Fixed ValueError: Attempt to have a second RNNCell use the weights of a variable scope that already has weights: 'rnn/basic_lstm_cell'

Author: energyfirefox

ch10_rnn/Concept02_rnn.ipynb

@@ -2,43 +2,39 @@
  "cells": [
   {
    "cell_type": "markdown",
-   "metadata": {
-    "deletable": true,
-    "editable": true
-   },
+   "metadata": {},
    "source": [
     "# Ch `10`: Concept `02`"
    ]
   },
   {
    "cell_type": "markdown",
-   "metadata": {
-    "deletable": true,
-    "editable": true
-   },
+   "metadata": {},
    "source": [
     "## Recurrent Neural Network"
    ]
   },
   {
    "cell_type": "markdown",
-   "metadata": {
-    "deletable": true,
-    "editable": true
-   },
+   "metadata": {},
    "source": [
     "Import the relevant libraries:"
    ]
   },
   {
    "cell_type": "code",
    "execution_count": 1,
-   "metadata": {
-    "collapsed": false,
-    "deletable": true,
-    "editable": true
-   },
-   "outputs": [],
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/Users/anastasiia/anaconda3/lib/python3.6/site-packages/h5py/__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n",
+      "  from ._conv import register_converters as _register_converters\n"
+     ]
+    }
+   ],
    "source": [
     "import numpy as np\n",
     "import tensorflow as tf\n",
@@ -47,22 +43,15 @@
   },
   {
    "cell_type": "markdown",
-   "metadata": {
-    "deletable": true,
-    "editable": true
-   },
+   "metadata": {},
    "source": [
     "Define the RNN model:"
    ]
   },
   {
    "cell_type": "code",
    "execution_count": 2,
-   "metadata": {
-    "collapsed": true,
-    "deletable": true,
-    "editable": true
-   },
+   "metadata": {},
    "outputs": [],
    "source": [
     "class SeriesPredictor:\n",
@@ -92,7 +81,7 @@
     "        :param W: matrix of fully-connected output layer weights\n",
     "        :param b: vector of fully-connected output layer biases\n",
     "        \"\"\"\n",
-    "        cell = rnn.BasicLSTMCell(self.hidden_dim)\n",
+    "        cell = rnn.BasicLSTMCell(self.hidden_dim, reuse=tf.get_variable_scope().reuse)\n",
     "        outputs, states = tf.nn.dynamic_rnn(cell, self.x, dtype=tf.float32)\n",
     "        num_examples = tf.shape(self.x)[0]\n",
     "        W_repeated = tf.tile(tf.expand_dims(self.W_out, 0), [num_examples, 1, 1])\n",
@@ -123,48 +112,42 @@
   },
   {
    "cell_type": "markdown",
-   "metadata": {
-    "deletable": true,
-    "editable": true
-   },
+   "metadata": {},
    "source": [
     "Now, we'll train a series predictor. Let's say we have a sequence of numbers `[a, b, c, d]` that we want to transform into `[a, a+b, b+c, c+d]`. We'll give the RNN a couple examples in the training data. Let's see how well it learns this intended transformation:"
    ]
   },
   {
    "cell_type": "code",
    "execution_count": 3,
-   "metadata": {
-    "collapsed": false,
-    "deletable": true,
-    "editable": true
-   },
+   "metadata": {},
    "outputs": [
     {
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "0 92.1852\n",
-      "100 61.1175\n",
-      "200 27.0341\n",
-      "300 13.9523\n",
-      "400 9.39037\n",
-      "500 7.08643\n",
-      "600 5.50997\n",
-      "700 4.12571\n",
-      "800 3.12016\n",
-      "900 2.42311\n",
+      "0 96.78678\n",
+      "100 61.329662\n",
+      "200 18.419907\n",
+      "300 7.646343\n",
+      "400 4.7979555\n",
+      "500 3.2019987\n",
+      "600 2.2661102\n",
+      "700 1.6707231\n",
+      "800 1.2424115\n",
+      "900 0.9125628\n",
       "Model saved to model.ckpt\n",
+      "INFO:tensorflow:Restoring parameters from ./model.ckpt\n",
       "\n",
       "Lets run some tests!\n",
       "\n",
       "When the input is [[1], [2], [3], [4]]\n",
       "The ground truth output should be [[1], [3], [5], [7]]\n",
-      "And the model thinks it is [ 0.96018004  2.76944828  5.35826826  7.3706851 ]\n",
+      "And the model thinks it is [1.037468 2.519481 4.514736 6.729595]\n",
       "\n",
       "When the input is [[4], [5], [6], [7]]\n",
       "The ground truth output should be [[4], [9], [11], [13]]\n",
-      "And the model thinks it is [  4.17302942   9.161376    11.13204765  11.64120388]\n",
+      "And the model thinks it is [ 4.5689063  9.189994  11.679442  12.760409 ]\n",
       "\n"
      ]
     }
@@ -211,7 +194,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.5.2"
+   "version": "3.6.5"
   }
  },
  "nbformat": 4,