1.梯度拟合sin函数.ipynb

{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"name":"1.梯度拟合sin函数","provenance":[],"collapsed_sections":[],"authorship_tag":"ABX9TyMZOQhrpc97jf5HKxyuTPb5"},"kernelspec":{"name":"python3","display_name":"Python 3"}},"cells":[{"cell_type":"code","metadata":{"id":"QP9ba_WMWLbY","executionInfo":{"status":"ok","timestamp":1613313294212,"user_tz":-480,"elapsed":982,"user":{"displayName":"白梦伟","photoUrl":"","userId":"14318251260292756104"}}},"source":["import torch\r\n","import numpy  as np\r\n","import matplotlib.pyplot as plt\r\n","import time"],"execution_count":155,"outputs":[]},{"cell_type":"code","metadata":{"id":"RkbbINoWWUbr","executionInfo":{"status":"ok","timestamp":1613315693254,"user_tz":-480,"elapsed":784,"user":{"displayName":"白梦伟","photoUrl":"","userId":"14318251260292756104"}}},"source":["class calculate():\r\n","  def __init__(self):\r\n","    self.target_function = self.target_function\r\n","    self.predict_function = self.predict_function \r\n","\r\n","    self.dimension = 4\r\n","    self.learning_rate = 0.001\r\n","    self.weight_array = np.array([np.random.uniform() for i in range(self.dimension)])\r\n","\r\n","  def start_execute(self, count):\r\n","    \r\n","    # import pdb; pdb.set_trace()\r\n","    for i in range(count):\r\n","      value_x = np.pi * (np.random.rand()-0.5)\r\n","      loss = np.square(self.target_function(value_x) - self.predict_function(value_x))\r\n","      value_array = [ np.power(value_x, i) for i in range(self.dimension) ]\r\n","      delta_loss = np.dot((self.predict_function(value_x)-self.target_function(value_x)) , value_array) * self.learning_rate\r\n","      self.weight_array -= delta_loss\r\n","      if i%5000 == 0:\r\n","        print('%20s, %20s, %80s'%(loss, self.predict_function(value_x), self.weight_array))\r\n","\r\n","  def predict_function(self, value_x):\r\n","    value_array = np.array([np.power(value_x, i) for i in range(self.dimension)])\r\n","    return np.dot(self.weight_array , value_array)\r\n","\r\n","  def target_function(self, value_x):\r\n","    return np.sin(value_x)\r\n","\r\n","\r\n"],"execution_count":189,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"J6kr0RaPWUZI","executionInfo":{"status":"ok","timestamp":1613315701475,"user_tz":-480,"elapsed":5894,"user":{"displayName":"白梦伟","photoUrl":"","userId":"14318251260292756104"}},"outputId":"04b6cefe-8f1b-4938-9e29-3520f32cd77e"},"source":["test_grad = calculate()\r\n","test_grad.start_execute(100000)\r\n","print('-----------')\r\n","print(test_grad.predict_function(np.pi/6))"],"execution_count":190,"outputs":[{"output_type":"stream","text":["   6.743795420158474,   3.5353289216450725,                                    [0.58942053 0.45044154 0.93574809 0.33253632]\n"," 0.03754519573219257,  0.31925739643978157,                                [-0.02657624  0.6107694   0.019351    0.07301092]\n","0.012320257809477016,   0.4159911143998379,                                [-0.0059934   0.76036794  0.00338716 -0.00562033]\n","1.5504235894538453e-05,   0.9722409092057246,                [-2.94457220e-03  8.46689381e-01  5.53358256e-04 -6.20150538e-02]\n","0.0008618745295826148, -0.29081079404845916,                [-1.63340844e-05  9.01976900e-01 -8.77675174e-04 -9.32908041e-02]\n","0.00014651140059193177,  0.16881372980801324,                [-7.33836714e-04  9.35813289e-01  4.47920550e-05 -1.13778388e-01]\n","0.00012426689243883848,   0.2621196835395808,                [ 9.51040921e-05  9.56425898e-01 -4.24745097e-04 -1.25362326e-01]\n","2.5673317776308437e-05,   0.1682656762710589,                [ 2.08964802e-04  9.68760791e-01 -1.07190066e-04 -1.33365484e-01]\n","1.7432310580940392e-06, -0.053526226627484554,                [ 4.60049602e-05  9.76712040e-01 -4.31631981e-05 -1.37838277e-01]\n","8.92174079643937e-06,   0.9477694930160595,                [-6.43395356e-06  9.81443811e-01  4.32024949e-05 -1.40710068e-01]\n","1.1243659106005188e-06, -0.05855202404323157,                [ 1.29587513e-04  9.84323520e-01 -5.15350272e-05 -1.42334722e-01]\n","7.6017791218980495e-06,   0.9064511103056696,                [-3.37636478e-05  9.85956949e-01  1.03658210e-04 -1.43503557e-01]\n","5.240282601998046e-06, -0.18522989438727328,                [ 5.53042504e-06  9.87108815e-01 -1.57601541e-05 -1.44039730e-01]\n","1.2172607316480706e-05,   0.5009144000053615,                [ 9.03536426e-05  9.87707904e-01 -1.87304403e-06 -1.44560827e-01]\n","1.287484919982015e-06,  0.10449469093576574,                [ 9.29153825e-05  9.88169801e-01 -4.39394941e-05 -1.44644858e-01]\n","6.67932426263041e-07,  -0.7122388536808053,                [ 5.92416264e-05  9.88559819e-01  1.30845761e-04 -1.44686334e-01]\n","8.710141823587269e-06,  -0.5345225554626177,                [-3.15229775e-05  9.88602248e-01 -1.83906081e-05 -1.44940031e-01]\n","6.18489896989477e-07,   0.7849199037327302,                [-3.51510731e-05  9.88620298e-01 -5.31325927e-05 -1.44840288e-01]\n","2.7579977351714246e-07,  0.04359154553652876,                [-2.31995809e-05  9.88595933e-01 -1.25909621e-05 -1.45111531e-01]\n","9.69189802337329e-06,  -0.4778828134959668,                [-6.30044007e-05  9.88698510e-01 -7.07775373e-05 -1.44943603e-01]\n","-----------\n","0.49693299312738654\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"LXQhGQrdWUWx","executionInfo":{"status":"ok","timestamp":1613309044070,"user_tz":-480,"elapsed":33309,"user":{"displayName":"白梦伟","photoUrl":"","userId":"14318251260292756104"}},"outputId":"0d5a2b4b-bf12-4dd0-9185-4c7104fd00a8"},"source":["%debug"],"execution_count":30,"outputs":[{"output_type":"stream","text":["> \u001b[0;32m<ipython-input-28-df20550947e1>\u001b[0m(15)\u001b[0;36mstart_execute\u001b[0;34m()\u001b[0m\n","\u001b[0;32m     13 \u001b[0;31m      \u001b[0mloss\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msquare\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtarget_function\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvalue_x\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict_function\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvalue_x\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[0m\u001b[0;32m     14 \u001b[0;31m      \u001b[0mvalue_array\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpower\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvalue_x\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mi\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mi\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdimension\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[0m\u001b[0;32m---> 15 \u001b[0;31m      \u001b[0mdelta_loss\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpredict_function\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvalue_x\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtarget_function\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mvalue_x\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m*\u001b[0m \u001b[0mvalue_array\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0m\u001b[0;32m     16 \u001b[0;31m      \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mweight_array\u001b[0m \u001b[0;34m-=\u001b[0m \u001b[0mdelta_loss\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0m\u001b[0;32m     17 \u001b[0;31m      \u001b[0;32mif\u001b[0m \u001b[0mi\u001b[0m\u001b[0;34m%\u001b[0m\u001b[0;36m10\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0m\n","ipdb> value_array \n","[1, 0, 0, 0]\n","ipdb> 2*(self.predict_function(value_x)-self.target_function(value_x))\n","1.5434175428024248\n","ipdb>  value_array\n","[1, 0, 0, 0]\n","ipdb> q\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":284},"id":"6JFf8RaeWUUZ","executionInfo":{"status":"ok","timestamp":1613315851472,"user_tz":-480,"elapsed":779,"user":{"displayName":"白梦伟","photoUrl":"","userId":"14318251260292756104"}},"outputId":"85389e3e-00df-463f-d95f-9f1cedefecd7"},"source":["x = np.linspace(-np.pi, np.pi, 1000)\r\n","y_predict = test_grad.predict_function(x)\r\n","y_target = test_grad.target_function(x)\r\n","plt.plot(x,y_predict)\r\n","plt.plot(x,y_target)\r\n","plt.legend(['predict', 'target'])"],"execution_count":193,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.legend.Legend at 0x7f4d3933a710>"]},"metadata":{"tags":[]},"execution_count":193},{"output_type":"display_data","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[],"needs_background":"light"}}]},{"cell_type":"code","metadata":{"id":"BE8I-NFGWUSI"},"source":[""],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"7ZAsRmDrWUPp"},"source":[""],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"gcWIM1GxWUNB"},"source":[""],"execution_count":null,"outputs":[]}]}