paddle_fast_handbook.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import onnx\n",
    "import paddle\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import onnxruntime as ort\n",
    "import paddle.nn.functional as F\n",
    "from paddle.metric import Accuracy\n",
    "from paddle.static import InputSpec\n",
    "from sklearn.metrics import accuracy_score"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Param Config"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "CUDAPlace(0)"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "N_EPOCH = 2\n",
    "N_BATCH = 64\n",
    "N_BATCH_NUM = 250\n",
    "S_DATA_PATH = r\"mnist_train.csv\"\n",
    "S_PADDLE_MODEL_PATH = r\"cnn_model\"\n",
    "S_ONNX_MODEL_PATH = r\"cnn_model_batch%d.onnx\" % N_BATCH\n",
    "S_DEVICE, N_DEVICE_ID, S_DEVICE_FULL = \"gpu\", 0, \"gpu:0\"\n",
    "# S_DEVICE, N_DEVICE_ID, S_DEVICE_FULL = \"cpu\", 0, \"cpu\"\n",
    "paddle.set_device(S_DEVICE_FULL)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Read Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(42000, 785)\n",
      "(42000, 785)\n",
      "(16000, 1, 28, 28)\n",
      "(16000,)\n",
      "(26000, 1, 28, 28)\n",
      "(26000,)\n"
     ]
    }
   ],
   "source": [
    "df = pd.read_csv(S_DATA_PATH, header=None)\n",
    "print(df.shape)\n",
    "np_mat = np.array(df)\n",
    "print(np_mat.shape)\n",
    "\n",
    "X = np_mat[:, 1:]\n",
    "Y = np_mat[:, 0]\n",
    "X = X.astype(np.float32) / 255\n",
    "\n",
    "X_train = X[:N_BATCH * N_BATCH_NUM]\n",
    "X_test = X[N_BATCH * N_BATCH_NUM:]\n",
    "Y_train = Y[:N_BATCH * N_BATCH_NUM]\n",
    "Y_test = Y[N_BATCH * N_BATCH_NUM:]\n",
    "X_train = X_train.reshape(X_train.shape[0], 1, 28, 28)\n",
    "X_test = X_test.reshape(X_test.shape[0], 1, 28, 28)\n",
    "print(X_train.shape)\n",
    "print(Y_train.shape)\n",
    "print(X_test.shape)\n",
    "print(Y_test.shape)\n",
    "\n",
    "class MnistDataSet(paddle.io.Dataset):\n",
    "    def __init__(self, X, Y):\n",
    "        self.l_data, self.l_label = [], []\n",
    "        for i in range(X.shape[0]):\n",
    "            self.l_data.append(X[i, :, :, :])\n",
    "            self.l_label.append(Y[i])\n",
    "\n",
    "    def __getitem__(self, index):\n",
    "        return self.l_data[index], self.l_label[index]\n",
    "\n",
    "    def __len__(self):\n",
    "        return len(self.l_data)\n",
    "\n",
    "train_loader = paddle.io.DataLoader(MnistDataSet(X_train, Y_train),  batch_size=N_BATCH, shuffle=True)\n",
    "test_loader = paddle.io.DataLoader(MnistDataSet(X_test, Y_test), batch_size=N_BATCH, shuffle=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Build Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Net(paddle.nn.Layer):\n",
    "    def __init__(self):\n",
    "        super(Net, self).__init__()\n",
    "        self.encoder = paddle.nn.Sequential(paddle.nn.Conv2D(1, 16, 3, 1),\n",
    "                                            paddle.nn.MaxPool2D(2),\n",
    "                                            paddle.nn.Flatten(1),\n",
    "                                            paddle.nn.Linear(2704, 128),\n",
    "                                            paddle.nn.ReLU(),\n",
    "                                            paddle.nn.Linear(128, 10))\n",
    "\n",
    "    def forward(self, x):\n",
    "        out = self.encoder(x)\n",
    "        return out"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Model Train and Model Save"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "model train\n"
     ]
    },
    {
     "ename": "RuntimeError",
     "evalue": "(PreconditionNotMet) The third-party dynamic library (cudnn64_7.dll) that Paddle depends on is not configured correctly. (error code is 126)\n  Suggestions:\n  1. Check if the third-party dynamic library (e.g. CUDA, CUDNN) is installed correctly and its version is matched with paddlepaddle you installed.\n  2. Configure third-party dynamic library environment variables as follows:\n  - Linux: set LD_LIBRARY_PATH by `export LD_LIBRARY_PATH=...`\n  - Windows: set PATH by `set PATH=XXX; (at ..\\paddle\\fluid\\platform\\dynload\\dynamic_loader.cc:285)\n  [operator < gaussian_random > error]",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mRuntimeError\u001b[0m                              Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-5-3f7fdffdb192>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m      1\u001b[0m \u001b[0mprint\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m\"model train\"\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 2\u001b[1;33m \u001b[0mmodel\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mpaddle\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mModel\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mNet\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mInputSpec\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;32mNone\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m1\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m28\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m28\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;34m'float32'\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;34m'x'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mInputSpec\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;32mNone\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m10\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;34m'float32'\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;34m'x'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      3\u001b[0m \u001b[0mmodel\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mprepare\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mpaddle\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0moptimizer\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mAdam\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mlearning_rate\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;36m0.001\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mparameters\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mmodel\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mparameters\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mpaddle\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mnn\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mCrossEntropyLoss\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mAccuracy\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      4\u001b[0m model.fit(train_loader,\n\u001b[0;32m      5\u001b[0m           \u001b[0mtest_loader\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m<ipython-input-4-d1343fd93f94>\u001b[0m in \u001b[0;36m__init__\u001b[1;34m(self)\u001b[0m\n\u001b[0;32m      2\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0m__init__\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      3\u001b[0m         \u001b[0msuper\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mNet\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m__init__\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 4\u001b[1;33m         self.encoder = paddle.nn.Sequential(paddle.nn.Conv2D(1, 16, 3, 1),\n\u001b[0m\u001b[0;32m      5\u001b[0m                                             \u001b[0mpaddle\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mnn\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mMaxPool2D\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m2\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      6\u001b[0m                                             \u001b[0mpaddle\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mnn\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mFlatten\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\AppData\\Roaming\\Python\\Python37\\site-packages\\paddle\\nn\\layer\\conv.py\u001b[0m in \u001b[0;36m__init__\u001b[1;34m(self, in_channels, out_channels, kernel_size, stride, padding, dilation, groups, padding_mode, weight_attr, bias_attr, data_format)\u001b[0m\n\u001b[0;32m    654\u001b[0m             \u001b[0mweight_attr\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mweight_attr\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    655\u001b[0m             \u001b[0mbias_attr\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mbias_attr\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 656\u001b[1;33m             data_format=data_format)\n\u001b[0m\u001b[0;32m    657\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    658\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0mforward\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mx\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\AppData\\Roaming\\Python\\Python37\\site-packages\\paddle\\nn\\layer\\conv.py\u001b[0m in \u001b[0;36m__init__\u001b[1;34m(self, in_channels, out_channels, kernel_size, transposed, dims, stride, padding, padding_mode, output_padding, dilation, groups, weight_attr, bias_attr, data_format)\u001b[0m\n\u001b[0;32m    133\u001b[0m             \u001b[0mshape\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mfilter_shape\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    134\u001b[0m             \u001b[0mattr\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_param_attr\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 135\u001b[1;33m             default_initializer=_get_default_param_initializer())\n\u001b[0m\u001b[0;32m    136\u001b[0m         self.bias = self.create_parameter(\n\u001b[0;32m    137\u001b[0m             attr=self._bias_attr, shape=[self._out_channels], is_bias=True)\n",
      "\u001b[1;32m~\\AppData\\Roaming\\Python\\Python37\\site-packages\\paddle\\fluid\\dygraph\\layers.py\u001b[0m in \u001b[0;36mcreate_parameter\u001b[1;34m(self, shape, attr, dtype, is_bias, default_initializer)\u001b[0m\n\u001b[0;32m    420\u001b[0m             \u001b[0mtemp_attr\u001b[0m \u001b[1;33m=\u001b[0m \u001b[1;32mNone\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    421\u001b[0m         return self._helper.create_parameter(temp_attr, shape, dtype, is_bias,\n\u001b[1;32m--> 422\u001b[1;33m                                              default_initializer)\n\u001b[0m\u001b[0;32m    423\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    424\u001b[0m     @deprecated(\n",
      "\u001b[1;32m~\\AppData\\Roaming\\Python\\Python37\\site-packages\\paddle\\fluid\\layer_helper_base.py\u001b[0m in \u001b[0;36mcreate_parameter\u001b[1;34m(self, attr, shape, dtype, is_bias, default_initializer, stop_gradient, type)\u001b[0m\n\u001b[0;32m    376\u001b[0m                 \u001b[0mtype\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mtype\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    377\u001b[0m                 \u001b[0mstop_gradient\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mstop_gradient\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 378\u001b[1;33m                 **attr._to_kwargs(with_initializer=True))\n\u001b[0m\u001b[0;32m    379\u001b[0m         \u001b[1;32melse\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    380\u001b[0m             self.startup_program.global_block().create_parameter(\n",
      "\u001b[1;32m~\\AppData\\Roaming\\Python\\Python37\\site-packages\\paddle\\fluid\\framework.py\u001b[0m in \u001b[0;36mcreate_parameter\u001b[1;34m(self, *args, **kwargs)\u001b[0m\n\u001b[0;32m   3135\u001b[0m                 \u001b[1;32mpass\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   3136\u001b[0m             \u001b[1;32melse\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 3137\u001b[1;33m                 \u001b[0minitializer\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mparam\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m   3138\u001b[0m         \u001b[1;32mreturn\u001b[0m \u001b[0mparam\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   3139\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\AppData\\Roaming\\Python\\Python37\\site-packages\\paddle\\fluid\\initializer.py\u001b[0m in \u001b[0;36m__call__\u001b[1;34m(self, var, block)\u001b[0m\n\u001b[0;32m    360\u001b[0m                 \u001b[1;34m\"use_mkldnn\"\u001b[0m\u001b[1;33m:\u001b[0m \u001b[1;32mFalse\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    361\u001b[0m             },\n\u001b[1;32m--> 362\u001b[1;33m             stop_gradient=True)\n\u001b[0m\u001b[0;32m    363\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    364\u001b[0m         \u001b[1;32mif\u001b[0m \u001b[0mvar\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mdtype\u001b[0m \u001b[1;32min\u001b[0m \u001b[1;33m[\u001b[0m\u001b[0mVarDesc\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mVarType\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mFP16\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mVarDesc\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mVarType\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mBF16\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\AppData\\Roaming\\Python\\Python37\\site-packages\\paddle\\fluid\\framework.py\u001b[0m in \u001b[0;36mappend_op\u001b[1;34m(self, *args, **kwargs)\u001b[0m\n\u001b[0;32m   3165\u001b[0m                                        \u001b[0mkwargs\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mget\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m\"outputs\"\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m{\u001b[0m\u001b[1;33m}\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mattrs\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   3166\u001b[0m                                        \u001b[1;32mif\u001b[0m \u001b[0mattrs\u001b[0m \u001b[1;32melse\u001b[0m \u001b[1;33m{\u001b[0m\u001b[1;33m}\u001b[0m\u001b[1;33m,\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m-> 3167\u001b[1;33m                                        kwargs.get(\"stop_gradient\", False))\n\u001b[0m\u001b[0;32m   3168\u001b[0m         \u001b[1;32melse\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m   3169\u001b[0m             \u001b[1;32mfrom\u001b[0m \u001b[0mpaddle\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mfluid\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mdygraph\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mbase\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0mparam_guard\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m~\\AppData\\Roaming\\Python\\Python37\\site-packages\\paddle\\fluid\\dygraph\\tracer.py\u001b[0m in \u001b[0;36mtrace_op\u001b[1;34m(self, type, inputs, outputs, attrs, stop_gradient)\u001b[0m\n\u001b[0;32m     43\u001b[0m         self.trace(type, inputs, outputs, attrs,\n\u001b[0;32m     44\u001b[0m                    \u001b[0mframework\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_current_expected_place\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_has_grad\u001b[0m \u001b[1;32mand\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 45\u001b[1;33m                    not stop_gradient)\n\u001b[0m\u001b[0;32m     46\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     47\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0mtrain_mode\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mRuntimeError\u001b[0m: (PreconditionNotMet) The third-party dynamic library (cudnn64_7.dll) that Paddle depends on is not configured correctly. (error code is 126)\n  Suggestions:\n  1. Check if the third-party dynamic library (e.g. CUDA, CUDNN) is installed correctly and its version is matched with paddlepaddle you installed.\n  2. Configure third-party dynamic library environment variables as follows:\n  - Linux: set LD_LIBRARY_PATH by `export LD_LIBRARY_PATH=...`\n  - Windows: set PATH by `set PATH=XXX; (at ..\\paddle\\fluid\\platform\\dynload\\dynamic_loader.cc:285)\n  [operator < gaussian_random > error]"
     ]
    }
   ],
   "source": [
    "print(\"model train\")\n",
    "model = paddle.Model(Net(), InputSpec([None, 1, 28, 28], 'float32', 'x'), InputSpec([None, 10], 'float32', 'x'))\n",
    "model.prepare(paddle.optimizer.Adam(learning_rate=0.001, parameters=model.parameters()), paddle.nn.CrossEntropyLoss(), Accuracy())\n",
    "model.fit(train_loader,\n",
    "          test_loader,\n",
    "          epochs=N_EPOCH,\n",
    "          batch_size=N_BATCH,\n",
    "          save_dir=S_PADDLE_MODEL_PATH + \"_iter\",\n",
    "          verbose=1)\n",
    "model.save(S_PADDLE_MODEL_PATH + \"_final_model\")\n",
    "print()\n",
    "# model.save(S_PADDLE_MODEL_PATH) # Model save\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Model Predict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "print(\"model pred\")\n",
    "model.evaluate(test_loader, batch_size=N_BATCH, verbose=1)\n",
    "print()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Model Load and Loaded Model Use"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "print(\"load model and pred test data\")\n",
    "model_load = paddle.Model(Net(), InputSpec([None, 1, 28, 28], 'float32', 'x'), InputSpec([None, 10], 'float32', 'x'))\n",
    "# model_load.load(S_PADDLE_MODEL_PATH + \"_iter/final\")\n",
    "model_load.load(S_PADDLE_MODEL_PATH + \"_final_model\")\n",
    "model_load.prepare(paddle.optimizer.Adam(learning_rate=0.001, parameters=model.parameters()), paddle.nn.loss.CrossEntropyLoss(), Accuracy())\n",
    "model_load.evaluate(test_loader, batch_size=N_BATCH, verbose=1)\n",
    "print()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Export Onnx"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_spec = InputSpec([None, 1, 28, 28], 'float32', 'x')\n",
    "paddle.onnx.export(Net(), S_ONNX_MODEL_PATH, input_spec=[x_spec])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Onnx Run"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "S_DEVICE = \"cuda\" if S_DEVICE == \"gpu\" else S_DEVICE\n",
    "model = onnx.load(S_ONNX_MODEL_PATH + \".onnx\")\n",
    "print(onnx.checker.check_model(model))  # Check that the model is well formed\n",
    "print(onnx.helper.printable_graph(model.graph))  # Print a human readable representation of the graph\n",
    "ls_input_name, ls_output_name = [input.name for input in model.graph.input], [output.name for output in model.graph.output]\n",
    "print(\"input name \", ls_input_name)\n",
    "print(\"output name \", ls_output_name)\n",
    "s_input_name = ls_input_name[0]\n",
    "\n",
    "x_input = X_train[:N_BATCH * 2, :, :, :].astype(np.float32)\n",
    "ort_val = ort.OrtValue.ortvalue_from_numpy(x_input, S_DEVICE, N_DEVICE_ID)\n",
    "print(\"val device \", ort_val.device_name())\n",
    "print(\"val shape \", ort_val.shape())\n",
    "print(\"val data type \", ort_val.data_type())\n",
    "print(\"is_tensor \", ort_val.is_tensor())\n",
    "print(\"array_equal \", np.array_equal(ort_val.numpy(), x_input))\n",
    "providers = 'CUDAExecutionProvider' if S_DEVICE == \"cuda\" else 'CPUExecutionProvider'\n",
    "print(\"providers \", providers)\n",
    "ort_session = ort.InferenceSession(S_ONNX_MODEL_PATH + \".onnx\", providers=[providers])  # gpu运行\n",
    "ort_session.set_providers([providers])\n",
    "outputs = ort_session.run(None, {s_input_name: ort_val})\n",
    "print(\"sess env \", ort_session.get_providers())\n",
    "print(type(outputs))\n",
    "print(outputs[0])\n",
    "'''\n",
    "For example ['CUDAExecutionProvider', 'CPUExecutionProvider']\n",
    "    means execute a node using CUDAExecutionProvider if capable, otherwise execute using CPUExecutionProvider.\n",
    "'''"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}