969, gender accuracy 1.000000, age diff 8.411442\n",
      "step 68000, loss 6.602285, gender accuracy 1.000000, age diff 6.585188\n",
      "step 68200, loss 8.226937, gender accuracy 1.000000, age diff 8.165344\n",
      "step 68400, loss 11.322997, gender accuracy 0.937500, age diff 11.200788\n",
      "step 68600, loss 6.024279, gender accuracy 1.000000, age diff 5.968598\n",
      "step 68800, loss 5.314990, gender accuracy 1.000000, age diff 5.264915\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "step 69000, loss 8.104418, gender accuracy 0.875000, age diff 7.787526\n",
      "step 69200, loss 7.490416, gender accuracy 1.000000, age diff 7.450109\n",
      "step 69400, loss 7.623877, gender accuracy 1.000000, age diff 7.601949\n",
      "step 69600, loss 9.306555, gender accuracy 0.875000, age diff 9.058871\n",
      "step 69800, loss 8.773368, gender accuracy 1.000000, age diff 8.746084\n",
      "step 70000, loss 6.332264, gender accuracy 1.000000, age diff 6.296140\n",
      "step 70200, loss 9.331511, gender accuracy 0.937500, age diff 9.176203\n",
      "step 70400, loss 8.718819, gender accuracy 0.875000, age diff 8.431681\n",
      "step 70600, loss 8.327847, gender accuracy 1.000000, age diff 8.273245\n",
      "step 70800, loss 8.698475, gender accuracy 0.937500, age diff 8.460566\n",
      "step 71000, loss 6.783820, gender accuracy 0.937500, age diff 6.682073\n",
      "step 71200, loss 7.230310, gender accuracy 1.000000, age diff 7.151170\n",
      "step 71400, loss 6.130629, gender accuracy 0.937500, age diff 6.067943\n",
      "step 71600, loss 7.964658, gender accuracy 1.000000, age diff 7.918936\n",
      "step 71800, loss 5.452634, gender accuracy 1.000000, age diff 5.441780\n",
      "step 72000, loss 8.750516, gender accuracy 1.000000, age diff 8.701813\n",
      "step 72200, loss 5.654587, gender accuracy 1.000000, age diff 5.602653\n",
      "step 72400, loss 7.005715, gender accuracy 0.937500, age diff 6.876373\n",
      "step 72600, loss 5.804044, gender accuracy 0.875000, age diff 5.369994\n",
      "step 72800, loss 6.747165, gender accuracy 0.875000, age diff 6.625289\n",
      "step 73000, loss 8.374841, gender accuracy 1.000000, age diff 8.335477\n",
      "step 73200, loss 6.291635, gender accuracy 1.000000, age diff 6.228549\n",
      "step 73400, loss 9.222164, gender accuracy 1.000000, age diff 9.180399\n",
      "step 73600, loss 8.400781, gender accuracy 0.875000, age diff 8.225361\n",
      "step 73800, loss 10.373840, gender accuracy 1.000000, age diff 10.339296\n",
      "step 74000, loss 6.665609, gender accuracy 1.000000, age diff 6.649319\n",
      "step 74200, loss 9.548489, gender accuracy 1.000000, age diff 9.514567\n",
      "step 74400, loss 9.479137, gender accuracy 1.000000, age diff 9.415430\n",
      "step 74600, loss 8.693264, gender accuracy 0.937500, age diff 8.424874\n",
      "step 74800, loss 6.220884, gender accuracy 0.937500, age diff 6.134798\n",
      "step 75000, loss 12.948354, gender accuracy 1.000000, age diff 12.916708\n",
      "step 75200, loss 6.637650, gender accuracy 1.000000, age diff 6.631224\n",
      "step 75400, loss 13.837984, gender accuracy 0.937500, age diff 13.729939\n",
      "step 75600, loss 9.741374, gender accuracy 1.000000, age diff 9.692165\n",
      "step 75800, loss 9.216219, gender accuracy 1.000000, age diff 9.174498\n",
      "step 76000, loss 8.630155, gender accuracy 1.000000, age diff 8.614008\n",
      "step 76200, loss 10.375215, gender accuracy 1.000000, age diff 10.316092\n",
      "step 76400, loss 10.285496, gender accuracy 0.937500, age diff 10.183577\n",
      "step 76600, loss 7.372294, gender accuracy 0.937500, age diff 7.307974\n",
      "step 76800, loss 7.561724, gender accuracy 0.937500, age diff 7.481084\n",
      "step 77000, loss 7.831446, gender accuracy 1.000000, age diff 7.791486\n",
      "step 77200, loss 7.919594, gender accuracy 1.000000, age diff 7.915070\n",
      "step 77400, loss 9.183257, gender accuracy 1.000000, age diff 9.124628\n",
      "step 77600, loss 11.012024, gender accuracy 1.000000, age diff 10.965557\n",
      "step 77800, loss 8.230931, gender accuracy 1.000000, age diff 8.225704\n",
      "step 78000, loss 11.695889, gender accuracy 1.000000, age diff 11.688189\n",
      "step 78200, loss 8.024947, gender accuracy 0.875000, age diff 7.824363\n",
      "step 78400, loss 6.004338, gender accuracy 0.875000, age diff 5.622754\n",
      "step 78600, loss 6.962766, gender accuracy 1.000000, age diff 6.941163\n",
      "step 78800, loss 6.090608, gender accuracy 1.000000, age diff 6.063759\n",
      "step 79000, loss 8.740350, gender accuracy 0.937500, age diff 8.624330\n",
      "step 79200, loss 8.867407, gender accuracy 1.000000, age diff 8.796577\n",
      "step 79400, loss 7.756358, gender accuracy 1.000000, age diff 7.747087\n",
      "step 79600, loss 7.098443, gender accuracy 0.937500, age diff 6.904164\n",
      "step 79800, loss 9.339297, gender accuracy 0.937500, age diff 9.247663\n",
      "step 80000, loss 8.051390, gender accuracy 1.000000, age diff 8.032579\n",
      "step 80200, loss 8.910893, gender accuracy 1.000000, age diff 8.852314\n",
      "step 80400, loss 8.750501, gender accuracy 0.937500, age diff 8.625196\n",
      "step 80600, loss 9.331690, gender accuracy 0.875000, age diff 9.186399\n",
      "step 80800, loss 7.504506, gender accuracy 1.000000, age diff 7.440892\n",
      "step 81000, loss 9.003839, gender accuracy 0.937500, age diff 8.911346\n",
      "step 81200, loss 10.803130, gender accuracy 0.937500, age diff 10.747492\n",
      "step 81400, loss 7.584488, gender accuracy 0.937500, age diff 7.517144\n",
      "step 81600, loss 7.929582, gender accuracy 0.937500, age diff 7.767557\n",
      "step 81800, loss 12.989425, gender accuracy 0.937500, age diff 12.909436\n",
      "step 82000, loss 7.937957, gender accuracy 0.812500, age diff 7.643112\n",
      "step 82200, loss 10.447308, gender accuracy 0.937500, age diff 10.314843\n",
      "step 82400, loss 6.762543, gender accuracy 1.000000, age diff 6.757178\n",
      "step 82600, loss 8.112280, gender accuracy 1.000000, age diff 8.105200\n",
      "step 82800, loss 9.847117, gender accuracy 0.937500, age diff 9.688221\n",
      "step 83000, loss 7.880388, gender accuracy 1.000000, age diff 7.857933\n",
      "step 83200, loss 9.142926, gender accuracy 0.937500, age diff 9.092386\n",
      "step 83400, loss 6.718220, gender accuracy 1.000000, age diff 6.693040\n",
      "step 83600, loss 9.194777, gender accuracy 1.000000, age diff 9.186998\n",
      "step 83800, loss 7.843067, gender accuracy 0.812500, age diff 7.089872\n",
      "step 84000, loss 5.455246, gender accuracy 1.000000, age diff 5.443883\n",
      "step 84200, loss 7.288983, gender accuracy 1.000000, age diff 7.265398\n",
      "step 84400, loss 8.674755, gender accuracy 0.875000, age diff 8.344488\n",
      "step 84600, loss 11.699527, gender accuracy 0.875000, age diff 11.335391\n",
      "step 84800, loss 4.522877, gender accuracy 1.000000, age diff 4.511458\n",
      "step 85000, loss 6.113622, gender accuracy 1.000000, age diff 6.074466\n",
      "step 85200, loss 10.715892, gender accuracy 0.937500, age diff 10.578498\n",
      "step 85400, loss 7.813317, gender accuracy 0.937500, age diff 7.702182\n",
      "step 85600, loss 11.463308, gender accuracy 0.875000, age diff 11.309225\n",
      "step 85800, loss 6.758701, gender accuracy 1.000000, age diff 6.732211\n",
      "step 86000, loss 11.937233, gender accuracy 0.937500, age diff 11.850545\n",
      "step 86200, loss 8.407109, gender accuracy 0.875000, age diff 8.132051\n",
      "step 86400, loss 10.205475, gender accuracy 1.000000, age diff 10.085249\n",
      "step 86600, loss 9.403728, gender accuracy 1.000000, age diff 9.363892\n",
      "step 86800, loss 5.404642, gender accuracy 1.000000, age diff 5.369199\n",
      "step 87000, loss 8.632974, gender accuracy 1.000000, age diff 8.601765\n",
      "step 87200, loss 9.744526, gender accuracy 0.875000, age diff 9.468904\n",
      "step 87400, loss 9.608854, gender accuracy 1.000000, age diff 9.581521\n",
      "step 87600, loss 9.387394, gender accuracy 0.937500, age diff 9.301455\n",
      "step 87800, loss 9.141389, gender accuracy 1.000000, age diff 9.087504\n",
      "step 88000, loss 7.706475, gender accuracy 0.937500, age diff 7.564975\n",
      "step 88200, loss 8.436590, gender accuracy 1.000000, age diff 8.409414\n",
      "step 88400, loss 9.203459, gender accuracy 1.000000, age diff 9.192836\n",
      "step 88600, loss 9.022256, gender accuracy 1.000000, age diff 9.007110\n",
      "step 88800, loss 9.437532, gender accuracy 0.937500, age diff 9.277614\n",
      "step 89000, loss 9.479275, gender accuracy 0.937500, age diff 9.285419\n",
      "step 89200, loss 8.637538, gender accuracy 0.937500, age diff 8.466461\n",
      "step 89400, loss 10.975183, gender accuracy 0.937500, age diff 10.913059\n",
      "step 89600, loss 7.677116, gender accuracy 0.937500, age diff 7.572819\n",
      "step 89800, loss 7.803854, gender accuracy 1.000000, age diff 7.754622\n",
      "step 90000, loss 5.804849, gender accuracy 1.000000, age diff 5.741963\n",
      "step 90200, loss 11.524823, gender accuracy 0.937500, age diff 11.440445\n",
      "step 90400, loss 7.424569, gender accuracy 1.000000, age diff 7.394546\n",
      "step 90600, loss 11.497972, gender accuracy 1.000000, age diff 11.428399\n",
      "step 90800, loss 11.057159, gender accuracy 1.000000, age diff 11.040185\n",
      "step 91000, loss 10.541770, gender accuracy 1.000000, age diff 10.531211\n",
      "step 91200, loss 7.324989, gender accuracy 0.937500, age diff 7.194443\n",
      "step 91400, loss 8.549691, gender accuracy 1.000000, age diff 8.538179\n",
      "step 91600, loss 5.293838, gender accuracy 1.000000, age diff 5.230196\n",
      "step 91800, loss 6.900919, gender accuracy 1.000000, age diff 6.863384\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "step 92000, loss 8.855661, gender accuracy 0.937500, age diff 8.764215\n",
      "step 92200, loss 12.185908, gender accuracy 1.000000, age diff 12.131969\n",
      "step 92400, loss 6.905859, gender accuracy 0.937500, age diff 6.770309\n",
      "step 92600, loss 6.295298, gender accuracy 0.937500, age diff 6.217623\n",
      "step 92800, loss 6.736632, gender accuracy 1.000000, age diff 6.699733\n",
      "step 93000, loss 8.034398, gender accuracy 1.000000, age diff 7.988052\n",
      "step 93200, loss 7.431229, gender accuracy 1.000000, age diff 7.378236\n",
      "step 93400, loss 8.856481, gender accuracy 1.000000, age diff 8.814498\n",
      "step 93600, loss 7.315168, gender accuracy 0.875000, age diff 7.003119\n",
      "step 93800, loss 7.883167, gender accuracy 1.000000, age diff 7.879380\n",
      "step 94000, loss 7.794802, gender accuracy 1.000000, age diff 7.781798\n",
      "step 94200, loss 7.506242, gender accuracy 1.000000, age diff 7.481086\n",
      "step 94400, loss 8.497822, gender accuracy 0.937500, age diff 8.394861\n",
      "step 94600, loss 6.447110, gender accuracy 1.000000, age diff 6.438249\n",
      "step 94800, loss 8.280281, gender accuracy 1.000000, age diff 8.260287\n",
      "step 95000, loss 8.702009, gender accuracy 0.937500, age diff 8.620671\n",
      "step 95200, loss 9.143239, gender accuracy 1.000000, age diff 9.096945\n",
      "step 95400, loss 12.478332, gender accuracy 1.000000, age diff 12.390280\n",
      "step 95600, loss 5.869580, gender accuracy 1.000000, age diff 5.864602\n",
      "step 95800, loss 10.599372, gender accuracy 0.937500, age diff 10.415194\n",
      "step 96000, loss 5.475683, gender accuracy 0.937500, age diff 5.364843\n",
      "step 96200, loss 6.270426, gender accuracy 1.000000, age diff 6.252716\n",
      "step 96400, loss 8.033444, gender accuracy 1.000000, age diff 8.012280\n",
      "step 96600, loss 9.331240, gender accuracy 1.000000, age diff 9.315922\n",
      "step 96800, loss 7.238855, gender accuracy 0.812500, age diff 6.743460\n",
      "step 97000, loss 11.738919, gender accuracy 1.000000, age diff 11.704371\n",
      "step 97200, loss 8.832468, gender accuracy 0.937500, age diff 8.581057\n",
      "step 97400, loss 5.239452, gender accuracy 1.000000, age diff 5.212788\n",
      "step 97600, loss 9.101378, gender accuracy 1.000000, age diff 9.031362\n",
      "step 97800, loss 9.310697, gender accuracy 0.937500, age diff 9.164575\n",
      "step 98000, loss 13.226105, gender accuracy 1.000000, age diff 13.143158\n",
      "step 98200, loss 8.373183, gender accuracy 1.000000, age diff 8.359764\n",
      "step 98400, loss 6.615827, gender accuracy 1.000000, age diff 6.580515\n",
      "step 98600, loss 8.801106, gender accuracy 1.000000, age diff 8.762043\n",
      "step 98800, loss 4.724977, gender accuracy 1.000000, age diff 4.699203\n",
      "step 99000, loss 8.894608, gender accuracy 1.000000, age diff 8.879417\n",
      "step 99200, loss 7.752672, gender accuracy 1.000000, age diff 7.743001\n",
      "step 99400, loss 8.734789, gender accuracy 0.937500, age diff 8.550119\n",
      "step 99600, loss 7.500532, gender accuracy 1.000000, age diff 7.471962\n",
      "step 99800, loss 7.410612, gender accuracy 1.000000, age diff 7.377259\n",
      "step 100000, loss 6.467261, gender accuracy 1.000000, age diff 6.432238\n",
      "step 100200, loss 9.416068, gender accuracy 1.000000, age diff 9.403502\n",
      "step 100400, loss 10.230927, gender accuracy 0.937500, age diff 10.165088\n",
      "step 100600, loss 4.720317, gender accuracy 1.000000, age diff 4.710711\n",
      "step 100800, loss 7.876378, gender accuracy 1.000000, age diff 7.828991\n",
      "step 101000, loss 7.817132, gender accuracy 0.937500, age diff 7.748930\n",
      "step 101200, loss 8.135967, gender accuracy 1.000000, age diff 8.111522\n",
      "step 101400, loss 5.724264, gender accuracy 0.937500, age diff 5.625422\n",
      "step 101600, loss 6.813982, gender accuracy 1.000000, age diff 6.768532\n",
      "step 101800, loss 5.464871, gender accuracy 0.937500, age diff 5.399529\n",
      "step 102000, loss 9.541992, gender accuracy 1.000000, age diff 9.511141\n",
      "step 102200, loss 6.373493, gender accuracy 1.000000, age diff 6.328023\n",
      "step 102400, loss 7.988796, gender accuracy 0.875000, age diff 7.493563\n",
      "step 102600, loss 8.930606, gender accuracy 1.000000, age diff 8.909601\n",
      "step 102800, loss 13.112194, gender accuracy 0.937500, age diff 12.829147\n",
      "step 103000, loss 8.496799, gender accuracy 1.000000, age diff 8.470049\n",
      "step 103200, loss 5.901297, gender accuracy 1.000000, age diff 5.889619\n",
      "step 103400, loss 6.127522, gender accuracy 1.000000, age diff 6.081241\n",
      "step 103600, loss 7.324094, gender accuracy 0.937500, age diff 7.236875\n",
      "step 103800, loss 9.492054, gender accuracy 0.875000, age diff 9.324160\n",
      "step 104000, loss 7.031245, gender accuracy 0.937500, age diff 6.877767\n",
      "step 104200, loss 9.968976, gender accuracy 1.000000, age diff 9.957813\n",
      "step 104400, loss 7.288937, gender accuracy 1.000000, age diff 7.279234\n",
      "step 104600, loss 11.416087, gender accuracy 1.000000, age diff 11.386993\n",
      "step 104800, loss 9.569148, gender accuracy 1.000000, age diff 9.559990\n",
      "step 105000, loss 7.400176, gender accuracy 1.000000, age diff 7.382407\n",
      "step 105200, loss 7.502656, gender accuracy 1.000000, age diff 7.480820\n",
      "step 105400, loss 6.631547, gender accuracy 1.000000, age diff 6.589752\n",
      "step 105600, loss 6.337424, gender accuracy 1.000000, age diff 6.300455\n",
      "step 105800, loss 5.214951, gender accuracy 1.000000, age diff 5.178566\n",
      "step 106000, loss 8.332656, gender accuracy 0.875000, age diff 8.087339\n",
      "step 106200, loss 9.536281, gender accuracy 1.000000, age diff 9.493986\n",
      "step 106400, loss 9.449852, gender accuracy 1.000000, age diff 9.415102\n",
      "step 106600, loss 10.345377, gender accuracy 1.000000, age diff 10.335379\n",
      "step 106800, loss 9.420160, gender accuracy 0.937500, age diff 9.315514\n",
      "step 107000, loss 7.662450, gender accuracy 1.000000, age diff 7.652154\n",
      "step 107200, loss 7.385978, gender accuracy 1.000000, age diff 7.337548\n",
      "step 107400, loss 8.936904, gender accuracy 1.000000, age diff 8.916621\n",
      "step 107600, loss 9.437737, gender accuracy 1.000000, age diff 9.436497\n",
      "step 107800, loss 8.917146, gender accuracy 0.937500, age diff 8.843936\n",
      "step 108000, loss 6.336901, gender accuracy 1.000000, age diff 6.328404\n",
      "step 108200, loss 8.480047, gender accuracy 1.000000, age diff 8.454434\n",
      "step 108400, loss 8.345156, gender accuracy 1.000000, age diff 8.335399\n",
      "step 108600, loss 7.349751, gender accuracy 1.000000, age diff 7.322936\n",
      "step 108800, loss 9.278611, gender accuracy 1.000000, age diff 9.275416\n",
      "step 109000, loss 8.223559, gender accuracy 1.000000, age diff 8.129379\n",
      "step 109200, loss 10.185675, gender accuracy 0.937500, age diff 10.071976\n",
      "step 109400, loss 7.904036, gender accuracy 1.000000, age diff 7.894999\n",
      "step 109600, loss 6.515789, gender accuracy 1.000000, age diff 6.482415\n",
      "step 109800, loss 5.585203, gender accuracy 1.000000, age diff 5.579433\n",
      "step 110000, loss 9.403558, gender accuracy 1.000000, age diff 9.397706\n",
      "step 110200, loss 8.120396, gender accuracy 1.000000, age diff 8.080069\n",
      "step 110400, loss 13.067122, gender accuracy 0.875000, age diff 12.848186\n",
      "step 110600, loss 7.883053, gender accuracy 1.000000, age diff 7.878918\n",
      "step 110800, loss 8.462273, gender accuracy 1.000000, age diff 8.431268\n",
      "step 111000, loss 5.111498, gender accuracy 0.875000, age diff 4.927286\n",
      "step 111200, loss 6.019779, gender accuracy 0.937500, age diff 5.964845\n",
      "step 111400, loss 8.196794, gender accuracy 1.000000, age diff 8.187279\n",
      "step 111600, loss 6.686246, gender accuracy 1.000000, age diff 6.674544\n",
      "step 111800, loss 5.969684, gender accuracy 0.937500, age diff 5.846092\n",
      "step 112000, loss 7.964385, gender accuracy 1.000000, age diff 7.948784\n",
      "step 112200, loss 7.937032, gender accuracy 1.000000, age diff 7.915489\n",
      "step 112400, loss 10.053832, gender accuracy 1.000000, age diff 10.036095\n",
      "step 112600, loss 7.732958, gender accuracy 1.000000, age diff 7.722422\n",
      "step 112800, loss 6.749847, gender accuracy 0.937500, age diff 6.660213\n",
      "step 113000, loss 8.183860, gender accuracy 1.000000, age diff 8.137207\n",
      "step 113200, loss 10.488445, gender accuracy 0.937500, age diff 10.432021\n",
      "step 113400, loss 6.373752, gender accuracy 1.000000, age diff 6.312112\n",
      "step 113600, loss 7.910875, gender accuracy 0.937500, age diff 7.736264\n",
      "step 113800, loss 7.067945, gender accuracy 0.937500, age diff 6.890388\n",
      "step 114000, loss 6.249219, gender accuracy 0.937500, age diff 6.198450\n",
      "step 114200, loss 11.698049, gender accuracy 0.937500, age diff 11.283796\n",
      "step 114400, loss 6.056697, gender accuracy 1.000000, age diff 6.049964\n",
      "step 114600, loss 7.755356, gender accuracy 1.000000, age diff 7.694701\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "step 114800, loss 8.022942, gender accuracy 1.000000, age diff 8.014832\n",
      "step 115000, loss 7.353242, gender accuracy 0.812500, age diff 7.075165\n",
      "step 115200, loss 7.703521, gender accuracy 0.812500, age diff 7.426696\n",
      "step 115400, loss 5.819754, gender accuracy 0.937500, age diff 5.751472\n",
      "step 115600, loss 9.929306, gender accuracy 0.937500, age diff 9.779552\n",
      "step 115800, loss 9.026073, gender accuracy 1.000000, age diff 9.008993\n",
      "step 116000, loss 11.720901, gender accuracy 0.937500, age diff 11.542569\n",
      "step 116200, loss 5.934225, gender accuracy 0.937500, age diff 5.815005\n",
      "step 116400, loss 5.637980, gender accuracy 1.000000, age diff 5.625785\n",
      "step 116600, loss 9.964754, gender accuracy 1.000000, age diff 9.958314\n",
      "step 116800, loss 7.739575, gender accuracy 1.000000, age diff 7.728171\n",
      "step 117000, loss 5.892016, gender accuracy 0.937500, age diff 5.748090\n",
      "step 117200, loss 6.507704, gender accuracy 0.937500, age diff 6.409097\n",
      "step 117400, loss 7.593054, gender accuracy 1.000000, age diff 7.585675\n",
      "step 117600, loss 5.461774, gender accuracy 1.000000, age diff 5.459355\n",
      "step 117800, loss 9.834227, gender accuracy 0.875000, age diff 9.681142\n",
      "step 118000, loss 5.324884, gender accuracy 1.000000, age diff 5.292599\n",
      "step 118200, loss 7.969670, gender accuracy 1.000000, age diff 7.932468\n",
      "step 118400, loss 7.474912, gender accuracy 1.000000, age diff 7.434650\n",
      "step 118600, loss 6.652211, gender accuracy 0.937500, age diff 6.533562\n",
      "step 118800, loss 9.138669, gender accuracy 1.000000, age diff 9.115049\n",
      "step 119000, loss 8.509133, gender accuracy 1.000000, age diff 8.502360\n",
      "step 119200, loss 8.730253, gender accuracy 1.000000, age diff 8.684659\n",
      "step 119400, loss 6.986328, gender accuracy 0.937500, age diff 6.795421\n",
      "step 119600, loss 6.790951, gender accuracy 1.000000, age diff 6.781649\n",
      "step 119800, loss 9.518317, gender accuracy 1.000000, age diff 9.510509\n",
      "step 120000, loss 7.785068, gender accuracy 1.000000, age diff 7.728055\n",
      "step 120200, loss 11.846684, gender accuracy 1.000000, age diff 11.833464\n",
      "step 120400, loss 7.763768, gender accuracy 0.875000, age diff 7.133253\n",
      "step 120600, loss 9.363786, gender accuracy 0.937500, age diff 9.256760\n",
      "step 120800, loss 13.613531, gender accuracy 0.937500, age diff 13.489479\n",
      "step 121000, loss 7.150536, gender accuracy 0.937500, age diff 7.090047\n",
      "step 121200, loss 11.549548, gender accuracy 1.000000, age diff 11.524632\n",
      "step 121400, loss 9.362104, gender accuracy 1.000000, age diff 9.329695\n",
      "step 121600, loss 9.951057, gender accuracy 1.000000, age diff 9.949822\n",
      "step 121800, loss 9.323411, gender accuracy 1.000000, age diff 9.296694\n",
      "step 122000, loss 7.776781, gender accuracy 1.000000, age diff 7.764176\n",
      "step 122200, loss 7.674196, gender accuracy 1.000000, age diff 7.668306\n",
      "step 122400, loss 12.101207, gender accuracy 1.000000, age diff 12.100130\n",
      "step 122600, loss 9.903998, gender accuracy 1.000000, age diff 9.885893\n",
      "step 122800, loss 10.278339, gender accuracy 1.000000, age diff 10.232093\n",
      "step 123000, loss 9.877871, gender accuracy 1.000000, age diff 9.828529\n",
      "step 123200, loss 6.943605, gender accuracy 0.937500, age diff 6.856890\n",
      "step 123400, loss 11.878610, gender accuracy 0.875000, age diff 11.635912\n",
      "step 123600, loss 13.097604, gender accuracy 0.875000, age diff 12.937443\n",
      "step 123800, loss 9.177689, gender accuracy 0.875000, age diff 8.865417\n",
      "step 124000, loss 7.939459, gender accuracy 0.937500, age diff 7.849524\n",
      "step 124200, loss 9.620318, gender accuracy 1.000000, age diff 9.593616\n",
      "step 124400, loss 12.381903, gender accuracy 0.937500, age diff 12.309892\n",
      "step 124600, loss 7.537154, gender accuracy 1.000000, age diff 7.532702\n",
      "step 124800, loss 9.983539, gender accuracy 0.937500, age diff 9.917250\n",
      "step 125000, loss 5.800871, gender accuracy 0.937500, age diff 5.723061\n",
      "step 125200, loss 5.580505, gender accuracy 1.000000, age diff 5.571015\n",
      "step 125400, loss 6.699613, gender accuracy 0.937500, age diff 6.594698\n",
      "step 125600, loss 9.624057, gender accuracy 1.000000, age diff 9.612926\n",
      "step 125800, loss 8.434508, gender accuracy 1.000000, age diff 8.431163\n",
      "step 126000, loss 6.776042, gender accuracy 1.000000, age diff 6.732807\n",
      "step 126200, loss 12.489306, gender accuracy 0.937500, age diff 12.421552\n",
      "step 126400, loss 9.240544, gender accuracy 1.000000, age diff 9.227272\n",
      "step 126600, loss 8.302567, gender accuracy 0.937500, age diff 8.168530\n",
      "step 126800, loss 9.955949, gender accuracy 1.000000, age diff 9.945265\n",
      "step 127000, loss 6.202791, gender accuracy 1.000000, age diff 6.200891\n",
      "step 127200, loss 7.130637, gender accuracy 1.000000, age diff 7.114053\n",
      "step 127400, loss 8.461212, gender accuracy 1.000000, age diff 8.414554\n",
      "step 127600, loss 6.611822, gender accuracy 1.000000, age diff 6.589984\n",
      "step 127800, loss 8.983258, gender accuracy 1.000000, age diff 8.834805\n",
      "step 128000, loss 6.411950, gender accuracy 1.000000, age diff 6.366004\n",
      "step 128200, loss 7.286648, gender accuracy 0.937500, age diff 7.204175\n",
      "step 128400, loss 9.973957, gender accuracy 1.000000, age diff 9.956760\n",
      "step 128600, loss 9.446201, gender accuracy 1.000000, age diff 9.404010\n",
      "step 128800, loss 8.529717, gender accuracy 1.000000, age diff 8.468798\n",
      "step 129000, loss 7.208059, gender accuracy 0.937500, age diff 7.038812\n",
      "step 129200, loss 9.909969, gender accuracy 0.937500, age diff 9.753938\n",
      "step 129400, loss 4.821955, gender accuracy 1.000000, age diff 4.818007\n",
      "step 129600, loss 10.394133, gender accuracy 0.937500, age diff 10.254278\n",
      "step 129800, loss 11.186420, gender accuracy 1.000000, age diff 11.178749\n",
      "step 130000, loss 6.911587, gender accuracy 1.000000, age diff 6.877303\n",
      "step 130200, loss 8.695795, gender accuracy 1.000000, age diff 8.690206\n",
      "step 130400, loss 8.278426, gender accuracy 0.937500, age diff 8.200901\n",
      "step 130600, loss 7.092091, gender accuracy 1.000000, age diff 7.047921\n",
      "step 130800, loss 5.049638, gender accuracy 1.000000, age diff 5.048598\n",
      "step 131000, loss 7.393443, gender accuracy 0.937500, age diff 7.267163\n",
      "step 131200, loss 6.555209, gender accuracy 0.937500, age diff 6.493764\n",
      "step 131400, loss 6.809535, gender accuracy 0.937500, age diff 6.747956\n",
      "step 131600, loss 8.617050, gender accuracy 0.812500, age diff 8.262313\n",
      "step 131800, loss 12.691343, gender accuracy 0.937500, age diff 12.551046\n",
      "step 132000, loss 8.131072, gender accuracy 1.000000, age diff 8.077009\n",
      "step 132200, loss 6.169206, gender accuracy 1.000000, age diff 6.136646\n",
      "step 132400, loss 8.765679, gender accuracy 1.000000, age diff 8.762737\n",
      "step 132600, loss 6.871077, gender accuracy 1.000000, age diff 6.865297\n",
      "step 132800, loss 9.580225, gender accuracy 0.937500, age diff 9.314956\n",
      "step 133000, loss 12.859928, gender accuracy 1.000000, age diff 12.796741\n",
      "step 133200, loss 6.507154, gender accuracy 0.937500, age diff 6.412052\n",
      "step 133400, loss 7.019851, gender accuracy 1.000000, age diff 6.958360\n",
      "step 133600, loss 7.393270, gender accuracy 1.000000, age diff 7.348155\n",
      "step 133800, loss 6.029770, gender accuracy 0.937500, age diff 5.960619\n",
      "step 134000, loss 9.278811, gender accuracy 0.937500, age diff 9.194540\n",
      "step 134200, loss 10.485735, gender accuracy 0.937500, age diff 10.414963\n",
      "step 134400, loss 12.292670, gender accuracy 1.000000, age diff 12.282597\n",
      "step 134600, loss 6.258669, gender accuracy 1.000000, age diff 6.249837\n",
      "step 134800, loss 14.108031, gender accuracy 1.000000, age diff 14.066525\n",
      "step 135000, loss 6.851789, gender accuracy 1.000000, age diff 6.836770\n",
      "step 135200, loss 7.419209, gender accuracy 1.000000, age diff 7.398914\n",
      "step 135400, loss 8.604650, gender accuracy 1.000000, age diff 8.588568\n",
      "step 135600, loss 10.825152, gender accuracy 1.000000, age diff 10.760801\n",
      "step 135800, loss 7.903250, gender accuracy 1.000000, age diff 7.850943\n",
      "step 136000, loss 11.176673, gender accuracy 0.875000, age diff 10.992109\n",
      "step 136200, loss 6.994963, gender accuracy 1.000000, age diff 6.971577\n",
      "step 136400, loss 7.201356, gender accuracy 0.875000, age diff 6.635717\n",
      "step 136600, loss 5.143927, gender accuracy 1.000000, age diff 5.137273\n",
      "step 136800, loss 9.024045, gender accuracy 1.000000, age diff 9.013090\n",
      "step 137000, loss 6.598188, gender accuracy 0.937500, age diff 6.514213\n",
      "step 137200, loss 7.595739, gender accuracy 0.937500, age diff 7.543144\n",
      "step 137400, loss 7.704581, gender accuracy 1.000000, age diff 7.679840\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "step 137600, loss 9.615824, gender accuracy 1.000000, age diff 9.605628\n",
      "step 137800, loss 7.104520, gender accuracy 1.000000, age diff 7.104218\n",
      "step 138000, loss 6.938059, gender accuracy 1.000000, age diff 6.933138\n",
      "step 138200, loss 8.244151, gender accuracy 1.000000, age diff 8.234560\n",
      "step 138400, loss 9.890263, gender accuracy 1.000000, age diff 9.855974\n",
      "step 138600, loss 8.033718, gender accuracy 1.000000, age diff 7.986648\n",
      "step 138800, loss 9.229598, gender accuracy 0.937500, age diff 9.115723\n",
      "step 139000, loss 9.186263, gender accuracy 1.000000, age diff 9.176540\n",
      "step 139200, loss 14.196956, gender accuracy 1.000000, age diff 14.184628\n",
      "step 139400, loss 9.307894, gender accuracy 0.937500, age diff 9.197001\n",
      "step 139600, loss 6.724680, gender accuracy 0.875000, age diff 6.290101\n",
      "step 139800, loss 8.596079, gender accuracy 0.937500, age diff 8.525202\n",
      "step 140000, loss 6.985217, gender accuracy 1.000000, age diff 6.915755\n",
      "step 140200, loss 8.731502, gender accuracy 0.875000, age diff 8.528210\n",
      "step 140400, loss 8.767225, gender accuracy 1.000000, age diff 8.761041\n",
      "step 140600, loss 4.165331, gender accuracy 1.000000, age diff 4.160753\n",
      "step 140800, loss 4.244242, gender accuracy 1.000000, age diff 4.243377\n",
      "step 141000, loss 9.049849, gender accuracy 0.937500, age diff 8.989119\n",
      "step 141200, loss 7.944057, gender accuracy 0.937500, age diff 7.817429\n",
      "step 141400, loss 9.110743, gender accuracy 1.000000, age diff 9.085693\n",
      "step 141600, loss 8.999773, gender accuracy 1.000000, age diff 8.986974\n",
      "step 141800, loss 6.884628, gender accuracy 0.937500, age diff 6.794849\n",
      "step 142000, loss 5.076415, gender accuracy 1.000000, age diff 5.026496\n",
      "step 142200, loss 8.505802, gender accuracy 1.000000, age diff 8.497104\n",
      "step 142400, loss 6.554451, gender accuracy 1.000000, age diff 6.552830\n",
      "step 142600, loss 6.889077, gender accuracy 1.000000, age diff 6.887264\n",
      "step 142800, loss 6.428863, gender accuracy 0.937500, age diff 6.336046\n",
      "step 143000, loss 7.713716, gender accuracy 1.000000, age diff 7.704368\n",
      "step 143200, loss 9.511399, gender accuracy 0.937500, age diff 9.446390\n",
      "step 143400, loss 7.326405, gender accuracy 1.000000, age diff 7.296760\n",
      "step 143600, loss 6.954816, gender accuracy 1.000000, age diff 6.902312\n",
      "step 143800, loss 6.174830, gender accuracy 1.000000, age diff 6.172154\n",
      "step 144000, loss 8.780239, gender accuracy 1.000000, age diff 8.778483\n",
      "step 144200, loss 9.239666, gender accuracy 0.937500, age diff 8.978600\n",
      "step 144400, loss 7.865602, gender accuracy 1.000000, age diff 7.864159\n",
      "step 144600, loss 6.700373, gender accuracy 1.000000, age diff 6.691596\n",
      "step 144800, loss 7.901319, gender accuracy 1.000000, age diff 7.868590\n",
      "step 145000, loss 9.053486, gender accuracy 1.000000, age diff 9.041204\n",
      "step 145200, loss 11.004427, gender accuracy 1.000000, age diff 10.993243\n",
      "step 145400, loss 6.793063, gender accuracy 0.875000, age diff 6.468919\n",
      "step 145600, loss 9.061129, gender accuracy 1.000000, age diff 9.046287\n",
      "step 145800, loss 7.994762, gender accuracy 1.000000, age diff 7.990566\n",
      "step 146000, loss 8.884769, gender accuracy 0.937500, age diff 8.671526\n",
      "step 146200, loss 6.939902, gender accuracy 1.000000, age diff 6.933052\n",
      "step 146400, loss 10.559910, gender accuracy 1.000000, age diff 10.546212\n",
      "step 146600, loss 6.853438, gender accuracy 1.000000, age diff 6.848284\n",
      "step 146800, loss 5.568335, gender accuracy 1.000000, age diff 5.556666\n",
      "step 147000, loss 11.615272, gender accuracy 1.000000, age diff 11.610954\n",
      "step 147200, loss 9.353249, gender accuracy 0.937500, age diff 9.155781\n",
      "step 147400, loss 8.905217, gender accuracy 0.937500, age diff 8.820541\n",
      "step 147600, loss 7.057611, gender accuracy 1.000000, age diff 7.035732\n",
      "step 147800, loss 8.443978, gender accuracy 1.000000, age diff 8.413351\n",
      "step 148000, loss 6.700104, gender accuracy 1.000000, age diff 6.650734\n",
      "step 148200, loss 7.565063, gender accuracy 0.937500, age diff 7.436032\n",
      "step 148400, loss 8.391232, gender accuracy 0.937500, age diff 8.282659\n",
      "step 148600, loss 6.272211, gender accuracy 1.000000, age diff 6.251835\n",
      "step 148800, loss 9.336271, gender accuracy 0.937500, age diff 9.272710\n",
      "step 149000, loss 11.713034, gender accuracy 1.000000, age diff 11.654945\n",
      "step 149200, loss 8.075349, gender accuracy 1.000000, age diff 8.072377\n",
      "step 149400, loss 10.108410, gender accuracy 0.937500, age diff 10.052864\n",
      "step 149600, loss 5.698771, gender accuracy 1.000000, age diff 5.676601\n",
      "step 149800, loss 8.546127, gender accuracy 1.000000, age diff 8.533916\n",
      "step 150000, loss 8.261959, gender accuracy 1.000000, age diff 8.256132\n",
      "step 150200, loss 7.804937, gender accuracy 1.000000, age diff 7.800841\n",
      "step 150400, loss 9.455000, gender accuracy 1.000000, age diff 9.448568\n",
      "step 150600, loss 10.053722, gender accuracy 1.000000, age diff 9.999157\n",
      "step 150800, loss 6.090718, gender accuracy 1.000000, age diff 6.082122\n",
      "step 151000, loss 10.403648, gender accuracy 0.937500, age diff 10.242688\n",
      "step 151200, loss 7.943539, gender accuracy 1.000000, age diff 7.921726\n",
      "step 151400, loss 9.795653, gender accuracy 1.000000, age diff 9.774766\n",
      "step 151600, loss 8.606734, gender accuracy 1.000000, age diff 8.567861\n",
      "step 151800, loss 8.653669, gender accuracy 1.000000, age diff 8.642865\n",
      "step 152000, loss 7.489598, gender accuracy 1.000000, age diff 7.470166\n",
      "step 152200, loss 6.408316, gender accuracy 0.937500, age diff 6.308057\n",
      "step 152400, loss 8.705414, gender accuracy 1.000000, age diff 8.703299\n",
      "step 152600, loss 7.678449, gender accuracy 1.000000, age diff 7.665136\n",
      "step 152800, loss 7.523343, gender accuracy 0.937500, age diff 7.459122\n",
      "step 153000, loss 9.703411, gender accuracy 0.937500, age diff 9.527250\n",
      "step 153200, loss 6.990594, gender accuracy 1.000000, age diff 6.988858\n",
      "step 153400, loss 4.817841, gender accuracy 1.000000, age diff 4.746365\n",
      "step 153600, loss 9.985363, gender accuracy 0.937500, age diff 9.887623\n",
      "step 153800, loss 11.264313, gender accuracy 1.000000, age diff 11.220433\n",
      "step 154000, loss 5.979525, gender accuracy 1.000000, age diff 5.954935\n",
      "step 154200, loss 7.378742, gender accuracy 0.937500, age diff 7.262357\n",
      "step 154400, loss 7.220317, gender accuracy 1.000000, age diff 7.220107\n",
      "step 154600, loss 6.960050, gender accuracy 1.000000, age diff 6.950545\n",
      "step 154800, loss 8.356505, gender accuracy 1.000000, age diff 8.320930\n",
      "step 155000, loss 7.081497, gender accuracy 0.937500, age diff 7.018026\n",
      "step 155200, loss 8.124657, gender accuracy 1.000000, age diff 8.118634\n",
      "step 155400, loss 9.363220, gender accuracy 1.000000, age diff 9.318222\n",
      "step 155600, loss 8.430692, gender accuracy 0.937500, age diff 8.287560\n",
      "step 155800, loss 9.209276, gender accuracy 1.000000, age diff 9.185698\n",
      "step 156000, loss 8.267164, gender accuracy 1.000000, age diff 8.246360\n",
      "step 156200, loss 14.744223, gender accuracy 1.000000, age diff 14.727535\n",
      "step 156400, loss 10.678102, gender accuracy 1.000000, age diff 10.665809\n",
      "step 156600, loss 4.988894, gender accuracy 1.000000, age diff 4.986482\n",
      "step 156800, loss 6.865891, gender accuracy 1.000000, age diff 6.826118\n",
      "step 157000, loss 8.139163, gender accuracy 1.000000, age diff 8.119761\n",
      "step 157200, loss 9.047878, gender accuracy 1.000000, age diff 9.042798\n",
      "step 157400, loss 7.097970, gender accuracy 1.000000, age diff 7.096591\n",
      "step 157600, loss 9.475001, gender accuracy 1.000000, age diff 9.444728\n",
      "step 157800, loss 6.823523, gender accuracy 1.000000, age diff 6.801561\n",
      "step 158000, loss 8.741677, gender accuracy 0.875000, age diff 8.540930\n",
      "step 158200, loss 6.288701, gender accuracy 1.000000, age diff 6.257953\n",
      "step 158400, loss 11.232243, gender accuracy 0.875000, age diff 11.116640\n",
      "step 158600, loss 7.582782, gender accuracy 0.937500, age diff 7.505848\n",
      "step 158800, loss 8.960816, gender accuracy 0.937500, age diff 8.806166\n",
      "step 159000, loss 8.413373, gender accuracy 1.000000, age diff 8.371575\n",
      "step 159200, loss 8.400713, gender accuracy 1.000000, age diff 8.398920\n",
      "step 159400, loss 5.668423, gender accuracy 0.875000, age diff 5.480899\n",
      "step 159600, loss 9.946557, gender accuracy 1.000000, age diff 9.894415\n",
      "step 159800, loss 8.710013, gender accuracy 1.000000, age diff 8.695892\n",
      "step 160000, loss 6.216054, gender accuracy 1.000000, age diff 6.215449\n",
      "step 160200, loss 8.763988, gender accuracy 0.937500, age diff 8.601255\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "step 160400, loss 8.557920, gender accuracy 0.937500, age diff 8.490454\n",
      "step 160600, loss 8.455893, gender accuracy 1.000000, age diff 8.449660\n",
      "step 160800, loss 8.039919, gender accuracy 1.000000, age diff 8.033752\n",
      "step 161000, loss 8.488459, gender accuracy 1.000000, age diff 8.449979\n",
      "step 161200, loss 9.871346, gender accuracy 1.000000, age diff 9.842533\n",
      "step 161400, loss 8.612929, gender accuracy 0.937500, age diff 8.460356\n",
      "step 161600, loss 12.863434, gender accuracy 0.937500, age diff 12.730770\n",
      "step 161800, loss 11.109612, gender accuracy 1.000000, age diff 11.103559\n",
      "step 162000, loss 10.782780, gender accuracy 1.000000, age diff 10.749151\n",
      "step 162200, loss 10.008420, gender accuracy 0.937500, age diff 9.758904\n",
      "step 162400, loss 9.696149, gender accuracy 0.937500, age diff 9.571918\n",
      "step 162600, loss 9.744928, gender accuracy 0.875000, age diff 9.394464\n",
      "step 162800, loss 8.230595, gender accuracy 0.937500, age diff 8.177048\n",
      "step 163000, loss 7.142080, gender accuracy 1.000000, age diff 7.135099\n",
      "step 163200, loss 9.721646, gender accuracy 0.937500, age diff 9.622692\n",
      "step 163400, loss 9.078221, gender accuracy 0.875000, age diff 8.552646\n",
      "step 163600, loss 7.203641, gender accuracy 1.000000, age diff 7.191111\n",
      "step 163800, loss 8.637218, gender accuracy 1.000000, age diff 8.606635\n",
      "step 164000, loss 11.073902, gender accuracy 1.000000, age diff 11.058959\n",
      "step 164200, loss 5.763429, gender accuracy 1.000000, age diff 5.754313\n",
      "step 164400, loss 8.942609, gender accuracy 0.937500, age diff 8.798666\n",
      "step 164600, loss 7.918653, gender accuracy 1.000000, age diff 7.871365\n",
      "step 164800, loss 11.460435, gender accuracy 1.000000, age diff 11.442510\n",
      "step 165000, loss 8.563274, gender accuracy 0.937500, age diff 8.223161\n",
      "step 165200, loss 9.187124, gender accuracy 0.875000, age diff 8.954637\n",
      "step 165400, loss 10.473811, gender accuracy 0.937500, age diff 10.313908\n",
      "step 165600, loss 6.340656, gender accuracy 1.000000, age diff 6.277623\n",
      "step 165800, loss 7.613875, gender accuracy 0.875000, age diff 7.272019\n",
      "step 166000, loss 7.768472, gender accuracy 1.000000, age diff 7.747803\n",
      "step 166200, loss 5.900865, gender accuracy 1.000000, age diff 5.857899\n",
      "step 166400, loss 9.430291, gender accuracy 1.000000, age diff 9.423361\n",
      "step 166600, loss 7.345419, gender accuracy 1.000000, age diff 7.317749\n",
      "step 166800, loss 8.597228, gender accuracy 1.000000, age diff 8.593409\n",
      "step 167000, loss 7.590113, gender accuracy 1.000000, age diff 7.581884\n",
      "step 167200, loss 6.766290, gender accuracy 1.000000, age diff 6.725199\n",
      "step 167400, loss 6.384877, gender accuracy 0.937500, age diff 6.248215\n",
      "step 167600, loss 5.580394, gender accuracy 1.000000, age diff 5.556723\n",
      "step 167800, loss 7.324903, gender accuracy 1.000000, age diff 7.321047\n",
      "step 168000, loss 9.116043, gender accuracy 1.000000, age diff 9.077657\n",
      "step 168200, loss 9.497367, gender accuracy 1.000000, age diff 9.476034\n",
      "step 168400, loss 7.470304, gender accuracy 0.937500, age diff 7.358855\n",
      "step 168600, loss 6.940229, gender accuracy 1.000000, age diff 6.937256\n",
      "step 168800, loss 6.173012, gender accuracy 1.000000, age diff 6.171173\n",
      "step 169000, loss 7.248559, gender accuracy 0.937500, age diff 7.094544\n",
      "step 169200, loss 6.620282, gender accuracy 0.937500, age diff 6.515200\n",
      "step 169400, loss 7.898011, gender accuracy 1.000000, age diff 7.831225\n",
      "step 169600, loss 8.909806, gender accuracy 1.000000, age diff 8.858093\n",
      "step 169800, loss 8.808673, gender accuracy 0.937500, age diff 8.602283\n",
      "step 170000, loss 9.848458, gender accuracy 1.000000, age diff 9.803429\n",
      "step 170200, loss 7.933381, gender accuracy 0.937500, age diff 7.790997\n",
      "step 170400, loss 7.179498, gender accuracy 1.000000, age diff 7.122337\n",
      "step 170600, loss 9.964635, gender accuracy 0.937500, age diff 9.736446\n",
      "step 170800, loss 7.099435, gender accuracy 1.000000, age diff 7.095982\n",
      "step 171000, loss 9.175014, gender accuracy 1.000000, age diff 9.172676\n",
      "step 171200, loss 14.148390, gender accuracy 1.000000, age diff 14.093594\n",
      "step 171400, loss 6.103836, gender accuracy 1.000000, age diff 6.083580\n",
      "step 171600, loss 8.117194, gender accuracy 1.000000, age diff 8.082897\n",
      "step 171800, loss 5.697287, gender accuracy 1.000000, age diff 5.695206\n",
      "step 172000, loss 11.118688, gender accuracy 1.000000, age diff 11.113128\n",
      "step 172200, loss 7.389159, gender accuracy 0.937500, age diff 7.239347\n",
      "step 172400, loss 9.339672, gender accuracy 1.000000, age diff 9.337729\n",
      "step 172600, loss 5.133790, gender accuracy 1.000000, age diff 5.130721\n",
      "step 172800, loss 8.205797, gender accuracy 0.937500, age diff 8.022605\n",
      "step 173000, loss 6.088538, gender accuracy 1.000000, age diff 6.086539\n",
      "step 173200, loss 7.659359, gender accuracy 0.937500, age diff 7.586005\n",
      "step 173400, loss 7.760313, gender accuracy 0.937500, age diff 7.687898\n",
      "step 173600, loss 6.081782, gender accuracy 1.000000, age diff 6.080007\n",
      "step 173800, loss 9.863470, gender accuracy 1.000000, age diff 9.848946\n",
      "step 174000, loss 9.048565, gender accuracy 1.000000, age diff 9.033570\n",
      "step 174200, loss 5.130873, gender accuracy 1.000000, age diff 5.129843\n",
      "step 174400, loss 8.548085, gender accuracy 0.937500, age diff 8.459410\n",
      "step 174600, loss 6.288393, gender accuracy 1.000000, age diff 6.283740\n",
      "step 174800, loss 7.919622, gender accuracy 0.937500, age diff 7.723314\n",
      "step 175000, loss 8.857099, gender accuracy 1.000000, age diff 8.837761\n",
      "step 175200, loss 5.790643, gender accuracy 0.875000, age diff 5.538416\n",
      "step 175400, loss 6.897979, gender accuracy 1.000000, age diff 6.889480\n",
      "step 175600, loss 7.903949, gender accuracy 1.000000, age diff 7.878229\n",
      "step 175800, loss 7.083535, gender accuracy 1.000000, age diff 7.075974\n",
      "step 176000, loss 8.620897, gender accuracy 0.937500, age diff 8.489676\n",
      "step 176200, loss 8.300529, gender accuracy 0.937500, age diff 8.216975\n",
      "step 176400, loss 9.266898, gender accuracy 1.000000, age diff 9.246203\n",
      "step 176600, loss 11.394277, gender accuracy 1.000000, age diff 11.361598\n",
      "step 176800, loss 9.394984, gender accuracy 1.000000, age diff 9.371500\n",
      "step 177000, loss 8.271034, gender accuracy 1.000000, age diff 8.268486\n",
      "step 177200, loss 12.250018, gender accuracy 0.937500, age diff 12.028669\n",
      "step 177400, loss 8.244469, gender accuracy 1.000000, age diff 8.181589\n",
      "step 177600, loss 8.020767, gender accuracy 1.000000, age diff 8.020054\n",
      "step 177800, loss 8.734465, gender accuracy 1.000000, age diff 8.719416\n",
      "step 178000, loss 6.518360, gender accuracy 1.000000, age diff 6.514336\n",
      "step 178200, loss 5.316280, gender accuracy 1.000000, age diff 5.314775\n",
      "step 178400, loss 9.643471, gender accuracy 1.000000, age diff 9.642592\n",
      "step 178600, loss 5.828476, gender accuracy 1.000000, age diff 5.824690\n",
      "step 178800, loss 8.490821, gender accuracy 1.000000, age diff 8.427329\n",
      "step 179000, loss 11.290755, gender accuracy 1.000000, age diff 11.270388\n",
      "step 179200, loss 7.059848, gender accuracy 1.000000, age diff 7.021748\n",
      "step 179400, loss 8.136628, gender accuracy 0.937500, age diff 8.035072\n",
      "step 179600, loss 6.663103, gender accuracy 1.000000, age diff 6.652157\n",
      "step 179800, loss 8.771205, gender accuracy 1.000000, age diff 8.768243\n",
      "step 180000, loss 8.744972, gender accuracy 0.875000, age diff 8.638916\n",
      "step 180200, loss 10.831136, gender accuracy 0.937500, age diff 10.704870\n",
      "step 180400, loss 6.587467, gender accuracy 1.000000, age diff 6.580476\n",
      "step 180600, loss 6.343978, gender accuracy 1.000000, age diff 6.327706\n",
      "step 180800, loss 8.794604, gender accuracy 1.000000, age diff 8.751173\n",
      "step 181000, loss 9.254616, gender accuracy 1.000000, age diff 9.239268\n",
      "step 181200, loss 6.314358, gender accuracy 0.937500, age diff 6.151122\n",
      "step 181400, loss 9.938460, gender accuracy 1.000000, age diff 9.913492\n",
      "step 181600, loss 10.299479, gender accuracy 1.000000, age diff 10.283459\n",
      "step 181800, loss 10.042737, gender accuracy 1.000000, age diff 10.033734\n",
      "step 182000, loss 8.557540, gender accuracy 1.000000, age diff 8.522440\n",
      "step 182200, loss 9.652408, gender accuracy 0.937500, age diff 9.566177\n",
      "step 182400, loss 5.177939, gender accuracy 1.000000, age diff 5.112818\n",
      "step 182600, loss 10.806955, gender accuracy 1.000000, age diff 10.800917\n",
      "step 182800, loss 6.488940, gender accuracy 1.000000, age diff 6.480366\n",
      "step 183000, loss 7.751612, gender accuracy 0.937500, age diff 7.624624\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "step 183200, loss 6.731630, gender accuracy 0.937500, age diff 6.658477\n",
      "step 183400, loss 10.153525, gender accuracy 1.000000, age diff 10.058887\n",
      "step 183600, loss 6.087200, gender accuracy 1.000000, age diff 6.064619\n",
      "step 183800, loss 5.958903, gender accuracy 1.000000, age diff 5.949659\n",
      "step 184000, loss 5.125942, gender accuracy 1.000000, age diff 5.107876\n",
      "step 184200, loss 8.714322, gender accuracy 1.000000, age diff 8.660130\n",
      "step 184400, loss 8.119495, gender accuracy 1.000000, age diff 8.115316\n",
      "step 184600, loss 9.706729, gender accuracy 1.000000, age diff 9.700070\n",
      "step 184800, loss 8.585682, gender accuracy 1.000000, age diff 8.583386\n",
      "step 185000, loss 6.921061, gender accuracy 1.000000, age diff 6.882768\n",
      "step 185200, loss 6.917100, gender accuracy 1.000000, age diff 6.914996\n",
      "step 185400, loss 6.067611, gender accuracy 1.000000, age diff 6.065495\n",
      "step 185600, loss 5.562921, gender accuracy 1.000000, age diff 5.509757\n",
      "step 185800, loss 8.839876, gender accuracy 1.000000, age diff 8.838106\n",
      "step 186000, loss 8.682486, gender accuracy 1.000000, age diff 8.664808\n",
      "step 186200, loss 7.008807, gender accuracy 1.000000, age diff 6.982947\n",
      "step 186400, loss 8.269827, gender accuracy 1.000000, age diff 8.258631\n",
      "step 186600, loss 6.726774, gender accuracy 1.000000, age diff 6.722114\n",
      "step 186800, loss 6.782200, gender accuracy 0.937500, age diff 6.721179\n",
      "step 187000, loss 7.583983, gender accuracy 1.000000, age diff 7.576203\n",
      "step 187200, loss 10.258929, gender accuracy 1.000000, age diff 10.250465\n",
      "step 187400, loss 7.498681, gender accuracy 1.000000, age diff 7.492508\n",
      "step 187600, loss 12.439420, gender accuracy 1.000000, age diff 12.407612\n",
      "step 187800, loss 6.844029, gender accuracy 1.000000, age diff 6.840669\n",
      "step 188000, loss 5.765983, gender accuracy 1.000000, age diff 5.755096\n",
      "step 188200, loss 5.737592, gender accuracy 1.000000, age diff 5.712668\n",
      "step 188400, loss 8.286836, gender accuracy 1.000000, age diff 8.264217\n",
      "step 188600, loss 8.746025, gender accuracy 0.937500, age diff 8.476918\n",
      "step 188800, loss 5.344805, gender accuracy 1.000000, age diff 5.342647\n",
      "step 189000, loss 8.748279, gender accuracy 0.937500, age diff 8.645491\n",
      "step 189200, loss 7.105865, gender accuracy 1.000000, age diff 7.076683\n",
      "step 189400, loss 7.002166, gender accuracy 0.937500, age diff 6.897015\n",
      "step 189600, loss 5.905755, gender accuracy 1.000000, age diff 5.897815\n",
      "step 189800, loss 9.622121, gender accuracy 1.000000, age diff 9.619614\n",
      "step 190000, loss 7.644209, gender accuracy 1.000000, age diff 7.635423\n",
      "step 190200, loss 8.726079, gender accuracy 1.000000, age diff 8.703241\n",
      "step 190400, loss 6.417935, gender accuracy 1.000000, age diff 6.393334\n",
      "step 190600, loss 6.909099, gender accuracy 0.875000, age diff 6.733144\n",
      "step 190800, loss 6.711022, gender accuracy 1.000000, age diff 6.702312\n",
      "step 191000, loss 8.038692, gender accuracy 1.000000, age diff 8.034231\n",
      "step 191200, loss 10.036996, gender accuracy 1.000000, age diff 10.024697\n",
      "step 191400, loss 6.915998, gender accuracy 1.000000, age diff 6.904511\n",
      "step 191600, loss 6.344744, gender accuracy 1.000000, age diff 6.344377\n",
      "step 191800, loss 10.663514, gender accuracy 1.000000, age diff 10.661005\n",
      "step 192000, loss 10.251433, gender accuracy 1.000000, age diff 10.238987\n",
      "step 192200, loss 8.894031, gender accuracy 0.937500, age diff 8.827406\n",
      "step 192400, loss 10.304559, gender accuracy 1.000000, age diff 10.291872\n",
      "step 192600, loss 8.877758, gender accuracy 1.000000, age diff 8.876254\n",
      "step 192800, loss 8.256572, gender accuracy 0.937500, age diff 8.191458\n",
      "step 193000, loss 6.810865, gender accuracy 1.000000, age diff 6.809363\n",
      "step 193200, loss 8.896144, gender accuracy 0.875000, age diff 8.614668\n",
      "step 193400, loss 5.866415, gender accuracy 1.000000, age diff 5.842484\n",
      "step 193600, loss 9.205517, gender accuracy 1.000000, age diff 9.157883\n",
      "step 193800, loss 6.275729, gender accuracy 1.000000, age diff 6.267649\n",
      "step 194000, loss 3.404477, gender accuracy 1.000000, age diff 3.389656\n",
      "step 194200, loss 6.258561, gender accuracy 1.000000, age diff 6.236271\n",
      "step 194400, loss 10.665432, gender accuracy 0.937500, age diff 10.571636\n",
      "step 194600, loss 7.007214, gender accuracy 0.937500, age diff 6.749454\n",
      "step 194800, loss 5.695860, gender accuracy 1.000000, age diff 5.679490\n",
      "step 195000, loss 7.324111, gender accuracy 1.000000, age diff 7.318556\n",
      "step 195200, loss 7.621419, gender accuracy 1.000000, age diff 7.617757\n",
      "step 195400, loss 4.923414, gender accuracy 1.000000, age diff 4.914629\n",
      "step 195600, loss 7.579553, gender accuracy 0.937500, age diff 7.389552\n",
      "step 195800, loss 5.503350, gender accuracy 1.000000, age diff 5.475663\n",
      "step 196000, loss 7.289141, gender accuracy 1.000000, age diff 7.260796\n",
      "step 196200, loss 5.959826, gender accuracy 1.000000, age diff 5.935042\n",
      "step 196400, loss 11.333404, gender accuracy 0.937500, age diff 11.054508\n",
      "step 196600, loss 5.677147, gender accuracy 1.000000, age diff 5.661432\n",
      "step 196800, loss 8.483445, gender accuracy 1.000000, age diff 8.433319\n",
      "step 197000, loss 6.882430, gender accuracy 1.000000, age diff 6.858825\n",
      "step 197200, loss 9.966264, gender accuracy 1.000000, age diff 9.931039\n",
      "step 197400, loss 9.945244, gender accuracy 1.000000, age diff 9.936743\n",
      "step 197600, loss 9.423821, gender accuracy 0.937500, age diff 9.311188\n",
      "step 197800, loss 10.774042, gender accuracy 1.000000, age diff 10.758724\n",
      "step 198000, loss 8.535386, gender accuracy 0.937500, age diff 8.453260\n",
      "step 198200, loss 7.652236, gender accuracy 0.937500, age diff 7.574062\n",
      "step 198400, loss 10.279028, gender accuracy 1.000000, age diff 10.235576\n",
      "step 198600, loss 6.980782, gender accuracy 0.937500, age diff 6.849851\n",
      "step 198800, loss 6.447334, gender accuracy 0.937500, age diff 6.388296\n",
      "step 199000, loss 7.161410, gender accuracy 0.937500, age diff 7.092002\n",
      "step 199200, loss 10.185787, gender accuracy 0.750000, age diff 9.893141\n",
      "step 199400, loss 6.270284, gender accuracy 1.000000, age diff 6.256597\n",
      "step 199600, loss 6.418670, gender accuracy 1.000000, age diff 6.383873\n",
      "step 199800, loss 8.126942, gender accuracy 1.000000, age diff 8.123451\n",
      "step 200000, loss 7.974893, gender accuracy 1.000000, age diff 7.934774\n",
      "step 200200, loss 8.483782, gender accuracy 1.000000, age diff 8.451902\n",
      "step 200400, loss 6.690321, gender accuracy 1.000000, age diff 6.642045\n",
      "step 200600, loss 6.336193, gender accuracy 1.000000, age diff 6.306432\n",
      "step 200800, loss 6.653264, gender accuracy 1.000000, age diff 6.652342\n",
      "step 201000, loss 6.949142, gender accuracy 1.000000, age diff 6.939923\n",
      "step 201200, loss 5.763567, gender accuracy 1.000000, age diff 5.757537\n",
      "step 201400, loss 6.554215, gender accuracy 1.000000, age diff 6.553177\n",
      "step 201600, loss 7.362416, gender accuracy 0.875000, age diff 7.161382\n",
      "step 201800, loss 7.798212, gender accuracy 1.000000, age diff 7.791096\n",
      "step 202000, loss 11.788491, gender accuracy 1.000000, age diff 11.772835\n",
      "step 202200, loss 8.291223, gender accuracy 0.937500, age diff 8.102807\n",
      "step 202400, loss 6.470019, gender accuracy 1.000000, age diff 6.463590\n",
      "step 202600, loss 7.591300, gender accuracy 1.000000, age diff 7.547256\n",
      "step 202800, loss 8.287903, gender accuracy 0.937500, age diff 8.215034\n",
      "step 203000, loss 7.945733, gender accuracy 1.000000, age diff 7.939001\n",
      "step 203200, loss 5.100916, gender accuracy 1.000000, age diff 5.089708\n",
      "step 203400, loss 7.486720, gender accuracy 1.000000, age diff 7.467712\n",
      "step 203600, loss 4.946569, gender accuracy 1.000000, age diff 4.940855\n",
      "step 203800, loss 7.746695, gender accuracy 1.000000, age diff 7.699522\n",
      "step 204000, loss 6.657220, gender accuracy 1.000000, age diff 6.632205\n",
      "step 204200, loss 9.268340, gender accuracy 1.000000, age diff 9.266905\n",
      "step 204400, loss 9.054580, gender accuracy 1.000000, age diff 9.054123\n",
      "step 204600, loss 7.337810, gender accuracy 1.000000, age diff 7.332010\n",
      "step 204800, loss 6.453564, gender accuracy 0.937500, age diff 6.356372\n",
      "step 205000, loss 5.333707, gender accuracy 1.000000, age diff 5.319791\n",
      "step 205200, loss 6.198360, gender accuracy 1.000000, age diff 6.195518\n",
      "step 205400, loss 7.583227, gender accuracy 1.000000, age diff 7.562242\n",
      "step 205600, loss 8.964543, gender accuracy 1.000000, age diff 8.949209\n",
      "step 205800, loss 7.994642, gender accuracy 1.000000, age diff 7.981892\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "step 206000, loss 8.177800, gender accuracy 1.000000, age diff 8.168010\n",
      "step 206200, loss 9.577551, gender accuracy 0.937500, age diff 9.424158\n",
      "step 206400, loss 8.224463, gender accuracy 1.000000, age diff 8.186447\n",
      "step 206600, loss 9.059581, gender accuracy 1.000000, age diff 9.050961\n",
      "step 206800, loss 10.735450, gender accuracy 1.000000, age diff 10.729000\n",
      "step 207000, loss 6.486341, gender accuracy 1.000000, age diff 6.453012\n",
      "step 207200, loss 7.464792, gender accuracy 0.937500, age diff 7.356671\n",
      "step 207400, loss 10.924529, gender accuracy 0.937500, age diff 10.831403\n",
      "step 207600, loss 7.665483, gender accuracy 1.000000, age diff 7.664765\n",
      "step 207800, loss 8.530147, gender accuracy 1.000000, age diff 8.514331\n",
      "step 208000, loss 8.303489, gender accuracy 1.000000, age diff 8.291576\n",
      "step 208200, loss 9.095933, gender accuracy 1.000000, age diff 9.051491\n",
      "step 208400, loss 7.374790, gender accuracy 1.000000, age diff 7.371163\n",
      "step 208600, loss 6.771121, gender accuracy 1.000000, age diff 6.760572\n",
      "step 208800, loss 6.854872, gender accuracy 1.000000, age diff 6.830676\n",
      "step 209000, loss 9.784907, gender accuracy 0.937500, age diff 9.617695\n",
      "step 209200, loss 7.843446, gender accuracy 1.000000, age diff 7.840117\n",
      "step 209400, loss 9.970499, gender accuracy 1.000000, age diff 9.952771\n",
      "step 209600, loss 8.956243, gender accuracy 1.000000, age diff 8.952063\n",
      "step 209800, loss 6.632020, gender accuracy 1.000000, age diff 6.628834\n",
      "step 210000, loss 6.899589, gender accuracy 1.000000, age diff 6.876948\n",
      "step 210200, loss 7.627881, gender accuracy 1.000000, age diff 7.627043\n",
      "step 210400, loss 7.124344, gender accuracy 1.000000, age diff 7.120944\n",
      "step 210600, loss 10.779412, gender accuracy 1.000000, age diff 10.772572\n",
      "step 210800, loss 9.661773, gender accuracy 1.000000, age diff 9.649176\n",
      "step 211000, loss 6.666479, gender accuracy 1.000000, age diff 6.658371\n",
      "step 211200, loss 9.406600, gender accuracy 0.937500, age diff 9.345611\n",
      "step 211400, loss 6.087696, gender accuracy 0.937500, age diff 5.919241\n",
      "step 211600, loss 5.311758, gender accuracy 1.000000, age diff 5.245240\n",
      "step 211800, loss 9.945077, gender accuracy 1.000000, age diff 9.934688\n",
      "step 212000, loss 7.590534, gender accuracy 0.937500, age diff 7.496319\n",
      "step 212200, loss 4.872062, gender accuracy 1.000000, age diff 4.871776\n",
      "step 212400, loss 9.460711, gender accuracy 1.000000, age diff 9.444243\n",
      "step 212600, loss 9.183855, gender accuracy 1.000000, age diff 9.138210\n",
      "step 212800, loss 7.906644, gender accuracy 1.000000, age diff 7.905982\n",
      "step 213000, loss 8.161210, gender accuracy 1.000000, age diff 8.152824\n",
      "step 213200, loss 9.096013, gender accuracy 0.937500, age diff 8.934396\n",
      "step 213400, loss 8.187267, gender accuracy 1.000000, age diff 8.163588\n",
      "step 213600, loss 6.491309, gender accuracy 1.000000, age diff 6.477467\n",
      "step 213800, loss 6.078267, gender accuracy 0.937500, age diff 6.014400\n",
      "step 214000, loss 9.715373, gender accuracy 1.000000, age diff 9.712254\n",
      "step 214200, loss 7.328104, gender accuracy 1.000000, age diff 7.327044\n",
      "step 214400, loss 6.508832, gender accuracy 1.000000, age diff 6.500762\n",
      "step 214600, loss 6.925911, gender accuracy 1.000000, age diff 6.910658\n",
      "step 214800, loss 9.088861, gender accuracy 0.875000, age diff 8.898113\n",
      "step 215000, loss 8.098495, gender accuracy 0.937500, age diff 8.029143\n",
      "step 215200, loss 7.884272, gender accuracy 1.000000, age diff 7.877917\n",
      "step 215400, loss 8.274812, gender accuracy 1.000000, age diff 8.269868\n",
      "step 215600, loss 7.408253, gender accuracy 1.000000, age diff 7.405311\n",
      "step 215800, loss 8.875605, gender accuracy 1.000000, age diff 8.874222\n",
      "step 216000, loss 7.805577, gender accuracy 1.000000, age diff 7.759893\n",
      "step 216200, loss 6.913502, gender accuracy 1.000000, age diff 6.877789\n",
      "step 216400, loss 4.550076, gender accuracy 1.000000, age diff 4.541698\n",
      "step 216600, loss 4.513929, gender accuracy 0.937500, age diff 4.436097\n",
      "step 216800, loss 8.640188, gender accuracy 1.000000, age diff 8.616266\n",
      "step 217000, loss 8.434321, gender accuracy 0.875000, age diff 8.240267\n",
      "step 217200, loss 9.016756, gender accuracy 1.000000, age diff 9.009596\n",
      "step 217400, loss 11.590766, gender accuracy 1.000000, age diff 11.574566\n",
      "step 217600, loss 6.826863, gender accuracy 0.937500, age diff 6.751484\n",
      "step 217800, loss 8.121031, gender accuracy 0.937500, age diff 7.600131\n",
      "step 218000, loss 6.434853, gender accuracy 1.000000, age diff 6.428608\n",
      "step 218200, loss 9.713477, gender accuracy 0.937500, age diff 9.552934\n",
      "step 218400, loss 8.133430, gender accuracy 0.937500, age diff 7.794885\n",
      "step 218600, loss 12.565640, gender accuracy 1.000000, age diff 12.565065\n",
      "step 218800, loss 8.039912, gender accuracy 1.000000, age diff 8.036873\n",
      "step 219000, loss 7.309692, gender accuracy 1.000000, age diff 7.308287\n",
      "step 219200, loss 7.418450, gender accuracy 1.000000, age diff 7.389261\n",
      "step 219400, loss 7.115819, gender accuracy 1.000000, age diff 7.107424\n",
      "step 219600, loss 4.499118, gender accuracy 1.000000, age diff 4.496863\n",
      "step 219800, loss 6.368398, gender accuracy 1.000000, age diff 6.361457\n",
      "step 220000, loss 8.099215, gender accuracy 1.000000, age diff 8.072711\n",
      "step 220200, loss 7.011522, gender accuracy 1.000000, age diff 6.983877\n",
      "step 220400, loss 7.390154, gender accuracy 1.000000, age diff 7.388986\n",
      "step 220600, loss 8.884233, gender accuracy 0.937500, age diff 8.696196\n",
      "step 220800, loss 6.723676, gender accuracy 0.937500, age diff 6.654075\n",
      "step 221000, loss 6.483935, gender accuracy 1.000000, age diff 6.465827\n",
      "step 221200, loss 8.493301, gender accuracy 1.000000, age diff 8.490558\n",
      "step 221400, loss 10.307206, gender accuracy 0.937500, age diff 10.230846\n",
      "step 221600, loss 5.952089, gender accuracy 1.000000, age diff 5.951670\n",
      "step 221800, loss 4.176046, gender accuracy 0.937500, age diff 4.025640\n",
      "step 222000, loss 7.470011, gender accuracy 0.875000, age diff 7.148922\n",
      "step 222200, loss 7.410141, gender accuracy 0.937500, age diff 7.206267\n",
      "step 222400, loss 9.793465, gender accuracy 1.000000, age diff 9.771362\n",
      "step 222600, loss 5.218332, gender accuracy 1.000000, age diff 5.199233\n",
      "step 222800, loss 10.017040, gender accuracy 1.000000, age diff 10.004971\n",
      "step 223000, loss 10.000772, gender accuracy 0.937500, age diff 9.931243\n",
      "step 223200, loss 13.913670, gender accuracy 0.937500, age diff 13.841319\n",
      "step 223400, loss 7.342439, gender accuracy 1.000000, age diff 7.340656\n",
      "step 223600, loss 5.424413, gender accuracy 1.000000, age diff 5.419381\n",
      "step 223800, loss 9.246657, gender accuracy 0.937500, age diff 9.151552\n",
      "step 224000, loss 10.036133, gender accuracy 1.000000, age diff 10.021149\n",
      "step 224200, loss 8.985355, gender accuracy 1.000000, age diff 8.981047\n",
      "step 224400, loss 9.591271, gender accuracy 1.000000, age diff 9.573219\n",
      "step 224600, loss 7.021554, gender accuracy 1.000000, age diff 7.017260\n",
      "step 224800, loss 7.284433, gender accuracy 1.000000, age diff 7.232333\n",
      "step 225000, loss 7.914372, gender accuracy 1.000000, age diff 7.899258\n",
      "step 225200, loss 7.472564, gender accuracy 0.875000, age diff 7.341827\n",
      "step 225400, loss 8.466914, gender accuracy 1.000000, age diff 8.449872\n",
      "step 225600, loss 5.622181, gender accuracy 1.000000, age diff 5.600802\n",
      "step 225800, loss 8.019690, gender accuracy 1.000000, age diff 8.010129\n",
      "step 226000, loss 8.727772, gender accuracy 0.937500, age diff 8.570467\n",
      "step 226200, loss 8.639039, gender accuracy 0.875000, age diff 8.469942\n",
      "step 226400, loss 9.004264, gender accuracy 1.000000, age diff 8.972164\n",
      "step 226600, loss 7.381210, gender accuracy 1.000000, age diff 7.378675\n",
      "step 226800, loss 7.320917, gender accuracy 0.937500, age diff 7.260292\n",
      "step 227000, loss 7.343020, gender accuracy 1.000000, age diff 7.337664\n",
      "step 227200, loss 7.960525, gender accuracy 0.937500, age diff 7.792017\n",
      "step 227400, loss 6.133419, gender accuracy 1.000000, age diff 6.118584\n",
      "step 227600, loss 6.951106, gender accuracy 1.000000, age diff 6.947649\n",
      "step 227800, loss 11.152567, gender accuracy 1.000000, age diff 11.151102\n",
      "step 228000, loss 7.363093, gender accuracy 1.000000, age diff 7.350860\n",
      "step 228200, loss 7.285684, gender accuracy 0.937500, age diff 7.203543\n",
      "step 228400, loss 11.441843, gender accuracy 1.000000, age diff 11.387247\n",
      "step 228600, loss 8.945763, gender accuracy 1.000000, age diff 8.933638\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "step 228800, loss 7.311576, gender accuracy 1.000000, age diff 7.308314\n",
      "step 229000, loss 6.817112, gender accuracy 0.937500, age diff 6.691503\n",
      "step 229200, loss 8.376264, gender accuracy 1.000000, age diff 8.333528\n",
      "step 229400, loss 7.617914, gender accuracy 1.000000, age diff 7.607024\n",
      "step 229600, loss 8.252808, gender accuracy 1.000000, age diff 8.245115\n",
      "step 229800, loss 7.283063, gender accuracy 1.000000, age diff 7.277230\n",
      "step 230000, loss 6.912627, gender accuracy 1.000000, age diff 6.876839\n",
      "step 230200, loss 7.810472, gender accuracy 1.000000, age diff 7.786442\n",
      "step 230400, loss 7.506257, gender accuracy 1.000000, age diff 7.495827\n",
      "step 230600, loss 8.289922, gender accuracy 0.937500, age diff 8.155134\n",
      "step 230800, loss 8.722666, gender accuracy 1.000000, age diff 8.716920\n",
      "step 231000, loss 8.938951, gender accuracy 1.000000, age diff 8.933043\n",
      "step 231200, loss 9.180080, gender accuracy 1.000000, age diff 9.133675\n",
      "step 231400, loss 8.399033, gender accuracy 0.937500, age diff 8.348324\n",
      "step 231600, loss 7.499829, gender accuracy 1.000000, age diff 7.493259\n",
      "step 231800, loss 8.823865, gender accuracy 1.000000, age diff 8.819277\n",
      "step 232000, loss 10.007868, gender accuracy 1.000000, age diff 10.000202\n",
      "step 232200, loss 10.147485, gender accuracy 1.000000, age diff 10.143831\n",
      "step 232400, loss 5.811151, gender accuracy 0.937500, age diff 5.653175\n",
      "step 232600, loss 6.141419, gender accuracy 1.000000, age diff 6.137147\n",
      "step 232800, loss 6.839266, gender accuracy 1.000000, age diff 6.804956\n",
      "step 233000, loss 7.079090, gender accuracy 1.000000, age diff 7.059046\n",
      "step 233200, loss 5.447722, gender accuracy 1.000000, age diff 5.434955\n",
      "step 233400, loss 9.067218, gender accuracy 1.000000, age diff 9.015663\n",
      "step 233600, loss 7.126757, gender accuracy 1.000000, age diff 7.126130\n",
      "step 233800, loss 7.312626, gender accuracy 1.000000, age diff 7.302231\n",
      "step 234000, loss 5.828674, gender accuracy 1.000000, age diff 5.819679\n",
      "step 234200, loss 7.129565, gender accuracy 1.000000, age diff 7.098133\n",
      "step 234400, loss 10.286143, gender accuracy 1.000000, age diff 10.230108\n",
      "step 234600, loss 9.639795, gender accuracy 1.000000, age diff 9.603819\n",
      "step 234800, loss 6.542358, gender accuracy 1.000000, age diff 6.540574\n",
      "step 235000, loss 9.828703, gender accuracy 0.875000, age diff 9.623297\n",
      "step 235200, loss 8.026816, gender accuracy 1.000000, age diff 8.022533\n",
      "step 235400, loss 7.412311, gender accuracy 0.875000, age diff 7.164531\n",
      "step 235600, loss 9.160599, gender accuracy 1.000000, age diff 9.116272\n",
      "step 235800, loss 7.372421, gender accuracy 1.000000, age diff 7.371688\n",
      "step 236000, loss 7.939376, gender accuracy 1.000000, age diff 7.904112\n",
      "step 236200, loss 5.323104, gender accuracy 1.000000, age diff 5.303069\n",
      "step 236400, loss 8.555676, gender accuracy 1.000000, age diff 8.525230\n",
      "step 236600, loss 6.106333, gender accuracy 1.000000, age diff 6.092586\n",
      "step 236800, loss 5.027088, gender accuracy 1.000000, age diff 5.004408\n",
      "step 237000, loss 13.378172, gender accuracy 1.000000, age diff 13.344372\n",
      "step 237200, loss 8.219594, gender accuracy 1.000000, age diff 8.215682\n",
      "step 237400, loss 8.743218, gender accuracy 1.000000, age diff 8.734499\n",
      "step 237600, loss 10.472108, gender accuracy 1.000000, age diff 10.442597\n",
      "step 237800, loss 6.424614, gender accuracy 1.000000, age diff 6.422174\n",
      "step 238000, loss 7.046142, gender accuracy 1.000000, age diff 7.016466\n",
      "step 238200, loss 8.535529, gender accuracy 1.000000, age diff 8.514409\n",
      "step 238400, loss 5.838644, gender accuracy 1.000000, age diff 5.831810\n",
      "step 238600, loss 8.240870, gender accuracy 1.000000, age diff 8.202229\n",
      "step 238800, loss 8.509125, gender accuracy 1.000000, age diff 8.490684\n",
      "step 239000, loss 10.250507, gender accuracy 1.000000, age diff 10.248989\n",
      "step 239200, loss 6.648940, gender accuracy 0.937500, age diff 6.592871\n",
      "step 239400, loss 9.479750, gender accuracy 1.000000, age diff 9.458023\n",
      "step 239600, loss 8.567107, gender accuracy 0.937500, age diff 8.339010\n",
      "step 239800, loss 9.286630, gender accuracy 1.000000, age diff 9.260559\n",
      "step 240000, loss 9.919176, gender accuracy 1.000000, age diff 9.894012\n",
      "step 240200, loss 6.987680, gender accuracy 0.875000, age diff 6.847528\n",
      "step 240400, loss 7.206211, gender accuracy 1.000000, age diff 7.204042\n",
      "step 240600, loss 8.344220, gender accuracy 0.937500, age diff 8.228956\n",
      "step 240800, loss 9.558677, gender accuracy 1.000000, age diff 9.553411\n",
      "step 241000, loss 13.365889, gender accuracy 0.937500, age diff 13.292967\n",
      "step 241200, loss 9.900417, gender accuracy 1.000000, age diff 9.892243\n",
      "step 241400, loss 7.301629, gender accuracy 0.937500, age diff 7.203815\n",
      "step 241600, loss 5.039966, gender accuracy 1.000000, age diff 5.012299\n",
      "step 241800, loss 9.037868, gender accuracy 0.937500, age diff 8.938633\n",
      "step 242000, loss 6.611309, gender accuracy 1.000000, age diff 6.609749\n",
      "step 242200, loss 10.075634, gender accuracy 0.937500, age diff 9.906454\n",
      "step 242400, loss 7.875820, gender accuracy 1.000000, age diff 7.869702\n",
      "step 242600, loss 6.979279, gender accuracy 1.000000, age diff 6.961784\n",
      "step 242800, loss 5.903893, gender accuracy 1.000000, age diff 5.866402\n",
      "step 243000, loss 10.949978, gender accuracy 0.937500, age diff 10.683800\n",
      "step 243200, loss 5.865663, gender accuracy 1.000000, age diff 5.860662\n",
      "step 243400, loss 11.490166, gender accuracy 1.000000, age diff 11.471354\n",
      "step 243600, loss 5.638364, gender accuracy 1.000000, age diff 5.634520\n",
      "step 243800, loss 9.538598, gender accuracy 1.000000, age diff 9.536088\n",
      "step 244000, loss 6.795088, gender accuracy 1.000000, age diff 6.792705\n",
      "step 244200, loss 5.032220, gender accuracy 1.000000, age diff 4.993727\n",
      "step 244400, loss 5.917125, gender accuracy 1.000000, age diff 5.910244\n",
      "step 244600, loss 9.179068, gender accuracy 1.000000, age diff 9.133684\n",
      "step 244800, loss 6.163114, gender accuracy 1.000000, age diff 6.156878\n",
      "step 245000, loss 4.406912, gender accuracy 1.000000, age diff 4.383542\n",
      "step 245200, loss 10.547810, gender accuracy 1.000000, age diff 10.547131\n",
      "step 245400, loss 8.752276, gender accuracy 1.000000, age diff 8.752193\n",
      "step 245600, loss 8.962461, gender accuracy 1.000000, age diff 8.946072\n",
      "step 245800, loss 9.444133, gender accuracy 1.000000, age diff 9.443004\n",
      "step 246000, loss 8.265758, gender accuracy 1.000000, age diff 8.260656\n",
      "step 246200, loss 7.182673, gender accuracy 1.000000, age diff 7.159306\n",
      "step 246400, loss 7.713595, gender accuracy 0.937500, age diff 7.472000\n",
      "step 246600, loss 6.702435, gender accuracy 1.000000, age diff 6.676049\n",
      "step 246800, loss 10.158701, gender accuracy 0.937500, age diff 10.067489\n",
      "step 247000, loss 6.769854, gender accuracy 0.937500, age diff 6.681817\n",
      "step 247200, loss 7.850679, gender accuracy 1.000000, age diff 7.836459\n",
      "step 247400, loss 6.484038, gender accuracy 1.000000, age diff 6.471746\n",
      "step 247600, loss 7.195683, gender accuracy 1.000000, age diff 7.164121\n",
      "step 247800, loss 7.079176, gender accuracy 1.000000, age diff 7.072827\n",
      "step 248000, loss 8.416927, gender accuracy 0.937500, age diff 8.317836\n",
      "step 248200, loss 6.767372, gender accuracy 1.000000, age diff 6.750343\n",
      "step 248400, loss 8.208143, gender accuracy 0.937500, age diff 8.150709\n",
      "step 248600, loss 10.618667, gender accuracy 1.000000, age diff 10.609739\n",
      "step 248800, loss 10.266754, gender accuracy 0.875000, age diff 9.940186\n",
      "step 249000, loss 9.111186, gender accuracy 1.000000, age diff 9.066481\n",
      "step 249200, loss 9.366486, gender accuracy 1.000000, age diff 9.361528\n",
      "step 249400, loss 7.976363, gender accuracy 0.937500, age diff 7.793484\n",
      "step 249600, loss 7.137416, gender accuracy 0.937500, age diff 6.940046\n",
      "step 249800, loss 9.502142, gender accuracy 0.875000, age diff 8.922703\n",
      "step 250000, loss 6.395625, gender accuracy 0.937500, age diff 5.915856\n",
      "step 250200, loss 7.872195, gender accuracy 1.000000, age diff 7.867325\n",
      "step 250400, loss 7.951013, gender accuracy 1.000000, age diff 7.948568\n",
      "step 250600, loss 14.063849, gender accuracy 0.875000, age diff 13.808428\n",
      "step 250800, loss 8.636520, gender accuracy 1.000000, age diff 8.605478\n",
      "step 251000, loss 8.247431, gender accuracy 1.000000, age diff 8.243319\n",
      "step 251200, loss 9.381353, gender accuracy 1.000000, age diff 9.352326\n",
      "step 251400, loss 10.154384, gender accuracy 1.000000, age diff 10.124771\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "step 251600, loss 9.177280, gender accuracy 1.000000, age diff 9.174450\n",
      "step 251800, loss 6.143126, gender accuracy 0.937500, age diff 6.045507\n",
      "step 252000, loss 4.954158, gender accuracy 0.937500, age diff 4.820352\n",
      "step 252200, loss 7.685936, gender accuracy 0.937500, age diff 7.634644\n",
      "step 252400, loss 7.922551, gender accuracy 1.000000, age diff 7.914960\n",
      "step 252600, loss 9.099501, gender accuracy 1.000000, age diff 9.089077\n",
      "step 252800, loss 5.848606, gender accuracy 0.937500, age diff 5.745379\n",
      "step 253000, loss 8.400634, gender accuracy 1.000000, age diff 8.396500\n",
      "step 253200, loss 8.137135, gender accuracy 0.937500, age diff 8.053553\n",
      "step 253400, loss 8.413190, gender accuracy 1.000000, age diff 8.381584\n",
      "step 253600, loss 7.188932, gender accuracy 1.000000, age diff 7.173276\n",
      "step 253800, loss 7.964612, gender accuracy 0.875000, age diff 7.816613\n",
      "step 254000, loss 9.131379, gender accuracy 1.000000, age diff 9.086411\n",
      "step 254200, loss 6.107254, gender accuracy 1.000000, age diff 6.101339\n",
      "step 254400, loss 7.646430, gender accuracy 1.000000, age diff 7.645908\n",
      "step 254600, loss 8.672650, gender accuracy 1.000000, age diff 8.657112\n",
      "step 254800, loss 8.202245, gender accuracy 1.000000, age diff 8.189294\n",
      "step 255000, loss 7.306008, gender accuracy 0.937500, age diff 7.245174\n",
      "step 255200, loss 6.333570, gender accuracy 0.875000, age diff 6.052362\n",
      "step 255400, loss 7.530056, gender accuracy 1.000000, age diff 7.523108\n",
      "step 255600, loss 8.884652, gender accuracy 1.000000, age diff 8.832254\n",
      "step 255800, loss 11.182837, gender accuracy 1.000000, age diff 11.178751\n",
      "step 256000, loss 8.180589, gender accuracy 1.000000, age diff 8.164574\n",
      "step 256200, loss 5.736269, gender accuracy 0.937500, age diff 5.685875\n",
      "step 256400, loss 9.058989, gender accuracy 1.000000, age diff 9.056671\n",
      "step 256600, loss 6.832440, gender accuracy 1.000000, age diff 6.761097\n",
      "step 256800, loss 7.069302, gender accuracy 0.937500, age diff 6.939098\n",
      "step 257000, loss 7.082488, gender accuracy 1.000000, age diff 7.047847\n",
      "step 257200, loss 7.782001, gender accuracy 1.000000, age diff 7.780817\n",
      "step 257400, loss 9.601723, gender accuracy 0.937500, age diff 9.415379\n",
      "step 257600, loss 5.413681, gender accuracy 0.937500, age diff 5.317283\n",
      "step 257800, loss 6.901372, gender accuracy 1.000000, age diff 6.888435\n",
      "step 258000, loss 9.228710, gender accuracy 0.937500, age diff 9.167780\n",
      "step 258200, loss 9.475404, gender accuracy 1.000000, age diff 9.472862\n",
      "step 258400, loss 9.328800, gender accuracy 1.000000, age diff 9.316924\n",
      "step 258600, loss 10.862686, gender accuracy 0.875000, age diff 10.570473\n",
      "step 258800, loss 6.321765, gender accuracy 1.000000, age diff 6.319643\n",
      "step 259000, loss 7.759873, gender accuracy 1.000000, age diff 7.758679\n",
      "step 259200, loss 10.888775, gender accuracy 1.000000, age diff 10.874060\n",
      "step 259400, loss 7.343301, gender accuracy 1.000000, age diff 7.330456\n",
      "step 259600, loss 8.491897, gender accuracy 0.937500, age diff 8.267433\n",
      "step 259800, loss 7.196956, gender accuracy 0.937500, age diff 7.131474\n",
      "step 260000, loss 9.279975, gender accuracy 1.000000, age diff 9.271572\n",
      "step 260200, loss 8.797648, gender accuracy 0.812500, age diff 8.268872\n",
      "step 260400, loss 7.811798, gender accuracy 0.937500, age diff 7.729484\n",
      "step 260600, loss 9.082702, gender accuracy 0.937500, age diff 8.983803\n",
      "step 260800, loss 10.552416, gender accuracy 1.000000, age diff 10.499673\n",
      "step 261000, loss 10.425314, gender accuracy 0.937500, age diff 10.378313\n",
      "step 261200, loss 7.333384, gender accuracy 1.000000, age diff 7.331250\n",
      "step 261400, loss 5.380524, gender accuracy 1.000000, age diff 5.357842\n",
      "step 261600, loss 5.206444, gender accuracy 1.000000, age diff 5.179494\n",
      "step 261800, loss 7.627532, gender accuracy 0.937500, age diff 7.545828\n",
      "step 262000, loss 7.643034, gender accuracy 1.000000, age diff 7.636188\n",
      "step 262200, loss 7.445053, gender accuracy 0.937500, age diff 7.382282\n",
      "step 262400, loss 7.261174, gender accuracy 1.000000, age diff 7.260124\n",
      "step 262600, loss 8.813075, gender accuracy 1.000000, age diff 8.805952\n",
      "step 262800, loss 9.972329, gender accuracy 1.000000, age diff 9.962774\n",
      "step 263000, loss 5.233505, gender accuracy 0.937500, age diff 5.177989\n",
      "step 263200, loss 5.964453, gender accuracy 0.937500, age diff 5.883694\n",
      "step 263400, loss 7.532591, gender accuracy 1.000000, age diff 7.530046\n",
      "step 263600, loss 7.451490, gender accuracy 1.000000, age diff 7.432995\n",
      "step 263800, loss 6.882195, gender accuracy 0.875000, age diff 6.655364\n",
      "step 264000, loss 7.814436, gender accuracy 1.000000, age diff 7.762815\n",
      "step 264200, loss 9.334114, gender accuracy 1.000000, age diff 9.296124\n",
      "step 264400, loss 8.640154, gender accuracy 0.937500, age diff 8.582645\n",
      "step 264600, loss 9.529600, gender accuracy 1.000000, age diff 9.528539\n",
      "step 264800, loss 7.274445, gender accuracy 1.000000, age diff 7.258282\n",
      "step 265000, loss 7.200013, gender accuracy 1.000000, age diff 7.198245\n",
      "step 265200, loss 8.434973, gender accuracy 1.000000, age diff 8.414327\n",
      "step 265400, loss 5.663892, gender accuracy 1.000000, age diff 5.630215\n",
      "step 265600, loss 6.951454, gender accuracy 1.000000, age diff 6.944764\n",
      "step 265800, loss 8.396178, gender accuracy 1.000000, age diff 8.359463\n",
      "step 266000, loss 6.871388, gender accuracy 1.000000, age diff 6.850342\n",
      "step 266200, loss 9.687610, gender accuracy 0.937500, age diff 9.542668\n",
      "step 266400, loss 7.172052, gender accuracy 0.937500, age diff 7.019052\n",
      "step 266600, loss 8.754777, gender accuracy 1.000000, age diff 8.753894\n",
      "step 266800, loss 7.533783, gender accuracy 1.000000, age diff 7.475414\n",
      "step 267000, loss 7.175023, gender accuracy 1.000000, age diff 7.161695\n",
      "step 267200, loss 10.005872, gender accuracy 0.875000, age diff 9.746173\n",
      "step 267400, loss 9.418699, gender accuracy 1.000000, age diff 9.404127\n",
      "step 267600, loss 9.546713, gender accuracy 1.000000, age diff 9.537775\n",
      "step 267800, loss 6.322299, gender accuracy 0.875000, age diff 6.077231\n",
      "step 268000, loss 6.606326, gender accuracy 0.937500, age diff 6.532142\n",
      "step 268200, loss 8.917109, gender accuracy 1.000000, age diff 8.917004\n",
      "step 268400, loss 11.466688, gender accuracy 0.875000, age diff 11.210434\n",
      "step 268600, loss 8.978081, gender accuracy 1.000000, age diff 8.948483\n",
      "step 268800, loss 9.070782, gender accuracy 0.937500, age diff 9.018843\n",
      "step 269000, loss 8.214236, gender accuracy 1.000000, age diff 8.208393\n",
      "step 269200, loss 8.121437, gender accuracy 1.000000, age diff 8.100309\n",
      "step 269400, loss 5.515823, gender accuracy 0.937500, age diff 5.456207\n",
      "step 269600, loss 7.942514, gender accuracy 1.000000, age diff 7.932469\n",
      "step 269800, loss 9.628269, gender accuracy 0.937500, age diff 9.488055\n",
      "step 270000, loss 6.175582, gender accuracy 1.000000, age diff 6.148279\n",
      "step 270200, loss 5.935362, gender accuracy 1.000000, age diff 5.929662\n",
      "step 270400, loss 6.867306, gender accuracy 1.000000, age diff 6.851068\n",
      "step 270600, loss 9.421705, gender accuracy 1.000000, age diff 9.384007\n",
      "step 270800, loss 9.531298, gender accuracy 1.000000, age diff 9.452816\n",
      "step 271000, loss 7.293723, gender accuracy 1.000000, age diff 7.267639\n",
      "step 271200, loss 5.319633, gender accuracy 1.000000, age diff 5.267080\n",
      "step 271400, loss 11.493483, gender accuracy 1.000000, age diff 11.487795\n",
      "step 271600, loss 5.754310, gender accuracy 0.937500, age diff 5.654405\n",
      "step 271800, loss 7.651546, gender accuracy 1.000000, age diff 7.641535\n",
      "step 272000, loss 9.468305, gender accuracy 1.000000, age diff 9.463657\n",
      "step 272200, loss 9.258546, gender accuracy 1.000000, age diff 9.214815\n",
      "step 272400, loss 8.223691, gender accuracy 1.000000, age diff 8.195713\n",
      "step 272600, loss 9.858692, gender accuracy 1.000000, age diff 9.848656\n",
      "step 272800, loss 8.211564, gender accuracy 1.000000, age diff 8.169472\n",
      "step 273000, loss 7.104413, gender accuracy 1.000000, age diff 7.103190\n",
      "step 273200, loss 7.865839, gender accuracy 1.000000, age diff 7.864428\n",
      "step 273400, loss 8.088142, gender accuracy 1.000000, age diff 8.083199\n",
      "step 273600, loss 7.252307, gender accuracy 1.000000, age diff 7.244150\n",
      "step 273800, loss 9.100216, gender accuracy 1.000000, age diff 9.096451\n",
      "step 274000, loss 7.293458, gender accuracy 1.000000, age diff 7.291717\n",
      "step 274200, loss 8.175042, gender accuracy 0.812500, age diff 8.001091\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "step 274400, loss 6.670650, gender accuracy 0.937500, age diff 6.604142\n",
      "step 274600, loss 7.659454, gender accuracy 1.000000, age diff 7.644194\n",
      "step 274800, loss 9.573100, gender accuracy 1.000000, age diff 9.571390\n",
      "step 275000, loss 8.796489, gender accuracy 1.000000, age diff 8.768267\n",
      "step 275200, loss 9.257168, gender accuracy 1.000000, age diff 9.254782\n",
      "step 275400, loss 9.707297, gender accuracy 1.000000, age diff 9.697639\n",
      "step 275600, loss 10.059240, gender accuracy 1.000000, age diff 10.049588\n",
      "step 275800, loss 9.057195, gender accuracy 1.000000, age diff 9.055000\n",
      "step 276000, loss 7.462698, gender accuracy 1.000000, age diff 7.459299\n",
      "step 276200, loss 9.889241, gender accuracy 1.000000, age diff 9.831132\n",
      "step 276400, loss 7.895123, gender accuracy 1.000000, age diff 7.881807\n",
      "step 276600, loss 8.012107, gender accuracy 1.000000, age diff 8.006328\n",
      "step 276800, loss 7.055398, gender accuracy 1.000000, age diff 7.052896\n",
      "step 277000, loss 7.275758, gender accuracy 1.000000, age diff 7.271384\n",
      "step 277200, loss 7.923656, gender accuracy 1.000000, age diff 7.918539\n",
      "step 277400, loss 8.076016, gender accuracy 0.937500, age diff 7.920177\n",
      "step 277600, loss 5.794302, gender accuracy 1.000000, age diff 5.777608\n",
      "step 277800, loss 7.306483, gender accuracy 0.937500, age diff 7.202170\n",
      "step 278000, loss 9.791549, gender accuracy 0.875000, age diff 9.345721\n",
      "step 278200, loss 6.038836, gender accuracy 1.000000, age diff 6.007905\n",
      "step 278400, loss 8.390044, gender accuracy 1.000000, age diff 8.365497\n",
      "step 278600, loss 8.874063, gender accuracy 0.937500, age diff 8.747650\n",
      "step 278800, loss 6.070538, gender accuracy 1.000000, age diff 6.065940\n",
      "step 279000, loss 7.334885, gender accuracy 1.000000, age diff 7.310753\n",
      "step 279200, loss 6.875444, gender accuracy 0.937500, age diff 6.821285\n",
      "step 279400, loss 10.776991, gender accuracy 1.000000, age diff 10.774604\n",
      "step 279600, loss 11.982840, gender accuracy 1.000000, age diff 11.960969\n",
      "step 279800, loss 7.282628, gender accuracy 0.937500, age diff 7.169864\n",
      "step 280000, loss 6.662892, gender accuracy 1.000000, age diff 6.614885\n",
      "step 280200, loss 8.217483, gender accuracy 0.937500, age diff 8.151472\n",
      "step 280400, loss 7.456007, gender accuracy 1.000000, age diff 7.440499\n",
      "step 280600, loss 5.692677, gender accuracy 1.000000, age diff 5.690084\n",
      "step 280800, loss 6.593702, gender accuracy 1.000000, age diff 6.590250\n",
      "step 281000, loss 7.480694, gender accuracy 1.000000, age diff 7.454286\n",
      "step 281200, loss 6.910612, gender accuracy 1.000000, age diff 6.902470\n",
      "step 281400, loss 10.566182, gender accuracy 1.000000, age diff 10.559319\n",
      "step 281600, loss 9.421804, gender accuracy 0.937500, age diff 9.174091\n",
      "step 281800, loss 7.037385, gender accuracy 1.000000, age diff 7.036654\n",
      "step 282000, loss 7.560024, gender accuracy 1.000000, age diff 7.510126\n",
      "step 282200, loss 6.777657, gender accuracy 1.000000, age diff 6.769631\n",
      "step 282400, loss 10.074259, gender accuracy 0.937500, age diff 9.965788\n",
      "step 282600, loss 8.493738, gender accuracy 0.937500, age diff 8.366518\n",
      "step 282800, loss 9.511621, gender accuracy 1.000000, age diff 9.488317\n",
      "step 283000, loss 8.931302, gender accuracy 1.000000, age diff 8.928499\n",
      "step 283200, loss 10.453250, gender accuracy 0.937500, age diff 10.266797\n",
      "step 283400, loss 7.425556, gender accuracy 1.000000, age diff 7.409242\n",
      "step 283600, loss 11.414351, gender accuracy 1.000000, age diff 11.369688\n",
      "step 283800, loss 8.496432, gender accuracy 1.000000, age diff 8.484395\n",
      "step 284000, loss 7.275096, gender accuracy 0.937500, age diff 7.199906\n",
      "step 284200, loss 7.212637, gender accuracy 1.000000, age diff 7.180806\n",
      "step 284400, loss 8.533761, gender accuracy 1.000000, age diff 8.520425\n",
      "step 284600, loss 6.917760, gender accuracy 1.000000, age diff 6.916034\n",
      "step 284800, loss 8.723825, gender accuracy 0.937500, age diff 8.643072\n",
      "step 285000, loss 7.876999, gender accuracy 1.000000, age diff 7.857306\n",
      "step 285200, loss 10.085276, gender accuracy 1.000000, age diff 10.077726\n",
      "step 285400, loss 7.188955, gender accuracy 1.000000, age diff 7.134329\n",
      "step 285600, loss 8.590742, gender accuracy 1.000000, age diff 8.588055\n",
      "step 285800, loss 9.870788, gender accuracy 0.937500, age diff 9.811069\n",
      "step 286000, loss 9.977771, gender accuracy 1.000000, age diff 9.947880\n",
      "step 286200, loss 10.987851, gender accuracy 1.000000, age diff 10.965355\n",
      "step 286400, loss 9.610785, gender accuracy 1.000000, age diff 9.597252\n",
      "step 286600, loss 7.227264, gender accuracy 1.000000, age diff 7.191497\n",
      "step 286800, loss 9.016652, gender accuracy 1.000000, age diff 9.001344\n",
      "step 287000, loss 4.994239, gender accuracy 1.000000, age diff 4.989320\n",
      "step 287200, loss 6.191830, gender accuracy 1.000000, age diff 6.165989\n",
      "step 287400, loss 8.130143, gender accuracy 0.875000, age diff 7.911232\n",
      "step 287600, loss 7.968143, gender accuracy 1.000000, age diff 7.954070\n",
      "step 287800, loss 7.052571, gender accuracy 1.000000, age diff 7.032535\n",
      "step 288000, loss 5.688191, gender accuracy 1.000000, age diff 5.687452\n",
      "step 288200, loss 6.844571, gender accuracy 0.812500, age diff 6.552190\n",
      "step 288400, loss 8.281188, gender accuracy 1.000000, age diff 8.271337\n",
      "step 288600, loss 8.054494, gender accuracy 1.000000, age diff 8.054165\n",
      "step 288800, loss 10.596973, gender accuracy 0.937500, age diff 10.506037\n",
      "step 289000, loss 6.383417, gender accuracy 1.000000, age diff 6.382464\n",
      "step 289200, loss 11.718906, gender accuracy 1.000000, age diff 11.709118\n",
      "step 289400, loss 10.477090, gender accuracy 1.000000, age diff 10.466604\n",
      "step 289600, loss 10.805203, gender accuracy 0.937500, age diff 10.601413\n",
      "step 289800, loss 9.649763, gender accuracy 1.000000, age diff 9.643076\n",
      "step 290000, loss 8.179214, gender accuracy 1.000000, age diff 8.163204\n",
      "step 290200, loss 5.845927, gender accuracy 1.000000, age diff 5.839633\n",
      "step 290400, loss 7.173263, gender accuracy 1.000000, age diff 7.171253\n",
      "step 290600, loss 8.473869, gender accuracy 1.000000, age diff 8.471597\n",
      "step 290800, loss 7.934654, gender accuracy 1.000000, age diff 7.912626\n",
      "step 291000, loss 10.607347, gender accuracy 1.000000, age diff 10.564228\n",
      "step 291200, loss 7.643897, gender accuracy 1.000000, age diff 7.633350\n",
      "step 291400, loss 8.767943, gender accuracy 1.000000, age diff 8.761591\n",
      "step 291600, loss 7.999913, gender accuracy 0.937500, age diff 7.843078\n",
      "step 291800, loss 9.829245, gender accuracy 1.000000, age diff 9.806278\n",
      "step 292000, loss 6.722264, gender accuracy 1.000000, age diff 6.718801\n",
      "step 292200, loss 10.485450, gender accuracy 1.000000, age diff 10.390539\n",
      "step 292400, loss 8.414264, gender accuracy 1.000000, age diff 8.408083\n",
      "step 292600, loss 8.509404, gender accuracy 1.000000, age diff 8.507780\n",
      "step 292800, loss 10.464730, gender accuracy 0.937500, age diff 10.350603\n",
      "step 293000, loss 10.542014, gender accuracy 1.000000, age diff 10.536608\n",
      "step 293200, loss 9.038635, gender accuracy 1.000000, age diff 9.005936\n",
      "step 293400, loss 9.446659, gender accuracy 1.000000, age diff 9.407927\n",
      "step 293600, loss 8.844360, gender accuracy 0.937500, age diff 8.511845\n",
      "step 293800, loss 10.129354, gender accuracy 1.000000, age diff 10.129238\n",
      "step 294000, loss 7.860837, gender accuracy 1.000000, age diff 7.860413\n",
      "step 294200, loss 6.787797, gender accuracy 1.000000, age diff 6.786001\n",
      "step 294400, loss 7.847029, gender accuracy 1.000000, age diff 7.839443\n",
      "step 294600, loss 6.994220, gender accuracy 0.937500, age diff 6.893448\n",
      "step 294800, loss 9.190681, gender accuracy 0.937500, age diff 9.142334\n",
      "step 295000, loss 9.768710, gender accuracy 1.000000, age diff 9.744347\n",
      "step 295200, loss 7.299705, gender accuracy 1.000000, age diff 7.252835\n",
      "step 295400, loss 6.344508, gender accuracy 0.937500, age diff 6.297470\n",
      "step 295600, loss 6.699601, gender accuracy 1.000000, age diff 6.694759\n",
      "step 295800, loss 6.548653, gender accuracy 1.000000, age diff 6.523390\n",
      "step 296000, loss 9.200828, gender accuracy 1.000000, age diff 9.200380\n",
      "step 296200, loss 7.213267, gender accuracy 1.000000, age diff 7.193965\n",
      "step 296400, loss 8.217886, gender accuracy 1.000000, age diff 8.173814\n",
      "step 296600, loss 6.713902, gender accuracy 1.000000, age diff 6.703218\n",
      "step 296800, loss 10.933778, gender accuracy 1.000000, age diff 10.932581\n",
      "step 297000, loss 7.332269, gender accuracy 1.000000, age diff 7.330775\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "step 297200, loss 10.061082, gender accuracy 0.937500, age diff 9.942039\n",
      "step 297400, loss 8.756095, gender accuracy 1.000000, age diff 8.754551\n",
      "step 297600, loss 7.435338, gender accuracy 1.000000, age diff 7.421965\n",
      "step 297800, loss 8.586569, gender accuracy 0.875000, age diff 8.367607\n",
      "step 298000, loss 12.294220, gender accuracy 0.937500, age diff 12.212612\n",
      "step 298200, loss 8.155967, gender accuracy 1.000000, age diff 8.143949\n",
      "step 298400, loss 10.268917, gender accuracy 1.000000, age diff 10.260633\n",
      "step 298600, loss 7.121571, gender accuracy 1.000000, age diff 7.106814\n",
      "step 298800, loss 7.681357, gender accuracy 1.000000, age diff 7.670802\n",
      "step 299000, loss 8.880280, gender accuracy 1.000000, age diff 8.878162\n",
      "step 299200, loss 6.614847, gender accuracy 1.000000, age diff 6.579890\n",
      "step 299400, loss 9.254182, gender accuracy 1.000000, age diff 9.212382\n",
      "step 299600, loss 7.613643, gender accuracy 0.937500, age diff 7.532262\n",
      "step 299800, loss 8.637527, gender accuracy 0.875000, age diff 8.370836\n",
      "step 300000, loss 8.931973, gender accuracy 1.000000, age diff 8.920485\n",
      "step 300200, loss 8.361745, gender accuracy 1.000000, age diff 8.314716\n",
      "step 300400, loss 12.049638, gender accuracy 0.937500, age diff 11.963313\n",
      "step 300600, loss 7.141843, gender accuracy 1.000000, age diff 7.133742\n",
      "step 300800, loss 7.662731, gender accuracy 1.000000, age diff 7.662445\n",
      "step 301000, loss 10.347706, gender accuracy 0.875000, age diff 10.161444\n",
      "step 301200, loss 7.119832, gender accuracy 1.000000, age diff 7.108594\n",
      "step 301400, loss 9.265050, gender accuracy 1.000000, age diff 9.262804\n",
      "step 301600, loss 5.140121, gender accuracy 1.000000, age diff 5.138967\n",
      "step 301800, loss 8.537775, gender accuracy 1.000000, age diff 8.527044\n",
      "step 302000, loss 6.285744, gender accuracy 1.000000, age diff 6.272815\n",
      "step 302200, loss 8.597342, gender accuracy 1.000000, age diff 8.564713\n",
      "step 302400, loss 7.281245, gender accuracy 0.937500, age diff 7.206584\n",
      "step 302600, loss 7.492892, gender accuracy 1.000000, age diff 7.488015\n",
      "step 302800, loss 7.505441, gender accuracy 1.000000, age diff 7.478384\n",
      "step 303000, loss 7.240863, gender accuracy 1.000000, age diff 7.197366\n",
      "step 303200, loss 7.832987, gender accuracy 1.000000, age diff 7.805360\n",
      "step 303400, loss 6.727291, gender accuracy 1.000000, age diff 6.715083\n",
      "step 303600, loss 11.907757, gender accuracy 1.000000, age diff 11.861670\n",
      "step 303800, loss 7.606199, gender accuracy 1.000000, age diff 7.598015\n",
      "step 304000, loss 6.710051, gender accuracy 1.000000, age diff 6.708486\n",
      "step 304200, loss 8.651627, gender accuracy 1.000000, age diff 8.639849\n",
      "step 304400, loss 6.415270, gender accuracy 1.000000, age diff 6.380321\n",
      "step 304600, loss 9.358945, gender accuracy 1.000000, age diff 9.354560\n",
      "step 304800, loss 7.416296, gender accuracy 1.000000, age diff 7.413542\n",
      "step 305000, loss 6.098296, gender accuracy 1.000000, age diff 6.065941\n",
      "step 305200, loss 8.735803, gender accuracy 0.937500, age diff 8.570390\n",
      "step 305400, loss 8.612218, gender accuracy 0.875000, age diff 8.222486\n",
      "step 305600, loss 8.500837, gender accuracy 1.000000, age diff 8.493660\n",
      "step 305800, loss 10.331933, gender accuracy 1.000000, age diff 10.330908\n",
      "step 306000, loss 13.806643, gender accuracy 0.937500, age diff 13.694080\n",
      "step 306200, loss 9.298844, gender accuracy 1.000000, age diff 9.296986\n",
      "step 306400, loss 7.162929, gender accuracy 0.937500, age diff 7.112657\n",
      "step 306600, loss 10.337294, gender accuracy 1.000000, age diff 10.335373\n",
      "step 306800, loss 6.969024, gender accuracy 0.937500, age diff 6.809700\n",
      "step 307000, loss 9.835697, gender accuracy 1.000000, age diff 9.816763\n",
      "step 307200, loss 9.121856, gender accuracy 1.000000, age diff 9.116367\n",
      "step 307400, loss 8.308539, gender accuracy 1.000000, age diff 8.290920\n",
      "step 307600, loss 6.839656, gender accuracy 1.000000, age diff 6.824224\n",
      "step 307800, loss 8.200355, gender accuracy 1.000000, age diff 8.133440\n",
      "step 308000, loss 11.347458, gender accuracy 0.937500, age diff 11.239076\n",
      "step 308200, loss 7.566074, gender accuracy 1.000000, age diff 7.562881\n",
      "step 308400, loss 5.876929, gender accuracy 0.937500, age diff 5.795816\n",
      "step 308600, loss 12.465634, gender accuracy 1.000000, age diff 12.454074\n",
      "step 308800, loss 4.916563, gender accuracy 1.000000, age diff 4.914485\n",
      "step 309000, loss 7.407005, gender accuracy 1.000000, age diff 7.397433\n",
      "step 309200, loss 8.676541, gender accuracy 1.000000, age diff 8.675810\n",
      "step 309400, loss 6.424666, gender accuracy 1.000000, age diff 6.422958\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-9-84a6b1f3b199>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      2\u001b[0m     \u001b[0mstep\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msess\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrun\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mglobal_step\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m     \u001b[0;32mwhile\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mcoord\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshould_stop\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 4\u001b[0;31m         \u001b[0macc_gender\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mage_loss\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0mloss\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0m_\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msess\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrun\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mgender_accuracy\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mabs_loss\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcost\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0moptimizer\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      5\u001b[0m         \u001b[0msummary\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msess\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrun\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmerged\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      6\u001b[0m         \u001b[0mwriter\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd_summary\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msummary\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mglobal_step\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mstep\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python3.5/dist-packages/tensorflow/python/client/session.py\u001b[0m in \u001b[0;36mrun\u001b[0;34m(self, fetches, feed_dict, options, run_metadata)\u001b[0m\n\u001b[1;32m    893\u001b[0m     \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    894\u001b[0m       result = self._run(None, fetches, feed_dict, options_ptr,\n\u001b[0;32m--> 895\u001b[0;31m                          run_metadata_ptr)\n\u001b[0m\u001b[1;32m    896\u001b[0m       \u001b[0;32mif\u001b[0m \u001b[0mrun_metadata\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    897\u001b[0m         \u001b[0mproto_data\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtf_session\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mTF_GetBuffer\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mrun_metadata_ptr\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python3.5/dist-packages/tensorflow/python/client/session.py\u001b[0m in \u001b[0;36m_run\u001b[0;34m(self, handle, fetches, feed_dict, options, run_metadata)\u001b[0m\n\u001b[1;32m   1126\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mfinal_fetches\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0mfinal_targets\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mhandle\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0mfeed_dict_tensor\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1127\u001b[0m       results = self._do_run(handle, final_targets, final_fetches,\n\u001b[0;32m-> 1128\u001b[0;31m                              feed_dict_tensor, options, run_metadata)\n\u001b[0m\u001b[1;32m   1129\u001b[0m     \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1130\u001b[0m       \u001b[0mresults\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python3.5/dist-packages/tensorflow/python/client/session.py\u001b[0m in \u001b[0;36m_do_run\u001b[0;34m(self, handle, target_list, fetch_list, feed_dict, options, run_metadata)\u001b[0m\n\u001b[1;32m   1342\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mhandle\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1343\u001b[0m       return self._do_call(_run_fn, self._session, feeds, fetches, targets,\n\u001b[0;32m-> 1344\u001b[0;31m                            options, run_metadata)\n\u001b[0m\u001b[1;32m   1345\u001b[0m     \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1346\u001b[0m       \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_do_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0m_prun_fn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_session\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mhandle\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfeeds\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfetches\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python3.5/dist-packages/tensorflow/python/client/session.py\u001b[0m in \u001b[0;36m_do_call\u001b[0;34m(self, fn, *args)\u001b[0m\n\u001b[1;32m   1348\u001b[0m   \u001b[0;32mdef\u001b[0m \u001b[0m_do_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1349\u001b[0m     \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1350\u001b[0;31m       \u001b[0;32mreturn\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1351\u001b[0m     \u001b[0;32mexcept\u001b[0m \u001b[0merrors\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mOpError\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1352\u001b[0m       \u001b[0mmessage\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcompat\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mas_text\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0me\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmessage\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python3.5/dist-packages/tensorflow/python/client/session.py\u001b[0m in \u001b[0;36m_run_fn\u001b[0;34m(session, feed_dict, fetch_list, target_list, options, run_metadata)\u001b[0m\n\u001b[1;32m   1327\u001b[0m           return tf_session.TF_Run(session, options,\n\u001b[1;32m   1328\u001b[0m                                    \u001b[0mfeed_dict\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfetch_list\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtarget_list\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1329\u001b[0;31m                                    status, run_metadata)\n\u001b[0m\u001b[1;32m   1330\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1331\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0m_prun_fn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msession\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mhandle\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfeed_dict\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfetch_list\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "try:\n",
    "    step = sess.run(global_step)\n",
    "    while not coord.should_stop():\n",
    "        acc_gender, age_loss,loss,_ = sess.run([gender_accuracy, abs_loss, cost, optimizer])\n",
    "        summary = sess.run(merged)\n",
    "        writer.add_summary(summary, global_step=step)\n",
    "        if step % 200 == 0:\n",
    "            print('step %d, loss %f, gender accuracy %f, age diff %f'%(step,loss,acc_gender, age_loss))\n",
    "            saver.save(sess, \"checkpoints/combine-checkpoint-mobilenet.ckpt\")\n",
    "        step = sess.run(global_step)\n",
    "except tf.errors.OutOfRangeError:\n",
    "    print('Done training for %d epochs, %d steps.' % (epoch, step))\n",
    "finally:\n",
    "    coord.request_stop()\n",
    "coord.join(threads)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}