polyaxon_lstm_sin_cos.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import tensorflow as tf\n",
    "from matplotlib import pyplot as plt\n",
    "\n",
    "from tensorflow.contrib import learn\n",
    "from sklearn.metrics import mean_squared_error\n",
    "\n",
    "from data_processing import generate_data, sin_cos\n",
    "from polyaxon_experiement import create_experiment\n",
    "from polyaxon.processing.input_fn import numpy_input_fn\n",
    "\n",
    "tf.logging.set_verbosity(tf.logging.INFO)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Using config: {'_task_id': 0, '_cluster_spec': <tensorflow.python.training.server_lib.ClusterSpec object at 0x114683dd8>, '_keep_checkpoint_every_n_hours': 4, '_save_summary_steps': 100, '_session_config': None, '_save_checkpoints_steps': 100, '_master': '', '_tf_random_seed': 1, '_num_ps_replicas': 0, '_keep_checkpoint_max': 5, '_save_checkpoints_secs': None, '_model_dir': None, '_num_worker_replicas': 0, '_is_chief': True, '_evaluation_master': '', '_environment': 'local', '_tf_config': gpu_options {\n",
      "  per_process_gpu_memory_fraction: 1.0\n",
      "}\n",
      ", '_task_type': None}\n"
     ]
    }
   ],
   "source": [
    "TIMESTEPS = 7\n",
    "X, y = generate_data(sin_cos, np.linspace(0, 100, 10000, dtype=np.float32), TIMESTEPS, seperate=False)\n",
    "xp = create_experiment('/tmp/polyaxon_logs/sin_cos', X, y, output_units=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Training model for 100 steps\n",
      "WARNING:tensorflow:enqueue_data was called with shuffle=False and num_threads > 1. This will create multiple threads, all reading the array/dataframe in order. If you want examples read in order, use one thread; if you want multiple threads, enable shuffling.\n",
      "INFO:tensorflow:Create CheckpointSaverHook.\n",
      "INFO:tensorflow:Saving checkpoints for 1 into /tmp/polyaxon_logs/sin_cos/model.ckpt.\n",
      "INFO:tensorflow:loss = 0.498125, step = 1\n",
      "INFO:tensorflow:Saving checkpoints for 100 into /tmp/polyaxon_logs/sin_cos/model.ckpt.\n",
      "INFO:tensorflow:Loss for final step: 0.00243695.\n",
      "INFO:tensorflow:Evaluating model now.\n",
      "WARNING:tensorflow:enqueue_data was called with shuffle=False and num_threads > 1. This will create multiple threads, all reading the array/dataframe in order. If you want examples read in order, use one thread; if you want multiple threads, enable shuffling.\n",
      "INFO:tensorflow:Starting evaluation at 2017-06-15-16:06:40\n",
      "INFO:tensorflow:Restoring parameters from /tmp/polyaxon_logs/sin_cos/model.ckpt-100\n",
      "INFO:tensorflow:Evaluation [1/10]\n",
      "INFO:tensorflow:Evaluation [2/10]\n",
      "INFO:tensorflow:Evaluation [3/10]\n",
      "INFO:tensorflow:Evaluation [4/10]\n",
      "INFO:tensorflow:Evaluation [5/10]\n",
      "INFO:tensorflow:Evaluation [6/10]\n",
      "INFO:tensorflow:Evaluation [7/10]\n",
      "INFO:tensorflow:Evaluation [8/10]\n",
      "INFO:tensorflow:Evaluation [9/10]\n",
      "INFO:tensorflow:Evaluation [10/10]\n",
      "INFO:tensorflow:Finished evaluation at 2017-06-15-16:06:40\n",
      "INFO:tensorflow:Saving dict for global step 100: global_step = 100, loss = 0.00041879, streaming_mean_absolute_error = 0.0182897, streaming_root_mean_squared_error = 0.0204644\n",
      "INFO:tensorflow:Training model for 100 steps\n",
      "WARNING:tensorflow:enqueue_data was called with shuffle=False and num_threads > 1. This will create multiple threads, all reading the array/dataframe in order. If you want examples read in order, use one thread; if you want multiple threads, enable shuffling.\n",
      "INFO:tensorflow:Create CheckpointSaverHook.\n",
      "INFO:tensorflow:Restoring parameters from /tmp/polyaxon_logs/sin_cos/model.ckpt-100\n",
      "INFO:tensorflow:Saving checkpoints for 101 into /tmp/polyaxon_logs/sin_cos/model.ckpt.\n",
      "INFO:tensorflow:loss = 0.000188615, step = 101\n",
      "INFO:tensorflow:Saving checkpoints for 200 into /tmp/polyaxon_logs/sin_cos/model.ckpt.\n",
      "INFO:tensorflow:Loss for final step: 0.000223352.\n",
      "INFO:tensorflow:Evaluating model now.\n",
      "WARNING:tensorflow:enqueue_data was called with shuffle=False and num_threads > 1. This will create multiple threads, all reading the array/dataframe in order. If you want examples read in order, use one thread; if you want multiple threads, enable shuffling.\n",
      "INFO:tensorflow:Starting evaluation at 2017-06-15-16:06:52\n",
      "INFO:tensorflow:Restoring parameters from /tmp/polyaxon_logs/sin_cos/model.ckpt-200\n",
      "INFO:tensorflow:Evaluation [1/10]\n",
      "INFO:tensorflow:Evaluation [2/10]\n",
      "INFO:tensorflow:Evaluation [3/10]\n",
      "INFO:tensorflow:Evaluation [4/10]\n",
      "INFO:tensorflow:Evaluation [5/10]\n",
      "INFO:tensorflow:Evaluation [6/10]\n",
      "INFO:tensorflow:Evaluation [7/10]\n",
      "INFO:tensorflow:Evaluation [8/10]\n",
      "INFO:tensorflow:Evaluation [9/10]\n",
      "INFO:tensorflow:Evaluation [10/10]\n",
      "INFO:tensorflow:Finished evaluation at 2017-06-15-16:06:53\n",
      "INFO:tensorflow:Saving dict for global step 200: global_step = 200, loss = 0.000239881, streaming_mean_absolute_error = 0.0136209, streaming_root_mean_squared_error = 0.0154881\n",
      "INFO:tensorflow:Training model for 100 steps\n",
      "WARNING:tensorflow:enqueue_data was called with shuffle=False and num_threads > 1. This will create multiple threads, all reading the array/dataframe in order. If you want examples read in order, use one thread; if you want multiple threads, enable shuffling.\n",
      "INFO:tensorflow:Create CheckpointSaverHook.\n",
      "INFO:tensorflow:Restoring parameters from /tmp/polyaxon_logs/sin_cos/model.ckpt-200\n",
      "INFO:tensorflow:Saving checkpoints for 201 into /tmp/polyaxon_logs/sin_cos/model.ckpt.\n",
      "INFO:tensorflow:loss = 6.16938e-05, step = 201\n",
      "INFO:tensorflow:Saving checkpoints for 300 into /tmp/polyaxon_logs/sin_cos/model.ckpt.\n",
      "INFO:tensorflow:Loss for final step: 0.000174256.\n",
      "INFO:tensorflow:Evaluating model now.\n",
      "WARNING:tensorflow:enqueue_data was called with shuffle=False and num_threads > 1. This will create multiple threads, all reading the array/dataframe in order. If you want examples read in order, use one thread; if you want multiple threads, enable shuffling.\n",
      "INFO:tensorflow:Starting evaluation at 2017-06-15-16:07:04\n",
      "INFO:tensorflow:Restoring parameters from /tmp/polyaxon_logs/sin_cos/model.ckpt-300\n",
      "INFO:tensorflow:Evaluation [1/10]\n",
      "INFO:tensorflow:Evaluation [2/10]\n",
      "INFO:tensorflow:Evaluation [3/10]\n",
      "INFO:tensorflow:Evaluation [4/10]\n",
      "INFO:tensorflow:Evaluation [5/10]\n",
      "INFO:tensorflow:Evaluation [6/10]\n",
      "INFO:tensorflow:Evaluation [7/10]\n",
      "INFO:tensorflow:Evaluation [8/10]\n",
      "INFO:tensorflow:Evaluation [9/10]\n",
      "INFO:tensorflow:Evaluation [10/10]\n",
      "INFO:tensorflow:Finished evaluation at 2017-06-15-16:07:05\n",
      "INFO:tensorflow:Saving dict for global step 300: global_step = 300, loss = 0.000104413, streaming_mean_absolute_error = 0.00873271, streaming_root_mean_squared_error = 0.0102182\n",
      "INFO:tensorflow:Training model for 100 steps\n",
      "WARNING:tensorflow:enqueue_data was called with shuffle=False and num_threads > 1. This will create multiple threads, all reading the array/dataframe in order. If you want examples read in order, use one thread; if you want multiple threads, enable shuffling.\n",
      "INFO:tensorflow:Create CheckpointSaverHook.\n",
      "INFO:tensorflow:Restoring parameters from /tmp/polyaxon_logs/sin_cos/model.ckpt-300\n",
      "INFO:tensorflow:Saving checkpoints for 301 into /tmp/polyaxon_logs/sin_cos/model.ckpt.\n",
      "INFO:tensorflow:loss = 3.99327e-05, step = 301\n",
      "INFO:tensorflow:Saving checkpoints for 400 into /tmp/polyaxon_logs/sin_cos/model.ckpt.\n",
      "INFO:tensorflow:Loss for final step: 7.79159e-05.\n",
      "INFO:tensorflow:Evaluating model now.\n",
      "WARNING:tensorflow:enqueue_data was called with shuffle=False and num_threads > 1. This will create multiple threads, all reading the array/dataframe in order. If you want examples read in order, use one thread; if you want multiple threads, enable shuffling.\n",
      "INFO:tensorflow:Starting evaluation at 2017-06-15-16:07:17\n",
      "INFO:tensorflow:Restoring parameters from /tmp/polyaxon_logs/sin_cos/model.ckpt-400\n",
      "INFO:tensorflow:Evaluation [1/10]\n",
      "INFO:tensorflow:Evaluation [2/10]\n",
      "INFO:tensorflow:Evaluation [3/10]\n",
      "INFO:tensorflow:Evaluation [4/10]\n",
      "INFO:tensorflow:Evaluation [5/10]\n",
      "INFO:tensorflow:Evaluation [6/10]\n",
      "INFO:tensorflow:Evaluation [7/10]\n",
      "INFO:tensorflow:Evaluation [8/10]\n",
      "INFO:tensorflow:Evaluation [9/10]\n",
      "INFO:tensorflow:Evaluation [10/10]\n",
      "INFO:tensorflow:Finished evaluation at 2017-06-15-16:07:18\n",
      "INFO:tensorflow:Saving dict for global step 400: global_step = 400, loss = 7.78833e-05, streaming_mean_absolute_error = 0.00762699, streaming_root_mean_squared_error = 0.00882515\n",
      "INFO:tensorflow:Training model for 100 steps\n",
      "WARNING:tensorflow:enqueue_data was called with shuffle=False and num_threads > 1. This will create multiple threads, all reading the array/dataframe in order. If you want examples read in order, use one thread; if you want multiple threads, enable shuffling.\n",
      "INFO:tensorflow:Create CheckpointSaverHook.\n",
      "INFO:tensorflow:Restoring parameters from /tmp/polyaxon_logs/sin_cos/model.ckpt-400\n",
      "INFO:tensorflow:Saving checkpoints for 401 into /tmp/polyaxon_logs/sin_cos/model.ckpt.\n",
      "INFO:tensorflow:loss = 3.27785e-05, step = 401\n",
      "INFO:tensorflow:Saving checkpoints for 500 into /tmp/polyaxon_logs/sin_cos/model.ckpt.\n",
      "INFO:tensorflow:Loss for final step: 0.000145839.\n",
      "INFO:tensorflow:Evaluating model now.\n",
      "WARNING:tensorflow:enqueue_data was called with shuffle=False and num_threads > 1. This will create multiple threads, all reading the array/dataframe in order. If you want examples read in order, use one thread; if you want multiple threads, enable shuffling.\n",
      "INFO:tensorflow:Starting evaluation at 2017-06-15-16:07:29\n",
      "INFO:tensorflow:Restoring parameters from /tmp/polyaxon_logs/sin_cos/model.ckpt-500\n",
      "INFO:tensorflow:Evaluation [1/10]\n",
      "INFO:tensorflow:Evaluation [2/10]\n",
      "INFO:tensorflow:Evaluation [3/10]\n",
      "INFO:tensorflow:Evaluation [4/10]\n",
      "INFO:tensorflow:Evaluation [5/10]\n",
      "INFO:tensorflow:Evaluation [6/10]\n",
      "INFO:tensorflow:Evaluation [7/10]\n",
      "INFO:tensorflow:Evaluation [8/10]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Evaluation [9/10]\n",
      "INFO:tensorflow:Evaluation [10/10]\n",
      "INFO:tensorflow:Finished evaluation at 2017-06-15-16:07:30\n",
      "INFO:tensorflow:Saving dict for global step 500: global_step = 500, loss = 5.35004e-05, streaming_mean_absolute_error = 0.00611612, streaming_root_mean_squared_error = 0.0073144\n",
      "INFO:tensorflow:Training model for 100 steps\n",
      "WARNING:tensorflow:enqueue_data was called with shuffle=False and num_threads > 1. This will create multiple threads, all reading the array/dataframe in order. If you want examples read in order, use one thread; if you want multiple threads, enable shuffling.\n",
      "INFO:tensorflow:Create CheckpointSaverHook.\n",
      "INFO:tensorflow:Restoring parameters from /tmp/polyaxon_logs/sin_cos/model.ckpt-500\n",
      "INFO:tensorflow:Saving checkpoints for 501 into /tmp/polyaxon_logs/sin_cos/model.ckpt.\n",
      "INFO:tensorflow:loss = 1.14016e-05, step = 501\n",
      "INFO:tensorflow:Saving checkpoints for 600 into /tmp/polyaxon_logs/sin_cos/model.ckpt.\n",
      "INFO:tensorflow:Loss for final step: 8.65954e-05.\n",
      "INFO:tensorflow:Evaluating model now.\n",
      "WARNING:tensorflow:enqueue_data was called with shuffle=False and num_threads > 1. This will create multiple threads, all reading the array/dataframe in order. If you want examples read in order, use one thread; if you want multiple threads, enable shuffling.\n",
      "INFO:tensorflow:Starting evaluation at 2017-06-15-16:07:42\n",
      "INFO:tensorflow:Restoring parameters from /tmp/polyaxon_logs/sin_cos/model.ckpt-600\n",
      "INFO:tensorflow:Evaluation [1/10]\n",
      "INFO:tensorflow:Evaluation [2/10]\n",
      "INFO:tensorflow:Evaluation [3/10]\n",
      "INFO:tensorflow:Evaluation [4/10]\n",
      "INFO:tensorflow:Evaluation [5/10]\n",
      "INFO:tensorflow:Evaluation [6/10]\n",
      "INFO:tensorflow:Evaluation [7/10]\n",
      "INFO:tensorflow:Evaluation [8/10]\n",
      "INFO:tensorflow:Evaluation [9/10]\n",
      "INFO:tensorflow:Evaluation [10/10]\n",
      "INFO:tensorflow:Finished evaluation at 2017-06-15-16:07:43\n",
      "INFO:tensorflow:Saving dict for global step 600: global_step = 600, loss = 4.85077e-05, streaming_mean_absolute_error = 0.00603464, streaming_root_mean_squared_error = 0.00696475\n",
      "INFO:tensorflow:Training model for 100 steps\n",
      "WARNING:tensorflow:enqueue_data was called with shuffle=False and num_threads > 1. This will create multiple threads, all reading the array/dataframe in order. If you want examples read in order, use one thread; if you want multiple threads, enable shuffling.\n",
      "INFO:tensorflow:Create CheckpointSaverHook.\n",
      "INFO:tensorflow:Restoring parameters from /tmp/polyaxon_logs/sin_cos/model.ckpt-600\n",
      "INFO:tensorflow:Saving checkpoints for 601 into /tmp/polyaxon_logs/sin_cos/model.ckpt.\n",
      "INFO:tensorflow:loss = 1.74002e-05, step = 601\n",
      "INFO:tensorflow:Saving checkpoints for 700 into /tmp/polyaxon_logs/sin_cos/model.ckpt.\n",
      "INFO:tensorflow:Loss for final step: 7.57762e-05.\n",
      "INFO:tensorflow:Evaluating model now.\n",
      "WARNING:tensorflow:enqueue_data was called with shuffle=False and num_threads > 1. This will create multiple threads, all reading the array/dataframe in order. If you want examples read in order, use one thread; if you want multiple threads, enable shuffling.\n",
      "INFO:tensorflow:Starting evaluation at 2017-06-15-16:07:55\n",
      "INFO:tensorflow:Restoring parameters from /tmp/polyaxon_logs/sin_cos/model.ckpt-700\n",
      "INFO:tensorflow:Evaluation [1/10]\n",
      "INFO:tensorflow:Evaluation [2/10]\n",
      "INFO:tensorflow:Evaluation [3/10]\n",
      "INFO:tensorflow:Evaluation [4/10]\n",
      "INFO:tensorflow:Evaluation [5/10]\n",
      "INFO:tensorflow:Evaluation [6/10]\n",
      "INFO:tensorflow:Evaluation [7/10]\n",
      "INFO:tensorflow:Evaluation [8/10]\n",
      "INFO:tensorflow:Evaluation [9/10]\n",
      "INFO:tensorflow:Evaluation [10/10]\n",
      "INFO:tensorflow:Finished evaluation at 2017-06-15-16:07:55\n",
      "INFO:tensorflow:Saving dict for global step 700: global_step = 700, loss = 4.27691e-05, streaming_mean_absolute_error = 0.0056428, streaming_root_mean_squared_error = 0.00653981\n",
      "INFO:tensorflow:Training model for 100 steps\n",
      "WARNING:tensorflow:enqueue_data was called with shuffle=False and num_threads > 1. This will create multiple threads, all reading the array/dataframe in order. If you want examples read in order, use one thread; if you want multiple threads, enable shuffling.\n",
      "INFO:tensorflow:Create CheckpointSaverHook.\n",
      "INFO:tensorflow:Restoring parameters from /tmp/polyaxon_logs/sin_cos/model.ckpt-700\n",
      "INFO:tensorflow:Saving checkpoints for 701 into /tmp/polyaxon_logs/sin_cos/model.ckpt.\n",
      "INFO:tensorflow:loss = 1.67958e-05, step = 701\n",
      "INFO:tensorflow:Saving checkpoints for 800 into /tmp/polyaxon_logs/sin_cos/model.ckpt.\n",
      "INFO:tensorflow:Loss for final step: 5.95159e-05.\n",
      "INFO:tensorflow:Evaluating model now.\n",
      "WARNING:tensorflow:enqueue_data was called with shuffle=False and num_threads > 1. This will create multiple threads, all reading the array/dataframe in order. If you want examples read in order, use one thread; if you want multiple threads, enable shuffling.\n",
      "INFO:tensorflow:Starting evaluation at 2017-06-15-16:08:06\n",
      "INFO:tensorflow:Restoring parameters from /tmp/polyaxon_logs/sin_cos/model.ckpt-800\n",
      "INFO:tensorflow:Evaluation [1/10]\n",
      "INFO:tensorflow:Evaluation [2/10]\n",
      "INFO:tensorflow:Evaluation [3/10]\n",
      "INFO:tensorflow:Evaluation [4/10]\n",
      "INFO:tensorflow:Evaluation [5/10]\n",
      "INFO:tensorflow:Evaluation [6/10]\n",
      "INFO:tensorflow:Evaluation [7/10]\n",
      "INFO:tensorflow:Evaluation [8/10]\n",
      "INFO:tensorflow:Evaluation [9/10]\n",
      "INFO:tensorflow:Evaluation [10/10]\n",
      "INFO:tensorflow:Finished evaluation at 2017-06-15-16:08:07\n",
      "INFO:tensorflow:Saving dict for global step 800: global_step = 800, loss = 3.63065e-05, streaming_mean_absolute_error = 0.0051756, streaming_root_mean_squared_error = 0.00602549\n",
      "INFO:tensorflow:Training model for 100 steps\n",
      "WARNING:tensorflow:enqueue_data was called with shuffle=False and num_threads > 1. This will create multiple threads, all reading the array/dataframe in order. If you want examples read in order, use one thread; if you want multiple threads, enable shuffling.\n",
      "INFO:tensorflow:Create CheckpointSaverHook.\n",
      "INFO:tensorflow:Restoring parameters from /tmp/polyaxon_logs/sin_cos/model.ckpt-800\n",
      "INFO:tensorflow:Saving checkpoints for 801 into /tmp/polyaxon_logs/sin_cos/model.ckpt.\n",
      "INFO:tensorflow:loss = 1.31804e-05, step = 801\n",
      "INFO:tensorflow:Saving checkpoints for 900 into /tmp/polyaxon_logs/sin_cos/model.ckpt.\n",
      "INFO:tensorflow:Loss for final step: 5.84989e-05.\n",
      "INFO:tensorflow:Evaluating model now.\n",
      "WARNING:tensorflow:enqueue_data was called with shuffle=False and num_threads > 1. This will create multiple threads, all reading the array/dataframe in order. If you want examples read in order, use one thread; if you want multiple threads, enable shuffling.\n",
      "INFO:tensorflow:Starting evaluation at 2017-06-15-16:08:17\n",
      "INFO:tensorflow:Restoring parameters from /tmp/polyaxon_logs/sin_cos/model.ckpt-900\n",
      "INFO:tensorflow:Evaluation [1/10]\n",
      "INFO:tensorflow:Evaluation [2/10]\n",
      "INFO:tensorflow:Evaluation [3/10]\n",
      "INFO:tensorflow:Evaluation [4/10]\n",
      "INFO:tensorflow:Evaluation [5/10]\n",
      "INFO:tensorflow:Evaluation [6/10]\n",
      "INFO:tensorflow:Evaluation [7/10]\n",
      "INFO:tensorflow:Evaluation [8/10]\n",
      "INFO:tensorflow:Evaluation [9/10]\n",
      "INFO:tensorflow:Evaluation [10/10]\n",
      "INFO:tensorflow:Finished evaluation at 2017-06-15-16:08:18\n",
      "INFO:tensorflow:Saving dict for global step 900: global_step = 900, loss = 3.41514e-05, streaming_mean_absolute_error = 0.00499401, streaming_root_mean_squared_error = 0.00584392\n",
      "INFO:tensorflow:Training model for 100 steps\n",
      "WARNING:tensorflow:enqueue_data was called with shuffle=False and num_threads > 1. This will create multiple threads, all reading the array/dataframe in order. If you want examples read in order, use one thread; if you want multiple threads, enable shuffling.\n",
      "INFO:tensorflow:Create CheckpointSaverHook.\n",
      "INFO:tensorflow:Restoring parameters from /tmp/polyaxon_logs/sin_cos/model.ckpt-900\n",
      "INFO:tensorflow:Saving checkpoints for 901 into /tmp/polyaxon_logs/sin_cos/model.ckpt.\n",
      "INFO:tensorflow:loss = 1.38461e-05, step = 901\n",
      "INFO:tensorflow:Saving checkpoints for 1000 into /tmp/polyaxon_logs/sin_cos/model.ckpt.\n",
      "INFO:tensorflow:Loss for final step: 1.79018e-05.\n",
      "INFO:tensorflow:Evaluating model now.\n",
      "WARNING:tensorflow:enqueue_data was called with shuffle=False and num_threads > 1. This will create multiple threads, all reading the array/dataframe in order. If you want examples read in order, use one thread; if you want multiple threads, enable shuffling.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Starting evaluation at 2017-06-15-16:08:27\n",
      "INFO:tensorflow:Restoring parameters from /tmp/polyaxon_logs/sin_cos/model.ckpt-1000\n",
      "INFO:tensorflow:Evaluation [1/10]\n",
      "INFO:tensorflow:Evaluation [2/10]\n",
      "INFO:tensorflow:Evaluation [3/10]\n",
      "INFO:tensorflow:Evaluation [4/10]\n",
      "INFO:tensorflow:Evaluation [5/10]\n",
      "INFO:tensorflow:Evaluation [6/10]\n",
      "INFO:tensorflow:Evaluation [7/10]\n",
      "INFO:tensorflow:Evaluation [8/10]\n",
      "INFO:tensorflow:Evaluation [9/10]\n",
      "INFO:tensorflow:Evaluation [10/10]\n",
      "INFO:tensorflow:Finished evaluation at 2017-06-15-16:08:28\n",
      "INFO:tensorflow:Saving dict for global step 1000: global_step = 1000, loss = 3.21573e-05, streaming_mean_absolute_error = 0.00481721, streaming_root_mean_squared_error = 0.00567074\n",
      "INFO:tensorflow:Stop training model as max steps reached\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "({'global_step': 1000,\n",
       "  'loss': 3.2157306e-05,\n",
       "  'streaming_mean_absolute_error': 0.0048172129,\n",
       "  'streaming_root_mean_squared_error': 0.0056707412},\n",
       " [])"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xp.continuous_train_and_evaluate()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Restoring parameters from /tmp/polyaxon_logs/sin_cos/model.ckpt-1000\n",
      "MSE: 0.000050\n"
     ]
    }
   ],
   "source": [
    "predicted = [i['results'] for i in xp.estimator.predict(numpy_input_fn({'x': X['test']}, shuffle=False))]\n",
    "rmse = np.sqrt(((predicted - y['test']) ** 2).mean(axis=0))\n",
    "score = mean_squared_error(predicted, y['test'])\n",
    "print (\"MSE: %f\" % score)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x1160e8a58>,\n",
       " <matplotlib.lines.Line2D at 0x1160e8c50>]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYYAAAEACAYAAAC3adEgAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x115f67f60>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(predicted)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x115a65278>,\n",
       " <matplotlib.lines.Line2D at 0x115a65128>]"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYYAAAEACAYAAAC3adEgAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1160e65c0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(y['test'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x103fed208>"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYYAAAEACAYAAAC3adEgAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x103fe70f0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_predicted = plt.plot(predicted, label='predicted')\n",
    "plot_test = plt.plot(y['test'], label='test')\n",
    "plt.legend(handles=plot_predicted+plot_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    ""
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3.0
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}