fine_tuning.ipynb

{
 "metadata": {
  "name": "",
  "signature": "sha256:73b63e1a61c0cd7334b08a7df568a00a0b1ab598753886af912cf86b0d4090e9"
 },
 "nbformat": 3,
 "nbformat_minor": 0,
 "worksheets": [
  {
   "cells": [
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# Import here\n",
      "%matplotlib inline\n",
      "import matplotlib.pylab as pylab\n",
      "import matplotlib.pyplot as plt\n",
      "import numpy as np\n",
      "import theano\n",
      "import theano.tensor as T\n",
      "\n",
      "pylab.rcParams['figure.figsize'] = (9.0, 6.0)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 1
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "# architecture params\n",
      "dim_unit = 64\n",
      "n_steps = 100\n",
      "n_batch = 1024\n",
      "n_iter = 1000000\n",
      "disp_every = 10\n",
      "acce_every = 250\n",
      "save_every = 1000\n",
      "sigma = np.float32(1. / np.sqrt(dim_unit))\n",
      "\n",
      "# learning rate params\n",
      "eta_A0 = np.float32(0.001)\n",
      "eta_V0 = np.float32(0.001)\n",
      "eta_W0 = np.float32(eta_V0 / n_steps)\n",
      "eta_S0 = np.float32(eta_W0 / n_steps)\n",
      "eta_c0 = np.float32(eta_V0 / dim_unit)\n",
      "eta_b0 = np.float32(eta_W0 / dim_unit)\n",
      "eta_ratio = np.float32(100)\n",
      "mu_max = np.float(0.999)\n",
      "mu_ending = np.float(0.9)\n",
      "\n",
      "# save file path\n",
      "tmp_save = '/Users/chayut/Dropbox/data/grid_cells/opt_coding_1d_fine_tuning_tmp_save.npz'\n",
      "fin_save = '/Users/chayut/Dropbox/data/grid_cells/opt_coding_1d_fine_tuning_fin_save.npz'\n",
      "\n",
      "# initialization\n",
      "A = np.load('/Users/chayut/Dropbox/data/grid_cells/cpc_decoding_vector.npy')\n",
      "with np.load('/Users/chayut/Dropbox/data/grid_cells/opt_coding_1d_fin_save.npz') as result:\n",
      "    S = result['S']\n",
      "    W = result['W']\n",
      "    V = result['V']\n",
      "    b = result['b']\n",
      "    c = result['c']\n",
      "\n",
      "vA = np.zeros_like(A)\n",
      "vS = np.zeros_like(S)\n",
      "vV = np.zeros_like(V)\n",
      "vW = np.zeros_like(W)\n",
      "vb = np.zeros_like(b)\n",
      "vc = np.zeros_like(c)\n",
      "\n",
      "eta_A = eta_A0\n",
      "eta_V = eta_V0\n",
      "eta_W = eta_W0\n",
      "eta_S = eta_S0\n",
      "eta_c = eta_c0\n",
      "eta_b = eta_b0\n",
      "mu = np.float(0.5)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 26
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "class VariableHolder(object):\n",
      "    pass\n",
      "    \n",
      "def rectify(x):\n",
      "    return T.maximum(x, 0)\n",
      "\n",
      "def one_step(z, S, b_eff):\n",
      "    return rectify(T.dot(S, z) + b_eff)\n",
      "\n",
      "sparse_basis = np.arange(dim_unit, dtype=np.float32).reshape((dim_unit, 1))\n",
      "\n",
      "t = VariableHolder()\n",
      "t.u = T.frow(name='u')\n",
      "t.nu = T.fmatrix(name='nu')\n",
      "t.A, t.S, t.W, t.V = T.fmatrices('A', 'S', 'W', 'V')\n",
      "t.b, t.c = T.cols('b', 'c')\n",
      "\n",
      "t.u_dist = t.u * np.float32(dim_unit - 1) - sparse_basis\n",
      "t.x0 = T.exp(-(t.u_dist ** 2 / 2))\n",
      "t.x = t.x0 / T.sqrt(T.sum(t.x0 ** 2, axis=0, keepdims=True))\n",
      "t.y = rectify(T.dot(t.V, t.x) + t.c)\n",
      "t.b_eff = t.b + T.dot(t.W, t.y + t.nu)\n",
      "t.z_seq, t.updates = theano.scan(fn=one_step,\n",
      "                                 outputs_info=T.zeros_like(t.b_eff),\n",
      "                                 non_sequences=[t.S, t.b_eff],\n",
      "                                 n_steps=n_steps)\n",
      "t.x_hat = t.z_seq[-1]\n",
      "t.u_hat = T.nnet.sigmoid(T.dot(t.A, t.x_hat))\n",
      "t.cost = T.mean((t.u_hat - t.u) ** 2)\n",
      "\n",
      "t.dA = T.grad(t.cost, t.A)\n",
      "t.dS = T.grad(t.cost, t.S)\n",
      "t.dW = T.grad(t.cost, t.W)\n",
      "t.dV = T.grad(t.cost, t.V)\n",
      "t.db = T.grad(t.cost, t.b)\n",
      "t.dc = T.grad(t.cost, t.c)\n",
      "\n",
      "rnn_compute = theano.function(inputs=[t.u, t.nu, t.A, t.S, t.W, t.V, t.b, t.c],\n",
      "                              outputs=[t.cost, t.x, t.y, t.x_hat, t.u_hat,\n",
      "                                       t.dA, t.dS, t.dW, t.dV, t.db, t.dc],\n",
      "                              updates=t.updates)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [],
     "prompt_number": 27
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "for i in range(n_iter):\n",
      "    i1 = i + 1\n",
      "    if i1 % acce_every == 0:\n",
      "        eta_factor = 1.0 / (1.0 + (eta_ratio * i1) / n_iter)\n",
      "        eta_A = eta_A0 * eta_factor\n",
      "        eta_S = eta_S0 * eta_factor\n",
      "        eta_W = eta_W0 * eta_factor\n",
      "        eta_V = eta_V0 * eta_factor\n",
      "        eta_b = eta_b0 * eta_factor\n",
      "        eta_c = eta_c0 * eta_factor\n",
      "        if i1 < 0.9 * n_iter:\n",
      "            mu_unbounded = 1 - 0.5 / (1 + (i1 / acce_every))\n",
      "            mu = np.minimum(mu_max, mu_unbounded)\n",
      "        else:\n",
      "            mu = mu_ending\n",
      "    \n",
      "    # decay momentums\n",
      "    vA *= mu\n",
      "    vS *= mu\n",
      "    vW *= mu\n",
      "    vV *= mu\n",
      "    vb *= mu\n",
      "    vc *= mu\n",
      "    # update weights with momentums\n",
      "    A += vA\n",
      "    S += vS\n",
      "    W += vW\n",
      "    V += vV\n",
      "    b += vb\n",
      "    c += vc\n",
      "    V /= np.sqrt(np.sum(V ** 2, axis=1, keepdims=True))\n",
      "\n",
      "    u = np.random.uniform(size=(1, n_batch)).astype(np.float32)\n",
      "    nu = sigma * np.random.randn(dim_unit, n_batch).astype(np.float32)\n",
      "    cost, x, y, x_hat, u_hat, dA, dS, dW, dV, db, dc = rnn_compute(u, nu, A, S, W, V, b, c)\n",
      "    if i1 % disp_every == 0:\n",
      "        print i1, ':', np.sqrt(cost)\n",
      "    \n",
      "    # update momentums with new grads\n",
      "    vA -= eta_A * dA\n",
      "    vS -= eta_S * dS\n",
      "    vW -= eta_W * dW\n",
      "    vV -= eta_V * dV\n",
      "    vb -= eta_b * db\n",
      "    vc -= eta_c * dc\n",
      "    # update weights with new grads\n",
      "    A -= eta_A * dA\n",
      "    S -= eta_S * dS\n",
      "    W -= eta_W * dW\n",
      "    V -= eta_V * dV\n",
      "    b -= eta_b * db\n",
      "    c -= eta_c * dc\n",
      "    V /= np.sqrt(np.sum(V ** 2, axis=1, keepdims=True))\n",
      "    \n",
      "    if i1 % save_every == 0:\n",
      "        np.savez(tmp_save,\n",
      "                 cost=cost, u=u, u_hat=u_hat, x=x, x_hat=x_hat, y=y,\n",
      "                 A=A, S=S, W=W, V=V, b=b, c=c)"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        "10 : 0.00986188\n",
        "20"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " : 0.010486\n",
        "30"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " : 0.011162\n",
        "40"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " : 0.0096735\n",
        "50"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " : 0.0103571\n",
        "60"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " : 0.0105638\n",
        "70"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " : 0.00994699\n",
        "80"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " : 0.0099207\n",
        "90"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " : 0.0100439\n",
        "100"
       ]
      },
      {
       "output_type": "stream",
       "stream": "stdout",
       "text": [
        " : 0.0108662\n"
       ]
      },
      {
       "ename": "KeyboardInterrupt",
       "evalue": "",
       "output_type": "pyerr",
       "traceback": [
        "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m\n\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
        "\u001b[0;32m<ipython-input-28-a6eb478a3d61>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     33\u001b[0m     \u001b[0mu\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrandom\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0muniform\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0msize\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mn_batch\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mastype\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfloat32\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     34\u001b[0m     \u001b[0mnu\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msigma\u001b[0m \u001b[0;34m*\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrandom\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrandn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdim_unit\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mn_batch\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mastype\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfloat32\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 35\u001b[0;31m     \u001b[0mcost\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx_hat\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mu_hat\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdA\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdS\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdW\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdV\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdb\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdc\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mrnn_compute\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mu\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnu\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mA\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mS\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mW\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mV\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mb\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     36\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mi1\u001b[0m \u001b[0;34m%\u001b[0m \u001b[0mdisp_every\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     37\u001b[0m         \u001b[0;32mprint\u001b[0m \u001b[0mi1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m':'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msqrt\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mcost\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
        "\u001b[0;32m/Users/chayut/anaconda/lib/python2.7/site-packages/theano/compile/function_module.pyc\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m    593\u001b[0m         \u001b[0mt0_fn\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtime\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtime\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    594\u001b[0m         \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 595\u001b[0;31m             \u001b[0moutputs\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[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    596\u001b[0m         \u001b[0;32mexcept\u001b[0m \u001b[0mException\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    597\u001b[0m             \u001b[0;32mif\u001b[0m \u001b[0mhasattr\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'position_of_error'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
        "\u001b[0;32m/Users/chayut/anaconda/lib/python2.7/site-packages/theano/scan_module/scan_op.pyc\u001b[0m in \u001b[0;36mrval\u001b[0;34m(p, i, o, n, allow_gc)\u001b[0m\n\u001b[1;32m    635\u001b[0m         def rval(p=p, i=node_input_storage, o=node_output_storage, n=node,\n\u001b[1;32m    636\u001b[0m                  allow_gc=allow_gc):\n\u001b[0;32m--> 637\u001b[0;31m             \u001b[0mr\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mp\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mn\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mi\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mo\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    638\u001b[0m             \u001b[0;32mfor\u001b[0m \u001b[0mo\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mnode\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0moutputs\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    639\u001b[0m                 \u001b[0mcompute_map\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mo\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
        "\u001b[0;32m/Users/chayut/anaconda/lib/python2.7/site-packages/theano/scan_module/scan_op.pyc\u001b[0m in \u001b[0;36m<lambda>\u001b[0;34m(node, args, outs)\u001b[0m\n\u001b[1;32m    624\u001b[0m                         \u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    625\u001b[0m                         \u001b[0mouts\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 626\u001b[0;31m                         self, node)\n\u001b[0m\u001b[1;32m    627\u001b[0m         \u001b[0;32mexcept\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mImportError\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtheano\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgof\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcmodule\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mMissingGXX\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    628\u001b[0m             \u001b[0mp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mexecute\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
        "\u001b[0;32m/Users/chayut/.theano/compiledir_Darwin-14.1.0-x86_64-i386-64bit-i386-2.7.9-64/scan_perform/scan_perform.so\u001b[0m in \u001b[0;36mtheano.scan_module.scan_perform.perform (/Users/chayut/.theano/compiledir_Darwin-14.1.0-x86_64-i386-64bit-i386-2.7.9-64/scan_perform/mod.cpp:4578)\u001b[0;34m()\u001b[0m\n",
        "\u001b[0;32m/Users/chayut/anaconda/lib/python2.7/site-packages/theano/tensor/type.pyc\u001b[0m in \u001b[0;36mvalue_zeros\u001b[0;34m(self, shape)\u001b[0m\n\u001b[1;32m    581\u001b[0m             \u001b[0;32mreturn\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    582\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 583\u001b[0;31m     \u001b[0;32mdef\u001b[0m \u001b[0mvalue_zeros\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mshape\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    584\u001b[0m         \"\"\"\n\u001b[1;32m    585\u001b[0m         \u001b[0mCreate\u001b[0m \u001b[0man\u001b[0m \u001b[0mnumpy\u001b[0m \u001b[0mndarray\u001b[0m \u001b[0mfull\u001b[0m \u001b[0mof\u001b[0m \u001b[0;36m0\u001b[0m \u001b[0mvalues\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
        "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
       ]
      }
     ],
     "prompt_number": 28
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [
      "plt.imshow(V, interpolation='None')"
     ],
     "language": "python",
     "metadata": {},
     "outputs": [
      {
       "metadata": {},
       "output_type": "pyout",
       "prompt_number": 17,
       "text": [
        "<matplotlib.image.AxesImage at 0x11d676f90>"
       ]
      },
      {
       "metadata": {},
       "output_type": "display_data",
       "png": "iVBORw0KGgoAAAANSUhEUgAAAWwAAAFtCAYAAAAjw1AcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
       "text": [
        "<matplotlib.figure.Figure at 0x11d582310>"
       ]
      }
     ],
     "prompt_number": 17
    },
    {
     "cell_type": "code",
     "collapsed": false,
     "input": [],
     "language": "python",
     "metadata": {},
     "outputs": []
    }
   ],
   "metadata": {}
  }
 ]
}