diff --git a/chapter08_computer-vision/object-detection.ipynb b/chapter08_computer-vision/object-detection.ipynb
index 9fd6a6b..aad2f6e 100644
--- a/chapter08_computer-vision/object-detection.ipynb
+++ b/chapter08_computer-vision/object-detection.ipynb
@@ -25,10 +25,10 @@
"so we'll abuse terminology just a little.\n",
"Second, while classifiers need only to output probabilities over classes,\n",
"object detectors must output both probabilities of class membership\n",
- "and aso the coordinates that identify the location of the objects.\n",
+ "and also the coordinates that identify the location of the objects.\n",
"\n",
"\n",
- "On this chapter we'll demonstrate the single shot multiple object detector (SSD),\n",
+ "On this chapter we'll demonstrate the single shot multiple box object detector (SSD),\n",
"a popular model for object detection that was first described in [this paper](https://arxiv.org/abs/1512.02325),\n",
"and is straightforward to implement in MXNet Gluon.\n"
]
@@ -43,11 +43,10 @@
"\n",
"The SSD model predicts anchor boxes at multiple scales. The model architecture is illustrated in the following figure. \n",
"\n",
- "\n",
+ "\n",
"\n",
"We first use a `body` network to extract the image features, \n",
- "which are used as the input to the first scale (scale 0). T\n",
- "he class labels and the corresponding anchor boxes \n",
+ "which are used as the input to the first scale (scale 0). The class labels and the corresponding anchor boxes \n",
"are predicted by `class_predictor` and `box_predictor`, respectively. \n",
"We then downsample the representations to the next scale (scale 1). \n",
"Again, at this new resolution, we predict both classes and anchor boxes. \n",
@@ -657,7 +656,7 @@
" print('Found hash mismatch in file {}, possibly due to incomplete download.'.format(file_path))\n",
" return matched\n",
"\n",
- "url_format = 'https://apache-mxnet.s3-accelerate.amazonaws.com/gluon/datasets/pikachu/{}'\n",
+ "url_format = 'https://apache-mxnet.s3-accelerate.amazonaws.com/gluon/dataset/pikachu/{}'\n",
"hashes = {'train.rec': 'e6bcb6ffba1ac04ff8a9b1115e650af56ee969c8', \n",
" 'train.idx': 'dcf7318b2602c06428b9988470c731621716c393', \n",
" 'val.rec': 'd6c33f799b4d058e82f2cb5bd9a976f69d72d520'}\n",
@@ -1299,7 +1298,7 @@
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
- "version": "3.4.3"
+ "version": "3.6.2"
}
},
"nbformat": 4,
diff --git a/img/ssd.svg b/img/ssd.svg
index be4b6dd..5da321a 100644
--- a/img/ssd.svg
+++ b/img/ssd.svg
@@ -1,3 +1,3 @@
-
+