ResNet: GitLFS, MxNet batch normalization bug fix for spatial export

vision/classification/resnet/README.md

@@ -4,7 +4,7 @@
 ResNet models perform image classification - they take images as input and classify the major object in the image into a set of pre-defined classes. They are trained on ImageNet dataset which contains images from 1000 classes. ResNet models provide very high accuracies with affordable model sizes. They are ideal for cases when high accuracy of classification is required.
 
 ## Description
-Deeper neural networks are more difficult to train. Residual learning framework ease the training of networks that are substantially deeper. The research explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. It also provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. On the ImageNet dataset the residual nets were evaluated with a depth of up to 152 layers — 8× deeper than VGG nets but still having lower complexity. 
+Deeper neural networks are more difficult to train. Residual learning framework ease the training of networks that are substantially deeper. The research explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. It also provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. On the ImageNet dataset the residual nets were evaluated with a depth of up to 152 layers — 8× deeper than VGG nets but still having lower complexity.
 
 ## Model
 
@@ -13,35 +13,35 @@ ResNet v2 uses pre-activation function whereas ResNet v1  uses post-activation f
 
 * Version 1:
 
-|Model        |Download  |Checksum|Download (with sample test data)| ONNX version |Opset version|Top-1 accuracy (%)|Top-5 accuracy (%)| 
-|-------------|:--------------|:--------------|:--------------|:--------------|:--------------|:--------------|:--------------|
-|ResNet-18|    [44.7 MB](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet18v1/resnet18v1.onnx)    | [MD5](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet18v1/resnet18v1-md5.txt)|[42.9 MB](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet18v1/resnet18v1.tar.gz)    |  1.2.1  |7| 69.93         |    89.29|         
-|ResNet-34|    [83.3 MB](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet34v1/resnet34v1.onnx)    | [MD5](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet34v1/resnet34v1-md5.txt)| [78.6 MB](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet34v1/resnet34v1.tar.gz)    |  1.2.1   |7|73.73         |     91.40           |
-|ResNet-50|    [97.8 MB](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet50v1/resnet50v1.onnx)    | [MD5](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet50v1/resnet50v1-md5.txt)|[92.2 MB](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet50v1/resnet50v1.tar.gz)    |1.2.1    |7|74.93         |     92.38           |
-|ResNet-101|    [170.6 MB](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet101v1/resnet101v1.onnx)    | [MD5](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet101v1/resnet101v1-md5.txt)|[159.8 MB](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet101v1/resnet101v1.tar.gz)    |  1.2.1  |7  | 76.48         |     93.20           |
-|ResNet-152|    [230.6 MB](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet152v1/resnet152v1.onnx)    |[MD5](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet152v1/resnet152v1-md5.txt)|[217.2 MB](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet152v1/resnet152v1.tar.gz)    | 1.2.1  |7 |77.11         |     93.61           |
+|Model        |Download  |Download (with sample test data)| ONNX version |Opset version|Top-1 accuracy (%)|Top-5 accuracy (%)|
+|-------------|:--------------|:--------------|:--------------|:--------------|:--------------|:--------------|
+|ResNet-18|    [44.7 MB](model/resnet18-v1.onnx)    |[42.9 MB](model/resnet18-v1.tar.gz)    |  1.2.1  |7| 69.93         |    89.29|         
+|ResNet-34|    [83.3 MB](model/resnet34-v1.onnx)    | [78.6 MB](model/resnet34-v1.tar.gz)    |  1.2.1   |7|73.73         |     91.40           |
+|ResNet-50|    [97.8 MB](model/resnet50-v1.onnx)    |[92.2 MB](model/resnet50-v1.tar.gz)    |1.2.1    |7|74.93         |     92.38           |
+|ResNet-101|    [170.6 MB](model/resnet101-v1.onnx)   | [159.8 MB](model/resnet101-v1.tar.gz)    |  1.2.1  |7  | 76.48         |     93.20           |
+|ResNet-152|    [230.6 MB](model/resnet152-v1.onnx)    |[217.2 MB](model/resnet152-v1.tar.gz)    | 1.2.1  |7 |77.11         |     93.61           |
 
 
 * Version 2:
 
-|Model        |Download  |Checksum|Download (with sample test data)| ONNX version |Opset version|Top-1 accuracy (%)|Top-5 accuracy (%)| 
-|-------------|:--------------|:--------------|:--------------|:--------------|:--------------|:--------------|:--------------|
-|ResNet-18|    [44.6 MB](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet18v2/resnet18v2.onnx)    | [MD5](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet18v2/resnet18v2-md5.txt)| [42.9 MB](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet18v2/resnet18v2.tar.gz)    | 1.2.1  |7 |    69.70         |     89.49          |
-|ResNet-34|    [83.2 MB](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet34v2/resnet34v2.onnx)    | [MD5](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet34v2/resnet34v2-md5.txt)|[78.6 MB](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet34v2/resnet34v2.tar.gz)    |  1.2.1   |7| 73.36         |     91.43           |
-|ResNet-50|    [97.7 MB](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet50v2/resnet50v2.onnx)    |  [MD5](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet50v2/resnet50v2-md5.txt)|[92.0 MB](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet50v2/resnet50v2.tar.gz)    | 1.2.1 |7|75.81         |     92.82           |
-|ResNet-101|    [170.4 MB](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet101v2/resnet101v2.onnx)    |[MD5](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet101v2/resnet101v2-md5.txt)|[159.4 MB](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet101v2/resnet101v2.tar.gz)    |  1.2.1  |7 | 77.42         |     93.61           |
-|ResNet-152|    [230.3 MB](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet152v2/resnet152v2.onnx)    | [MD5](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet152v2/resnet152v2-md5.txt)|[216.0 MB](https://s3.amazonaws.com/onnx-model-zoo/resnet/resnet152v2/resnet152v2.tar.gz)    | 1.2.1   |7 | 78.20         |     94.21           |
+|Model        |Download  |Download (with sample test data)| ONNX version |Opset version|Top-1 accuracy (%)|Top-5 accuracy (%)|
+|-------------|:--------------|:--------------|:--------------|:--------------|:--------------|:--------------|
+|ResNet-18|    [44.6 MB](model/resnet18-v2.onnx)    | [42.9 MB](model/resnet18-v2.tar.gz)    | 1.2.1  |7 |    69.70         |     89.49          |
+|ResNet-34|    [83.2 MB](model/resnet34-v2.onnx)    |[78.6 MB](model/resnet34-v2.tar.gz)    |  1.2.1   |7| 73.36         |     91.43           |
+|ResNet-50|    [97.7 MB](model/resnet50-v2.onnx)   |[92.0 MB](model/resnet50-v2.tar.gz)    | 1.2.1 |7|75.81         |     92.82           |
+|ResNet-101|    [170.4 MB](model/resnet101-v2.onnx)    |[159.4 MB](model/resnet101-v2.tar.gz)    |  1.2.1  |7 | 77.42         |     93.61           |
+|ResNet-152|    [230.3 MB](model/resnet152-v2.onnx)    |[216.0 MB](model/resnet152-v2.tar.gz)    | 1.2.1   |7 | 78.20         |     94.21           |
 
 
 ## Inference
-We used MXNet as framework with gluon APIs to perform inference. View the notebook [imagenet_inference](../imagenet_inference.ipynb) to understand how to use above models for doing inference. Make sure to specify the appropriate model name in the notebook. 
+We used MXNet as framework with gluon APIs to perform inference. View the notebook [imagenet_inference](../imagenet_inference.ipynb) to understand how to use above models for doing inference. Make sure to specify the appropriate model name in the notebook.
 
-### Input 
-All pre-trained models expect input images normalized in the same way, i.e. mini-batches of 3-channel RGB images of shape (N x 3 x H x W), where N is the batch size, and H and W are expected to be at least 224. 
+### Input
+All pre-trained models expect input images normalized in the same way, i.e. mini-batches of 3-channel RGB images of shape (N x 3 x H x W), where N is the batch size, and H and W are expected to be at least 224.
 The inference was done using jpeg image.
 
 ### Preprocessing
-The images have to be loaded in to a range of [0, 1] and then normalized using mean = [0.485, 0.456, 0.406] and std = [0.229, 0.224, 0.225]. The transformation should preferrably happen at preprocessing. 
+The images have to be loaded in to a range of [0, 1] and then normalized using mean = [0.485, 0.456, 0.406] and std = [0.229, 0.224, 0.225]. The transformation should preferrably happen at preprocessing.
 
 The following code shows how to preprocess a NCHW tensor:
 

vision/classification/resnet/resnet50/README.md → vision/classification/resnet/resnet50-caffe2/README.md

@@ -10,7 +10,7 @@ Download:
 Model size: 103 MB
 
 ## Description
-ResNet-50 is a deep convolutional networks for classification.
+ResNet-50 is a deep convolutional network for classification.
 
 ### Paper
 [Deep Residual Learning for Image Recognition](https://arxiv.org/abs/1512.03385)