PyTorch RNN.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import numpy as np\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "import torch\n",
    "from torch.autograd import Variable\n",
    "import torch.nn as nn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(200,)"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = np.arange(-10,10,0.1)\n",
    "X.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "## Code taken from http://www.rigtorp.se/2011/01/01/rolling-statistics-numpy.html\n",
    "\n",
    "def rolling_window(a, window):\n",
    "    shape = a.shape[:-1] + (a.shape[-1] - window + 1, window)\n",
    "    strides = a.strides + (a.strides[-1],)\n",
    "    return np.lib.stride_tricks.as_strided(a, shape=shape, strides=strides)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[-10. ,  -9.9],\n",
       "       [ -9.9,  -9.8],\n",
       "       [ -9.8,  -9.7],\n",
       "       [ -9.7,  -9.6]])"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rolling_window(X[:5], 2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(196,)\n"
     ]
    }
   ],
   "source": [
    "X = np.arange(-10,10,0.1)\n",
    "X = np.cos(np.mean(rolling_window(X, 5), -1))\n",
    "#X = X[:-5+1]\n",
    "print(X.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7fd70cf1c1d0>]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYYAAAD8CAYAAABzTgP2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nGv5/P1/AGd0SKqUAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd70efac748>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(X)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Implementation from Pytorch tutorial"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "class RNN(nn.Module):\n",
    "    def __init__(self, input_size, hidden_size, output_size):\n",
    "        super(RNN, self).__init__()\n",
    "        self.hidden_size = hidden_size\n",
    "        self.i2h = nn.Linear(input_size+hidden_size, hidden_size)\n",
    "        self.h2o = nn.Linear(input_size+hidden_size, output_size)\n",
    "        self.tanh = nn.Tanh()\n",
    "        \n",
    "    def forward(self, x, h):\n",
    "        inp = torch.cat((x,h), 1)\n",
    "        hidden = self.tanh(self.i2h(inp))\n",
    "        output = self.h2o(inp)\n",
    "        return hidden, output\n",
    "    \n",
    "    \n",
    "    def get_output(self, X):\n",
    "        time_steps = X.size(0)\n",
    "        batch_size = X.size(1)\n",
    "        hidden = Variable(torch.zeros(batch_size, self.hidden_size))\n",
    "        if torch.cuda.is_available() and X.is_cuda:\n",
    "            hidden = hidden.cuda()\n",
    "        outputs = []\n",
    "        hiddens = []\n",
    "        for t in range(time_steps):\n",
    "            hidden, output = self.forward(X[t], hidden)\n",
    "            outputs.append(output)\n",
    "            hiddens.append(hidden)\n",
    "        return torch.cat(hiddens, 1), torch.cat(outputs, 1)\n",
    "    \n",
    "## Helper functions\n",
    "\n",
    "def get_variable_from_np(X):\n",
    "    return Variable(torch.from_numpy(X)).float()\n",
    "\n",
    "\n",
    "def get_training_data(X, batch_size=10, look_ahead=1):\n",
    "    ## Lookahead will always be one as the prediction is for 1 step ahead\n",
    "    inputs = []\n",
    "    targets = []\n",
    "    time_steps = X.shape[0]\n",
    "    for i in range(0, time_steps-batch_size-look_ahead):\n",
    "        inp = X[i:i+batch_size, np.newaxis, np.newaxis]\n",
    "        inputs.append(get_variable_from_np(inp))\n",
    "        target = X[i+look_ahead:i+batch_size+look_ahead, np.newaxis, np.newaxis]\n",
    "        targets.append(get_variable_from_np(target))\n",
    "        #print(inp.shape, target.shape)\n",
    "    return torch.cat(inputs, 1), torch.cat(targets, 1)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([5, 190, 1])\n",
      "torch.Size([5, 190, 1])\n"
     ]
    }
   ],
   "source": [
    "print(torch.cat([get_variable_from_np(X[i:i+5, np.newaxis, np.newaxis]) for i in range(X.shape[0]-5-1)], 1).size())\n",
    "print(torch.cat([get_variable_from_np(X[i:i+5, np.newaxis, np.newaxis]) for i in range(1, X.shape[0]-5)], 1).size())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(torch.Size([5, 190, 1]), torch.Size([5, 190, 1]))"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "inputs, targets = get_training_data(X, batch_size=5)\n",
    "inputs.size(), targets.size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(torch.Size([3, 192, 1]), torch.Size([3, 192, 1]))"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "inputs, targets = get_training_data(X, batch_size=3)\n",
    "inputs.size(), targets.size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "rnn = RNN(30, 20, 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "criterion = nn.MSELoss()\n",
    "batch_size = 10\n",
    "TIMESTEPS = 5\n",
    "\n",
    "batch = Variable(torch.randn(batch_size, 30))\n",
    "hidden = Variable(torch.randn(batch_size, 20))\n",
    "target = Variable(torch.randn(batch_size, 1))\n",
    "\n",
    "loss = 0\n",
    "for t in range(TIMESTEPS):\n",
    "    hidden, output = rnn(batch, hidden)\n",
    "    loss += criterion(output, target)\n",
    "    \n",
    "loss.backward()    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "rnn = RNN(30, 20, 1).cuda()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "criterion = nn.MSELoss()\n",
    "batch_size = 10\n",
    "TIMESTEPS = 5\n",
    "\n",
    "batch = Variable(torch.randn(batch_size, 30)).cuda()\n",
    "hidden = Variable(torch.randn(batch_size, 20)).cuda()\n",
    "target = Variable(torch.randn(batch_size, 1)).cuda()\n",
    "\n",
    "loss = 0\n",
    "for t in range(TIMESTEPS):\n",
    "    hidden, output = rnn(batch, hidden)\n",
    "    loss += criterion(output, target)\n",
    "    \n",
    "loss.backward()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Run on own data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "rnn = RNN(1,3,1).cuda()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "criterion = nn.MSELoss()\n",
    "#optimizer = torch.optim.SGD(rnn.parameters(), lr=0.001, momentum=0.9)\n",
    "optimizer = torch.optim.Adadelta(rnn.parameters())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(196, 1, 1)"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X[:, np.newaxis, np.newaxis].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "batch = get_variable_from_np(X[:, np.newaxis, np.newaxis]).cuda()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "batch.is_cuda"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "batch = get_variable_from_np(X[:, np.newaxis, np.newaxis]).cuda()\n",
    "hiddens, outputs = rnn.get_output(batch)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([1, 196])"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "outputs.size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([1, 196])"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "target = get_variable_from_np(X[np.newaxis, :])\n",
    "target.size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([10, 5, 1])"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "torch.cat([get_variable_from_np(X[i:i+10, np.newaxis, np.newaxis]) for i in range(5)], 1).size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([10, 5, 1])"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "torch.cat([get_variable_from_np(X[i:i+10, np.newaxis, np.newaxis]) for i in range(5)], 1).size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "rnn = RNN(1,3,1)\n",
    "if torch.cuda.is_available():\n",
    "    rnn = rnn.cuda()\n",
    "criterion = nn.MSELoss()\n",
    "optimizer = torch.optim.SGD(rnn.parameters(), lr=0.001, momentum=0.9)\n",
    "#optimizer = torch.optim.Adadelta(rnn.parameters())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "'RNN' object has no attribute 'is_cuda'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-43-66df024a9ff9>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0mprint_every\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m1000\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0minputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtargets\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mget_training_data\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mX\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbatch_size\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 6\u001b[0;31m \u001b[0;32mif\u001b[0m \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcuda\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mis_available\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0mrnn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mis_cuda\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      7\u001b[0m     \u001b[0minputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0minputs\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcuda\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      8\u001b[0m     \u001b[0mtargets\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtargets\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcuda\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/home/napsternxg/anaconda3/lib/python3.5/site-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m__getattr__\u001b[0;34m(self, name)\u001b[0m\n\u001b[1;32m    241\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mname\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mmodules\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    242\u001b[0m                 \u001b[0;32mreturn\u001b[0m \u001b[0mmodules\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mname\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 243\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mobject\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__getattribute__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    244\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    245\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0m__setattr__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mname\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvalue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mAttributeError\u001b[0m: 'RNN' object has no attribute 'is_cuda'"
     ]
    }
   ],
   "source": [
    "batch_size = 1\n",
    "TIMESTEPS = X.shape[0]\n",
    "epochs = 10000\n",
    "print_every = 1000\n",
    "inputs, targets = get_training_data(X, batch_size=100)\n",
    "if torch.cuda.is_available() and rnn.is_cuda:\n",
    "    inputs = inputs.cuda()\n",
    "    targets = targets.cuda()\n",
    "print(inputs.size(), targets.size())\n",
    "losses = []\n",
    "\n",
    "for i in range(epochs):\n",
    "    optimizer.zero_grad()\n",
    "    hiddens, outputs = rnn.get_output(inputs)\n",
    "    loss = criterion(outputs, targets)\n",
    "    loss.backward()\n",
    "    optimizer.step()\n",
    "    losses.append(loss.data[0])\n",
    "    if (i+1) % print_every == 0:\n",
    "        print(\"Loss at epoch [%s]: %.3f\" % (i, loss.data[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "inputs, targets = get_training_data(X, batch_size=5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([5, 190, 1])"
      ]
     },
     "execution_count": 143,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "inputs.size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "inputs = inputs.cuda()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 145,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "torch.cuda.is_available()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([190])"
      ]
     },
     "execution_count": 133,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "outputs[:, 0].size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((196,), (190,))"
      ]
     },
     "execution_count": 134,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.shape, outputs[:, 0].data.numpy().flatten().shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {
    "collapsed": false,
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f1c7d3506d8>"
      ]
     },
     "execution_count": 135,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYYAAAD8CAYAAABzTgP2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQSsGjUaj0bRBKwaNRqPRtOH/AxQ5iQgqiwBZAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f1c7d3f5550>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(X, '-b', label='data')\n",
    "plt.plot(outputs[:, 0].data.numpy().flatten(), '-r', label='rnn') # add some offset to view each curve\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Try to use the LSTM cells and LSTM layer (Work in progress)\n",
    "\n",
    "\n",
    "\n",
    "## Use LSTM cell"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "input_size, hidden_size, output_size = 1,3,1\n",
    "lstm = nn.LSTMCell(input_size=input_size, hidden_size=hidden_size)\n",
    "output_layer = nn.Linear(hidden_size, output_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "criterion = nn.MSELoss()\n",
    "optimizer = torch.optim.SGD([\n",
    "    {\"params\": lstm.parameters()},\n",
    "    {\"params\": output_layer.parameters()}\n",
    "], lr=0.001, momentum=0.9)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Use LSTM Layer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "lstm = nn.LSTM(input_size=input_size, hidden_size=hidden_size)\n",
    "output_layer = nn.Linear(hidden_size, output_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([196, 1, 1])"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "batch = get_variable_from_np(X[:, np.newaxis, np.newaxis])\n",
    "batch.size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "hidden = Variable(torch.zeros(1, batch.size(1), hidden_size))\n",
    "cell_state = Variable(torch.zeros(1, batch.size(1), hidden_size))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "hx = (hidden, cell_state)\n",
    "output, (h_n, c_n) = lstm.forward(batch, hx)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([196, 1, 3])"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "output.size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Variable containing:\n",
       "-0.1082\n",
       "[torch.FloatTensor of size 1x1]"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "out = output_layer.forward(output[0])\n",
    "out"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "criterion = nn.MSELoss()\n",
    "optimizer = torch.optim.SGD([\n",
    "    {\"params\": lstm.parameters()},\n",
    "    {\"params\": output_layer.parameters()}\n",
    "], lr=0.001, momentum=0.9)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loss at epoch [0]: 0.810\n",
      "Loss at epoch [1]: 0.808\n",
      "Loss at epoch [2]: 0.804\n",
      "Loss at epoch [3]: 0.801\n",
      "Loss at epoch [4]: 0.797\n",
      "Loss at epoch [5]: 0.792\n",
      "Loss at epoch [6]: 0.787\n",
      "Loss at epoch [7]: 0.782\n",
      "Loss at epoch [8]: 0.777\n",
      "Loss at epoch [9]: 0.771\n"
     ]
    }
   ],
   "source": [
    "batch_size = 1\n",
    "epochs = 10\n",
    "\n",
    "inputs, targets = get_training_data(X, max_steps=1)\n",
    "\n",
    "for i in range(epochs):\n",
    "    loss = 0\n",
    "    optimizer.zero_grad()\n",
    "    hidden = Variable(torch.zeros(1, inputs.size(1), hidden_size))\n",
    "    cell_state = Variable(torch.zeros(1, inputs.size(1), hidden_size))\n",
    "    hx = (hidden, cell_state)\n",
    "    output, (h_n, c_n) = lstm.forward(inputs, hx)\n",
    "    losses = []\n",
    "    for j in range(output.size()[0]):\n",
    "        out = output_layer.forward(output[j])\n",
    "        losses.append((out - targets[j])**2)\n",
    "        #loss += criterion(out, target[i])\n",
    "    loss = torch.mean(torch.cat(losses, 1))\n",
    "    loss.backward()\n",
    "    optimizer.step()\n",
    "    print(\"Loss at epoch [%s]: %.3f\" % (i, loss.squeeze().data[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([1, 195, 3])"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "output.size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([195, 1])"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "out.size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([1, 195, 1])"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "targets.size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "y_pred = []\n",
    "for i in range(output.size()[1]):\n",
    "        out = output_layer.forward(output[i])\n",
    "        y_pred.append(out.squeeze().data[0])\n",
    "y_pred = np.array(y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7ff2e857ea20>"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYYAAAD8CAYAAABzTgP2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff2e8625828>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(X, '-b', alpha=0.5, label='data')\n",
    "plt.plot(y_pred + 0.1, '-r', label='rnn')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([-0.34993339])"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [default]",
   "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
}