Clean notebook

python/basic/advanced_img_io.ipynb

@@ -31,10 +31,10 @@
     "import time\n",
     "# set the number of threads you want to use before importing mxnet\n",
     "os.environ['MXNET_CPU_WORKER_NTHREADS'] = '4'\n",
-    "import mxnet as mx\n",
     "import numpy as np\n",
     "import matplotlib.pyplot as plt\n",
-    "import cv2"
+    "import cv2\n",
+    "import mxnet as mx"
    ]
   },
   {
@@ -97,6 +97,7 @@
     }
    ],
    "source": [
+    "# @@@ AUTOTEST_OUTPUT_IGNORED_CELL\n",
     "# opencv\n",
     "N = 1000\n",
     "tic = time.time()\n",
@@ -133,6 +134,7 @@
     }
    ],
    "source": [
+    "# @@@ AUTOTEST_OUTPUT_IGNORED_CELL\n",
     "# mx.image\n",
     "tic = time.time()\n",
     "for i in range(N):\n",
@@ -225,6 +227,7 @@
     }
    ],
    "source": [
+    "# @@@ AUTOTEST_OUTPUT_IGNORED_CELL\n",
     "# crop a random w x h region from image\n",
     "tmp, coord = mx.image.random_crop(img, (150, 200))\n",
     "print(coord)\n",
@@ -273,7 +276,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython2",
-   "version": "2.7.12"
+   "version": "2.7.6"
   }
  },
  "nbformat": 4,

python/basic/image_io.ipynb

@@ -37,12 +37,12 @@
     "%matplotlib inline\n",
     "import os\n",
     "import subprocess\n",
-    "import mxnet as mx\n",
     "import numpy as np\n",
     "import matplotlib.pyplot as plt\n",
+    "import mxnet as mx\n",
     "\n",
     "# change this to your mxnet location\n",
-    "MXNET_HOME = '/scratch/mxnet'"
+    "MXNET_HOME = '/mxnet'"
    ]
   },
   {
@@ -165,7 +165,7 @@
    ],
    "source": [
     "data_iter = mx.io.ImageRecordIter(\n",
-    "    path_imgrec=\"./data/caltech_train.rec\", # the target record file\n",
+    "    path_imgrec=\"./data/caltech.rec\", # the target record file\n",
     "    data_shape=(3, 227, 227), # output data shape. An 227x227 region will be cropped from the original image.\n",
     "    batch_size=4, # number of samples per batch\n",
     "    resize=256 # resize the shorter edge to 256 before cropping\n",

python/basic/mixed.ipynb

@@ -821,7 +821,6 @@
   }
  ],
  "metadata": {
-  "anaconda-cloud": {},
   "kernelspec": {
    "display_name": "Python 2",
    "language": "python",

python/basic/ndarray.ipynb

@@ -765,6 +765,7 @@
     }
    ],
    "source": [
+    "# @@@ AUTOTEST_OUTPUT_IGNORED_CELL\n",
     "def f():\n",
     "    a = mx.nd.ones((100,100))\n",
     "    b = mx.nd.ones((100,100))\n",
@@ -833,6 +834,7 @@
     }
    ],
    "source": [
+    "# @@@ AUTOTEST_OUTPUT_IGNORED_CELL\n",
     "a = mx.nd.ones((100,100), mx.cpu())\n",
     "b = mx.nd.ones((100,100), mx.gpu())\n",
     "c = mx.nd.ones((100,100), mx.gpu())\n",
@@ -907,6 +909,7 @@
     }
    ],
    "source": [
+    "# @@@ AUTOTEST_OUTPUT_IGNORED_CELL\n",
     "a = mx.nd.ones((2,3))\n",
     "b = mx.nd.ones((5,6))               \n",
     "mx.nd.save(\"temp.ndarray\", [a,b])\n",
@@ -940,6 +943,7 @@
     }
    ],
    "source": [
+    "# @@@ AUTOTEST_OUTPUT_IGNORED_CELL\n",
     "d = {'a':a, 'b':b}\n",
     "mx.nd.save(\"temp.ndarray\", d)\n",
     "c = mx.nd.load(\"temp.ndarray\")\n",
@@ -1121,7 +1125,6 @@
   }
  ],
  "metadata": {
-  "anaconda-cloud": {},
   "kernelspec": {
    "display_name": "Python 2",
    "language": "python",

python/basic/record_io.ipynb

@@ -21,9 +21,9 @@
    "source": [
     "%matplotlib inline\n",
     "from __future__ import print_function\n",
-    "import mxnet as mx\n",
     "import numpy as np\n",
-    "import matplotlib.pyplot as plt"
+    "import matplotlib.pyplot as plt\n",
+    "import mxnet as mx"
    ]
   },
   {
@@ -82,7 +82,7 @@
     "    item = record.read()\n",
     "    if not item:\n",
     "        break\n",
-    "    print item\n",
+    "    print (item)\n",
     "record.close()"
    ]
   },

python/basic/symbol.ipynb

@@ -271,7 +271,7 @@
     }
    ],
    "source": [
-    "# Output may vary\n",
+    "# @@@ AUTOTEST_OUTPUT_IGNORED_CELL\n",
     "net = mx.sym.Variable('data')\n",
     "net = mx.sym.FullyConnected(data=net, name='fc1', num_hidden=128)\n",
     "net = mx.sym.Activation(data=net, name='relu1', act_type=\"relu\")\n",
@@ -402,7 +402,7 @@
     }
    ],
    "source": [
-    "# Output may vary\n",
+    "# @@@ AUTOTEST_OUTPUT_IGNORED_CELL\n",
     "def ConvFactory(data, num_filter, kernel, stride=(1,1), pad=(0, 0), name=None, suffix=''):\n",
     "    conv = mx.symbol.Convolution(data=data, num_filter=num_filter, kernel=kernel, stride=stride, pad=pad, name='conv_%s%s' %(name, suffix))\n",
     "    bn = mx.symbol.BatchNorm(data=conv, name='bn_%s%s' %(name, suffix))\n",
@@ -1478,7 +1478,6 @@
   }
  ],
  "metadata": {
-  "anaconda-cloud": {},
   "kernelspec": {
    "display_name": "Python 2",
    "language": "python",

python/cnn/mnist.ipynb

@@ -374,6 +374,7 @@
     }
    ],
    "source": [
+    "# @@@ AUTOTEST_OUTPUT_IGNORED_CELL\n",
     "model = mx.model.FeedForward(\n",
     "    ctx = mx.gpu(0),      # Run on GPU 0\n",
     "    symbol = mlp,         # Use the network we just defined\n",
@@ -384,7 +385,9 @@
     "model.fit(\n",
     "    X=train_iter,  # Training data set\n",
     "    eval_data=test_iter,  # Testing data set. MXNet computes scores on test set every epoch\n",
-    "    batch_end_callback = mx.callback.Speedometer(batch_size, 200))  # Logging module to print out progress"
+    "    batch_end_callback = mx.callback.Speedometer(batch_size, 200))  # Logging module to print out progress\n",
+    "assert model.score(train_iter) > 0.98, \"Low training accuracy.\"\n",
+    "assert model.score(test_iter) > 0.97, \"Low validation accuracy.\""
    ]
   },
   {
@@ -423,6 +426,7 @@
     }
    ],
    "source": [
+    "# @@@ AUTOTEST_OUTPUT_IGNORED_CELL\n",
     "plt.imshow((X_test[0].reshape((28,28))*255).astype(np.uint8), cmap='Greys_r')\n",
     "plt.show()\n",
     "print 'Result:', model.predict(X_test[0:1])[0].argmax()"
@@ -452,7 +456,9 @@
     }
    ],
    "source": [
-    "print 'Accuracy:', model.score(test_iter)*100, '%'"
+    "# @@@ AUTOTEST_OUTPUT_IGNORED_CELL\n",
+    "print 'Accuracy:', model.score(test_iter)*100, '%'\n",
+    "assert model.score(test_iter) > 0.97, \"Low prediction accuracy.\""
    ]
   },
   {
@@ -489,6 +495,7 @@
     }
    ],
    "source": [
+    "# @@@ AUTOTEST_OUTPUT_IGNORED_CELL\n",
     "# run hand drawing test\n",
     "from IPython.display import HTML\n",
     "\n",
@@ -574,6 +581,7 @@
     }
    ],
    "source": [
+    "# @@@ AUTOTEST_OUTPUT_IGNORED_CELL\n",
     "def norm_stat(d):\n",
     "    \"\"\"The statistics you want to see.\n",
     "    We compute the L2 norm here but you can change it to anything you like.\"\"\"\n",
@@ -586,7 +594,9 @@
     "model = mx.model.FeedForward(ctx = mx.gpu(0), symbol = mlp, num_epoch = 1,\n",
     "                             learning_rate = 0.1, momentum = 0.9, wd = 0.00001)\n",
     "model.fit(X=train_iter, eval_data=test_iter, monitor=mon,  # Set the monitor here\n",
-    "          batch_end_callback = mx.callback.Speedometer(100, 100))"
+    "          batch_end_callback = mx.callback.Speedometer(100, 100))\n",
+    "assert model.score(train_iter) > 0.94, \"Low training accuracy.\"\n",
+    "assert model.score(test_iter) > 0.94, \"Low validation accuracy.\""
    ]
   },
   {
@@ -676,6 +686,7 @@
     }
    ],
    "source": [
+    "# @@@ AUTOTEST_OUTPUT_IGNORED_CELL\n",
     "# ==================Binding=====================\n",
     "# The symbol we created is only a graph description.\n",
     "# To run it, we first need to allocate memory and create an executor by 'binding' it.\n",
@@ -740,7 +751,8 @@
     "        metric.update(batch.label, exe.outputs)\n",
     "        t += 1\n",
     "        if t % 100 == 0:\n",
-    "            print 'epoch:', epoch, 'iter:', t, 'metric:', metric.get()\n"
+    "            print 'epoch:', epoch, 'iter:', t, 'metric:', metric.get()\n",
+    "assert metric.get()[1] > 0.98, \"Low training accuracy.\""
    ]
   },
   {
@@ -786,6 +798,7 @@
     }
    ],
    "source": [
+    "# @@@ AUTOTEST_OUTPUT_IGNORED_CELL\n",
     "# Define custom softmax operator\n",
     "class NumpySoftmax(mx.operator.NumpyOp):\n",
     "    def __init__(self):\n",
@@ -840,7 +853,9 @@
     "model = mx.model.FeedForward(ctx = mx.gpu(0), symbol = mlp, num_epoch = 2,\n",
     "                             learning_rate = 0.1, momentum = 0.9, wd = 0.00001)\n",
     "model.fit(X=train_iter, eval_data=test_iter,\n",
-    "          batch_end_callback = mx.callback.Speedometer(100, 100))"
+    "          batch_end_callback = mx.callback.Speedometer(100, 100))\n",
+    "assert model.score(train_iter) > 0.96, \"Low training accuracy.\"\n",
+    "assert model.score(test_iter) > 0.95, \"Low validation accuracy.\""
    ]
   },
   {

python/how_to/predict.ipynb

@@ -25,6 +25,11 @@
    "outputs": [],
    "source": [
     "import os, urllib\n",
+    "%matplotlib inline\n",
+    "import matplotlib\n",
+    "matplotlib.rc(\"savefig\", dpi=100)\n",
+    "import matplotlib.pyplot as plt\n",
+    "import cv2\n",
     "def download(url):\n",
     "    filename = url.split(\"/\")[-1]\n",
     "    if not os.path.exists(filename):\n",
@@ -3500,11 +3505,6 @@
     }
    ],
    "source": [
-    "%matplotlib inline\n",
-    "import matplotlib\n",
-    "matplotlib.rc(\"savefig\", dpi=100)\n",
-    "import matplotlib.pyplot as plt\n",
-    "import cv2\n",
     "for i in range(0,8):\n",
     "    img = cv2.cvtColor(cv2.imread('val_1000/%d.jpg' % (i,)), cv2.COLOR_BGR2RGB)\n",
     "    plt.subplot(2,4,i+1)\n",
@@ -3531,7 +3531,6 @@
    "outputs": [],
    "source": [
     "import numpy as np\n",
-    "import cv2\n",
     "def get_image(filename):\n",
     "    img = cv2.imread(filename)  # read image in b,g,r order\n",
     "    img = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)   # change to r,g,b order\n",

python/tutorials/linear-regression.ipynb

@@ -48,8 +48,8 @@
    "outputs": [],
    "source": [
     "#Training data\n",
-    "train_data = np.array([[1,2],[3,4],[5,6],[3,2],[7,1],[6,9]])\n",
-    "train_label = np.array([5,11,17,7,9,24])\n",
+    "train_data = np.random.uniform(0, 1, [100, 2])\n",
+    "train_label = np.array([train_data[i][0] + 2 * train_data[i][1] for i in range(100)])\n",
     "batch_size = 1\n",
     "\n",
     "#Evaluation Data\n",
@@ -281,8 +281,9 @@
    },
    "outputs": [],
    "source": [
+    "# @@@ AUTOTEST_OUTPUT_IGNORED_CELL\n",
     "model.fit(train_iter, eval_iter,\n",
-    "            optimizer_params={'learning_rate':0.01, 'momentum': 0.9},\n",
+    "            optimizer_params={'learning_rate':0.005, 'momentum': 0.9},\n",
     "            num_epoch=1000,\n",
     "            batch_end_callback = mx.callback.Speedometer(batch_size, 2))"
    ]
@@ -326,6 +327,7 @@
     }
    ],
    "source": [
+    "# @@@ AUTOTEST_OUTPUT_IGNORED_CELL\n",
     "#Inference\n",
     "model.predict(eval_iter).asnumpy()"
    ]
@@ -356,9 +358,11 @@
     }
    ],
    "source": [
+    "# @@@ AUTOTEST_OUTPUT_IGNORED_CELL\n",
     "#Evaluation\n",
     "metric = mx.metric.MSE()\n",
-    "model.score(eval_iter, metric)"
+    "model.score(eval_iter, metric)\n",
+    "assert model.score(eval_iter, metric)[0][1] < 0.001, \"High mean squared error on evaluation data.\""
    ]
   },
   {
@@ -389,12 +393,14 @@
     }
    ],
    "source": [
+    "# @@@ AUTOTEST_OUTPUT_IGNORED_CELL\n",
     "#Evaluation Data\n",
     "eval_data = np.array([[7,2],[6,10],[12,2]])\n",
     "eval_label = np.array([11.1,26.1,16.1]) #Adding 0.1 to each of the values \n",
     "eval_iter = mx.io.NDArrayIter(eval_data, eval_label, batch_size, shuffle=False)\n",
     "\n",
-    "model.score(eval_iter, metric)"
+    "model.score(eval_iter, metric)\n",
+    "assert model.score(eval_iter, metric)[0][1] < 0.02, \"High mean squared error on evaluation data with noise.\""
    ]
   },
   {
@@ -430,7 +436,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython2",
-   "version": "2.7.12"
+   "version": "2.7.6"
   }
  },
  "nbformat": 4,