MCPcopy Index your code
hub / github.com/tensorpack/tensorpack / resnext32x4d_bottleneck

Function resnext32x4d_bottleneck

examples/ResNet/resnet_model.py:107–113  ·  view source on GitHub ↗
(l, ch_out, stride)

Source from the content-addressed store, hash-verified

105
106
107def resnext32x4d_bottleneck(l, ch_out, stride):
108 shortcut = l
109 l = Conv2D('conv1', l, ch_out * 2, 1, strides=1, activation=BNReLU)
110 l = Conv2D('conv2', l, ch_out * 2, 3, strides=stride, activation=BNReLU, split=32)
111 l = Conv2D('conv3', l, ch_out * 4, 1, activation=get_bn(zero_init=True))
112 out = l + resnet_shortcut(shortcut, ch_out * 4, stride, activation=get_bn(zero_init=False))
113 return tf.nn.relu(out)
114
115
116def resnet_group(name, l, block_func, features, count, stride):

Callers

nothing calls this directly

Calls 3

Conv2DFunction · 0.90
get_bnFunction · 0.70
resnet_shortcutFunction · 0.70

Tested by

no test coverage detected