After testing the release

README.md

@@ -25,7 +25,7 @@ torchvision >= 0.2.1
 - ResNet (['Deep Residual Learning for Image Recognition'](https://arxiv.org/abs/1512.03385))
 - PreResNet (['Identity Mappings in Deep Residual Networks'](https://arxiv.org/abs/1603.05027))
 - DenseNet (['Densely Connected Convolutional Networks'](https://arxiv.org/abs/1608.06993))
-- CondenseNet (['CondenseNet: An Efficient DenseNet using Learned Group Convolutions'](https://arxiv.org/abs/1711.09224))
+- CondenseNet (['Condense````Net: An Efficient DenseNet using Learned Group Convolutions'](https://arxiv.org/abs/1711.09224))
 - DarkNet (['Darknet: Open source neural networks in c'](https://github.com/pjreddie/darknet)) 
 - SqueezeNet (['SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model size'](https://arxiv.org/abs/1602.07360))
 - SqueezeNext (['SqueezeNext: Hardware-Aware Neural Network Design'](https://arxiv.org/abs/1803.10615))
@@ -76,8 +76,8 @@ bottleneck block. Respectively a network without b-suffix has the stride in the
 | DenseNet-161 | 22.40 | 6.18 | 28,681,000 | 7,761.25M | Converted from Gluon Model Zoo ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.3/densenet161-0618-52e30516.params.log)) |
 | DenseNet-169 | 23.89 | 6.89 | 14,149,480 | 3,381.48M | Converted from Gluon Model Zoo ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.3/densenet169-0689-281ec06b.params.log)) |
 | DenseNet-201 | 22.71 | 6.36 | 20,013,928 | 4,318.75M | Converted from Gluon Model Zoo ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.3/densenet201-0636-65b5d389.params.log)) |
-| CondenseNet-74 (C=G=4) | 26.82 | 8.64 | 4,773,944 | 533.64M | From [ShichenLiu/CondenseNet] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.4/codensenet74_c4_g4-0864-cde68fa2.params.log)) |
-| CondenseNet-74 (C=G=8) | 29.76 | 10.49 | 2,935,416 | 278.55M | From [ShichenLiu/CondenseNet] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.4/codensenet74_c8_g8-1049-4cf4a08e.params.log)) |
+| CondenseNet-74 (C=G=4) | 26.82 | 8.64 | 4,773,944 | 533.64M | From [ShichenLiu/CondenseNet] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.4/condensenett74_c4_g4-0864-cde68fa2.params.log)) |
+| CondenseNet-74 (C=G=8) | 29.76 | 10.49 | 2,935,416 | 278.55M | From [ShichenLiu/CondenseNet] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.4/condensenett74_c8_g8-1049-4cf4a08e.params.log)) |
 | SqueezeNet v1.0 | 42.81 | 19.98 | 1,248,424 | 828.30M | Converted from Gluon Model Zoo ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.5/squeezenet_v1_0-1998-1b771149.params.log)) |
 | SqueezeNet v1.1 | 43.06 | 20.23 | 1,235,496 | 354.88M | Converted from Gluon Model Zoo ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.5/squeezenet_v1_1-2023-ab455761.params.log)) |
 | 108-MENet-8x1 (g=3) | 46.11 | 22.37 | 654,516 | 40.64M | From [clavichord93/MENet] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.6/menet108_8x1_g3-2237-d3bb5a4f.params.log)) |
@@ -129,8 +129,8 @@ bottleneck block. Respectively a network without b-suffix has the stride in the
 | DenseNet-161 | 22.86 | 6.44 | 28,681,000 | 7,761.25M | Converted from Gluon Model Zoo ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.3/densenet161-0644-c0fb22c8.pth.log)) |
 | DenseNet-169 | 24.40 | 7.19 | 14,149,480 | 3,381.48M | Converted from Gluon Model Zoo ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.3/densenet169-0719-27139105.pth.log)) |
 | DenseNet-201 | 23,10 | 6.63 | 20,013,928 | 4,318.75M | Converted from Gluon Model Zoo ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.3/densenet201-0663-71ece4ad.pth.log)) |
-| CondenseNet-74 (C=G=4) | 26.25 | 8.28 | 4,773,944 | 533.64M | From [ShichenLiu/CondenseNet] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.4/codensenet74_c4_g4-0828-5ba55049.pth.log)) |
-| CondenseNet-74 (C=G=8) | 28.93 | 10.06 | 2,935,416 | 278.55M | From [ShichenLiu/CondenseNet] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.4/codensenet74_c8_g8-1006-3574d874.pth.log)) |
+| CondenseNet-74 (C=G=4) | 26.25 | 8.28 | 4,773,944 | 533.64M | From [ShichenLiu/CondenseNet] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.4/condensenett74_c4_g4-0828-5ba55049.pth.log)) |
+| CondenseNet-74 (C=G=8) | 28.93 | 10.06 | 2,935,416 | 278.55M | From [ShichenLiu/CondenseNet] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.4/condensenett74_c8_g8-1006-3574d874.pth.log)) |
 | SqueezeNet v1.0 | 41.91 | 19.58 | 1,248,424 | 828.30M | Converted from TorchVision ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.5/squeezenet_v1_0-1958-d6d59f9c.pth.log)) |
 | SqueezeNet v1.1 | 41.82 | 19.38 | 1,235,496 | 354.88M | Converted from TorchVision ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.5/squeezenet_v1_1-1938-8dcd1cc5.pth.log)) |
 | 108-MENet-8x1 (g=3) | 43.92 | 20.76 | 654,516 | 40.64M | From [clavichord93/MENet] ([log](https://github.com/osmr/imgclsmob/releases/download/v0.0.6/menet108_8x1_g3-2076-7f47b37e.pth.log)) |

gluon/models/condensenet.py

@@ -4,7 +4,7 @@
     https://arxiv.org/abs/1711.09224.
 """
 
-__all__ = ['CondenseNet', 'codensenet74_c4_g4', 'codensenet74_c8_g8']
+__all__ = ['CondenseNet', 'condensenet74_c4_g4', 'condensenet74_c8_g8']
 
 import os
 from mxnet import cpu
@@ -513,7 +513,7 @@ def get_condensenet(num_layers,
     return net
 
 
-def codensenet74_c4_g4(**kwargs):
+def condensenet74_c4_g4(**kwargs):
     """
     CondenseNet-74 (C=G=4) model (converted) from 'CondenseNet: An Efficient DenseNet using Learned Group Convolutions,'
     https://arxiv.org/abs/1711.09224.
@@ -527,10 +527,10 @@ def codensenet74_c4_g4(**kwargs):
     root : str, default '~/.mxnet/models'
         Location for keeping the model parameters.
     """
-    return get_condensenet(num_layers=74, groups=4, model_name="codensenet74_c4_g4", **kwargs)
+    return get_condensenet(num_layers=74, groups=4, model_name="condensenet74_c4_g4", **kwargs)
 
 
-def codensenet74_c8_g8(**kwargs):
+def condensenet74_c8_g8(**kwargs):
     """
     CondenseNet-74 (C=G=8) model (converted) from 'CondenseNet: An Efficient DenseNet using Learned Group Convolutions,'
     https://arxiv.org/abs/1711.09224.
@@ -544,7 +544,7 @@ def codensenet74_c8_g8(**kwargs):
     root : str, default '~/.mxnet/models'
         Location for keeping the model parameters.
     """
-    return get_condensenet(num_layers=74, groups=8, model_name="codensenet74_c8_g8", **kwargs)
+    return get_condensenet(num_layers=74, groups=8, model_name="condensenet74_c8_g8", **kwargs)
 
 
 def _test():
@@ -554,8 +554,8 @@ def _test():
     pretrained = True
 
     models = [
-        codensenet74_c4_g4,
-        codensenet74_c8_g8,
+        condensenet74_c4_g4,
+        condensenet74_c8_g8,
     ]
 
     for model in models:
@@ -573,8 +573,8 @@ def _test():
             if (param.shape is None) or (not param._differentiable):
                 continue
             weight_count += np.prod(param.shape)
-        assert (model != codensenet74_c4_g4 or weight_count == 4773944)
-        assert (model != codensenet74_c8_g8 or weight_count == 2935416)
+        assert (model != condensenet74_c4_g4 or weight_count == 4773944)
+        assert (model != condensenet74_c8_g8 or weight_count == 2935416)
 
         x = mx.nd.zeros((1, 3, 224, 224), ctx=ctx)
         y = net(x)

gluon/models/model_store.py

@@ -32,14 +32,13 @@
     ('preresnet101b', '0588', '1015145a6228aa16583a975b9c33f879ee2a6fc0', 'v0.0.2'),
     ('preresnet152', '0634', 'a509a38809f455f00bff244754a3e187323ea150', 'v0.0.2'),
     ('preresnet152b', '0575', 'dc303191ea47ca258f5abadd203b5de24d059d1a', 'v0.0.2'),
-
     ('densenet121', '0780', '49b72d04bace00bb1964b38cec13d19059a14e86', 'v0.0.3'),
     ('densenet161', '0618', '52e30516e566bdef53dcb417f86849530c83d0d1', 'v0.0.3'),
     ('densenet169', '0689', '281ec06b02f407b4523245622371da669a287044', 'v0.0.3'),
     ('densenet201', '0636', '65b5d389b1f2a18c62dc39f74960266c601fec76', 'v0.0.3'),
 
-    ('codensenet74_c4_g4', '0864', 'cde68fa2fcc9197e336717a17753a15a6efd7596', 'v0.0.4'),
-    ('codensenet74_c8_g8', '1049', '4cf4a08e7fb46f5821049dcae97ae442b0ceb546', 'v0.0.4'),
+    ('condensenet74_c4_g4', '0864', 'cde68fa2fcc9197e336717a17753a15a6efd7596', 'v0.0.4'),
+    ('condensenet74_c8_g8', '1049', '4cf4a08e7fb46f5821049dcae97ae442b0ceb546', 'v0.0.4'),
 
     ('squeezenet_v1_0', '1998', '1b771149cafb1631f70814bd40d6ee8642f30148', 'v0.0.5'),
     ('squeezenet_v1_1', '2023', 'ab45576120fa846c6e69a99ca9afe82083f0f89d', 'v0.0.5'),

pytorch/models/condensenet.py

@@ -4,7 +4,7 @@
     https://arxiv.org/abs/1711.09224.
 """
 
-__all__ = ['CondenseNet', 'codensenet74_c4_g4', 'codensenet74_c8_g8']
+__all__ = ['CondenseNet', 'condensenet74_c4_g4', 'condensenet74_c8_g8']
 
 import os
 import torch
@@ -476,7 +476,7 @@ def get_condensenet(num_layers,
     return net
 
 
-def codensenet74_c4_g4(**kwargs):
+def condensenet74_c4_g4(**kwargs):
     """
     CondenseNet-74 (C=G=4) model (converted) from 'CondenseNet: An Efficient DenseNet using Learned Group Convolutions,'
     https://arxiv.org/abs/1711.09224.
@@ -488,10 +488,10 @@ def codensenet74_c4_g4(**kwargs):
     root : str, default '~/.torch/models'
         Location for keeping the model parameters.
     """
-    return get_condensenet(num_layers=74, groups=4, model_name="codensenet74_c4_g4", **kwargs)
+    return get_condensenet(num_layers=74, groups=4, model_name="condensenet74_c4_g4", **kwargs)
 
 
-def codensenet74_c8_g8(**kwargs):
+def condensenet74_c8_g8(**kwargs):
     """
     CondenseNet-74 (C=G=8) model (converted) from 'CondenseNet: An Efficient DenseNet using Learned Group Convolutions,'
     https://arxiv.org/abs/1711.09224.
@@ -503,7 +503,7 @@ def codensenet74_c8_g8(**kwargs):
     root : str, default '~/.torch/models'
         Location for keeping the model parameters.
     """
-    return get_condensenet(num_layers=74, groups=8, model_name="codensenet74_c8_g8", **kwargs)
+    return get_condensenet(num_layers=74, groups=8, model_name="condensenet74_c8_g8", **kwargs)
 
 
 def _test():
@@ -514,8 +514,8 @@ def _test():
     pretrained = True
 
     models = [
-        codensenet74_c4_g4,
-        codensenet74_c8_g8,
+        condensenet74_c4_g4,
+        condensenet74_c8_g8,
     ]
 
     for model in models:
@@ -528,8 +528,8 @@ def _test():
         weight_count = 0
         for param in net_params:
             weight_count += np.prod(param.size())
-        assert (model != codensenet74_c4_g4 or weight_count == 4773944)
-        assert (model != codensenet74_c8_g8 or weight_count == 2935416)
+        assert (model != condensenet74_c4_g4 or weight_count == 4773944)
+        assert (model != condensenet74_c8_g8 or weight_count == 2935416)
 
         x = Variable(torch.randn(1, 3, 224, 224))
         y = net(x)

pytorch/models/model_store.py

@@ -33,14 +33,13 @@
     ('preresnet101b', '0603', 'b1e37a09424dde15ecba72365d46b1f59abd479b', 'v0.0.2'),
     ('preresnet152', '0653', '426962af4d9f6944611d36c29bcc83a3a24a268c', 'v0.0.2'),
     ('preresnet152b', '0591', '2c91ab2c8d90f3990e7c30fd6ee2184f6c2c3bee', 'v0.0.2'),
-
     ('densenet121', '0803', 'f994107a83aed162916ff89e2ded4c5af5bc6457', 'v0.0.3'),
     ('densenet161', '0644', 'c0fb22c83e8077a952ce1a0c9703d1a08b2b9e3a', 'v0.0.3'),
     ('densenet169', '0719', '271391051775ba9bbf458a6bd77af4b3007dc892', 'v0.0.3'),
     ('densenet201', '0663', '71ece4ad7be5d1e2aa4bbf6f1a6b32ac2562d847', 'v0.0.3'),
 
-    ('codensenet74_c4_g4', '0828', '5ba550494cae7081d12c14b02b2a02365539d377', 'v0.0.4'),
-    ('codensenet74_c8_g8', '1006', '3574d874fefc3307f241690bad51f20e61be1542', 'v0.0.4'),
+    ('condensenet74_c4_g4', '0828', '5ba550494cae7081d12c14b02b2a02365539d377', 'v0.0.4'),
+    ('condensenet74_c8_g8', '1006', '3574d874fefc3307f241690bad51f20e61be1542', 'v0.0.4'),
 
     ('squeezenet_v1_0', '1958', 'd6d59f9cc04c147d1f2eeeead9ac315391c3f028', 'v0.0.5'),
     ('squeezenet_v1_1', '1938', '8dcd1cc5d955f3d154bfa5be20cd278f3e77f21b', 'v0.0.5'),

pytorch/utils.py

@@ -122,8 +122,8 @@
     'sqnxt23v5_w3d2': sqnxt23v5_w3d2,
     'sqnxt23v5_w2': sqnxt23v5_w2,
 
-    'codensenet74_c4_g4': codensenet74_c4_g4,
-    'codensenet74_c8_g8': codensenet74_c8_g8,
+    'codensenet74_c4_g4': condensenet74_c4_g4,
+    'codensenet74_c8_g8': condensenet74_c8_g8,
 
     # 'oth_codensenet74_c4_g4': oth_codensenet74_c4_g4,
     # 'oth_codensenet74_c8_g8': oth_codensenet74_c8_g8,