define AlexNet architecture

Swap order of lrn and max pooling layers to reflect Krizhevsky
architecture. Initialize biases to 0.1 where Krizhevsky had 1, as
training by initialization with 1 has flat loss.

diff --git a/examples/imagenet/alexnet_deploy.prototxt b/examples/imagenet/alexnet_deploy.prototxt
new file mode 100644 (file)
index 0000000..4059fd5
--- /dev/null
+++ b/examples/imagenet/alexnet_deploy.prototxt
@@ -0,0 +1,260 @@
+name: "AlexNet"
+input: "data"
+input_dim: 10
+input_dim: 3
+input_dim: 227
+input_dim: 227
+layers {
+  layer {
+    name: "conv1"
+    type: "conv"
+    num_output: 96
+    kernelsize: 11
+    stride: 4
+    blobs_lr: 1.
+    blobs_lr: 2.
+    weight_decay: 1.
+    weight_decay: 0.
+  }
+  bottom: "data"
+  top: "conv1"
+}
+layers {
+  layer {
+    name: "relu1"
+    type: "relu"
+  }
+  bottom: "conv1"
+  top: "conv1"
+}
+layers {
+  layer {
+    name: "norm1"
+    type: "lrn"
+    local_size: 5
+    alpha: 0.0001
+    beta: 0.75
+  }
+  bottom: "conv1"
+  top: "norm1"
+}
+layers {
+  layer {
+    name: "pool1"
+    type: "pool"
+    pool: MAX
+    kernelsize: 3
+    stride: 2
+  }
+  bottom: "norm1"
+  top: "pool1"
+}
+layers {
+  layer {
+    name: "conv2"
+    type: "conv"
+    num_output: 256
+    group: 2
+    kernelsize: 5
+    pad: 2
+    blobs_lr: 1.
+    blobs_lr: 2.
+    weight_decay: 1.
+    weight_decay: 0.
+  }
+  bottom: "pool1"
+  top: "conv2"
+}
+layers {
+  layer {
+    name: "relu2"
+    type: "relu"
+  }
+  bottom: "conv2"
+  top: "conv2"
+}
+layers {
+  layer {
+    name: "norm2"
+    type: "lrn"
+    local_size: 5
+    alpha: 0.0001
+    beta: 0.75
+  }
+  bottom: "conv2"
+  top: "norm2"
+}
+layers {
+  layer {
+    name: "pool2"
+    type: "pool"
+    pool: MAX
+    kernelsize: 3
+    stride: 2
+  }
+  bottom: "norm2"
+  top: "pool2"
+}
+layers {
+  layer {
+    name: "conv3"
+    type: "conv"
+    num_output: 384
+    kernelsize: 3
+    pad: 1
+    blobs_lr: 1.
+    blobs_lr: 2.
+    weight_decay: 1.
+    weight_decay: 0.
+  }
+  bottom: "pool2"
+  top: "conv3"
+}
+layers {
+  layer {
+    name: "relu3"
+    type: "relu"
+  }
+  bottom: "conv3"
+  top: "conv3"
+}
+layers {
+  layer {
+    name: "conv4"
+    type: "conv"
+    num_output: 384
+    group: 2
+    kernelsize: 3
+    pad: 1
+    blobs_lr: 1.
+    blobs_lr: 2.
+    weight_decay: 1.
+    weight_decay: 0.
+  }
+  bottom: "conv3"
+  top: "conv4"
+}
+layers {
+  layer {
+    name: "relu4"
+    type: "relu"
+  }
+  bottom: "conv4"
+  top: "conv4"
+}
+layers {
+  layer {
+    name: "conv5"
+    type: "conv"
+    num_output: 256
+    group: 2
+    kernelsize: 3
+    pad: 1
+    blobs_lr: 1.
+    blobs_lr: 2.
+    weight_decay: 1.
+    weight_decay: 0.
+  }
+  bottom: "conv4"
+  top: "conv5"
+}
+layers {
+  layer {
+    name: "relu5"
+    type: "relu"
+  }
+  bottom: "conv5"
+  top: "conv5"
+}
+layers {
+  layer {
+    name: "pool5"
+    type: "pool"
+    kernelsize: 3
+    pool: MAX
+    stride: 2
+  }
+  bottom: "conv5"
+  top: "pool5"
+}
+layers {
+  layer {
+    name: "fc6"
+    type: "innerproduct"
+    num_output: 4096
+    blobs_lr: 1.
+    blobs_lr: 2.
+    weight_decay: 1.
+    weight_decay: 0.
+  }
+  bottom: "pool5"
+  top: "fc6"
+}
+layers {
+  layer {
+    name: "relu6"
+    type: "relu"
+  }
+  bottom: "fc6"
+  top: "fc6"
+}
+layers {
+  layer {
+    name: "drop6"
+    type: "dropout"
+    dropout_ratio: 0.5
+  }
+  bottom: "fc6"
+  top: "fc6"
+}
+layers {
+  layer {
+    name: "fc7"
+    type: "innerproduct"
+    num_output: 4096
+    blobs_lr: 1.
+    blobs_lr: 2.
+    weight_decay: 1.
+    weight_decay: 0.
+  }
+  bottom: "fc6"
+  top: "fc7"
+}
+layers {
+  layer {
+    name: "relu7"
+    type: "relu"
+  }
+  bottom: "fc7"
+  top: "fc7"
+}
+layers {
+  layer {
+    name: "drop7"
+    type: "dropout"
+    dropout_ratio: 0.5
+  }
+  bottom: "fc7"
+  top: "fc7"
+}
+layers {
+  layer {
+    name: "fc8"
+    type: "innerproduct"
+    num_output: 1000
+    blobs_lr: 1.
+    blobs_lr: 2.
+    weight_decay: 1.
+    weight_decay: 0.
+  }
+  bottom: "fc7"
+  top: "fc8"
+}
+layers {
+  layer {
+    name: "prob"
+    type: "softmax"
+  }
+  bottom: "fc8"
+  top: "prob"
+}

diff --git a/examples/imagenet/alexnet_solver.prototxt b/examples/imagenet/alexnet_solver.prototxt
new file mode 100644 (file)
index 0000000..75d0d5d
--- /dev/null
+++ b/examples/imagenet/alexnet_solver.prototxt
@@ -0,0 +1,14 @@
+train_net: "alexnet_train.prototxt"
+test_net: "alexnet_val.prototxt"
+test_iter: 1000
+test_interval: 1000
+base_lr: 0.01
+lr_policy: "step"
+gamma: 0.1
+stepsize: 100000
+display: 20
+max_iter: 450000
+momentum: 0.9
+weight_decay: 0.0005
+snapshot: 10000
+snapshot_prefix: "caffe_alexnet_train"

diff --git a/examples/imagenet/alexnet_train.prototxt b/examples/imagenet/alexnet_train.prototxt
new file mode 100644 (file)
index 0000000..c5394dc
--- /dev/null
+++ b/examples/imagenet/alexnet_train.prototxt
@@ -0,0 +1,332 @@
+name: "AlexNet"
+layers {
+  layer {
+    name: "data"
+    type: "data"
+    source: "ilsvrc12_train_leveldb"
+    meanfile: "../../data/ilsvrc12/imagenet_mean.binaryproto"
+    batchsize: 256
+    cropsize: 227
+    mirror: true
+  }
+  top: "data"
+  top: "label"
+}
+layers {
+  layer {
+    name: "conv1"
+    type: "conv"
+    num_output: 96
+    kernelsize: 11
+    stride: 4
+    weight_filler {
+      type: "gaussian"
+      std: 0.01
+    }
+    bias_filler {
+      type: "constant"
+      value: 0.
+    }
+    blobs_lr: 1.
+    blobs_lr: 2.
+    weight_decay: 1.
+    weight_decay: 0.
+  }
+  bottom: "data"
+  top: "conv1"
+}
+layers {
+  layer {
+    name: "relu1"
+    type: "relu"
+  }
+  bottom: "conv1"
+  top: "conv1"
+}
+layers {
+  layer {
+    name: "norm1"
+    type: "lrn"
+    local_size: 5
+    alpha: 0.0001
+    beta: 0.75
+  }
+  bottom: "conv1"
+  top: "norm1"
+}
+layers {
+  layer {
+    name: "pool1"
+    type: "pool"
+    pool: MAX
+    kernelsize: 3
+    stride: 2
+  }
+  bottom: "norm1"
+  top: "pool1"
+}
+layers {
+  layer {
+    name: "conv2"
+    type: "conv"
+    num_output: 256
+    group: 2
+    kernelsize: 5
+    pad: 2
+    weight_filler {
+      type: "gaussian"
+      std: 0.01
+    }
+    bias_filler {
+      type: "constant"
+      value: 0.1
+    }
+    blobs_lr: 1.
+    blobs_lr: 2.
+    weight_decay: 1.
+    weight_decay: 0.
+  }
+  bottom: "pool1"
+  top: "conv2"
+}
+layers {
+  layer {
+    name: "relu2"
+    type: "relu"
+  }
+  bottom: "conv2"
+  top: "conv2"
+}
+layers {
+  layer {
+    name: "norm2"
+    type: "lrn"
+    local_size: 5
+    alpha: 0.0001
+    beta: 0.75
+  }
+  bottom: "conv2"
+  top: "norm2"
+}
+layers {
+  layer {
+    name: "pool2"
+    type: "pool"
+    pool: MAX
+    kernelsize: 3
+    stride: 2
+  }
+  bottom: "norm2"
+  top: "pool2"
+}
+layers {
+  layer {
+    name: "conv3"
+    type: "conv"
+    num_output: 384
+    kernelsize: 3
+    pad: 1
+    weight_filler {
+      type: "gaussian"
+      std: 0.01
+    }
+    bias_filler {
+      type: "constant"
+      value: 0.
+    }
+    blobs_lr: 1.
+    blobs_lr: 2.
+    weight_decay: 1.
+    weight_decay: 0.
+  }
+  bottom: "pool2"
+  top: "conv3"
+}
+layers {
+  layer {
+    name: "relu3"
+    type: "relu"
+  }
+  bottom: "conv3"
+  top: "conv3"
+}
+layers {
+  layer {
+    name: "conv4"
+    type: "conv"
+    num_output: 384
+    group: 2
+    kernelsize: 3
+    pad: 1
+    weight_filler {
+      type: "gaussian"
+      std: 0.01
+    }
+    bias_filler {
+      type: "constant"
+      value: 0.1
+    }
+    blobs_lr: 1.
+    blobs_lr: 2.
+    weight_decay: 1.
+    weight_decay: 0.
+  }
+  bottom: "conv3"
+  top: "conv4"
+}
+layers {
+  layer {
+    name: "relu4"
+    type: "relu"
+  }
+  bottom: "conv4"
+  top: "conv4"
+}
+layers {
+  layer {
+    name: "conv5"
+    type: "conv"
+    num_output: 256
+    group: 2
+    kernelsize: 3
+    pad: 1
+    weight_filler {
+      type: "gaussian"
+      std: 0.01
+    }
+    bias_filler {
+      type: "constant"
+      value: 0.1
+    }
+    blobs_lr: 1.
+    blobs_lr: 2.
+    weight_decay: 1.
+    weight_decay: 0.
+  }
+  bottom: "conv4"
+  top: "conv5"
+}
+layers {
+  layer {
+    name: "relu5"
+    type: "relu"
+  }
+  bottom: "conv5"
+  top: "conv5"
+}
+layers {
+  layer {
+    name: "pool5"
+    type: "pool"
+    kernelsize: 3
+    pool: MAX
+    stride: 2
+  }
+  bottom: "conv5"
+  top: "pool5"
+}
+layers {
+  layer {
+    name: "fc6"
+    type: "innerproduct"
+    num_output: 4096
+    weight_filler {
+      type: "gaussian"
+      std: 0.005
+    }
+    bias_filler {
+      type: "constant"
+      value: 0.1
+    }
+    blobs_lr: 1.
+    blobs_lr: 2.
+    weight_decay: 1.
+    weight_decay: 0.
+  }
+  bottom: "pool5"
+  top: "fc6"
+}
+layers {
+  layer {
+    name: "relu6"
+    type: "relu"
+  }
+  bottom: "fc6"
+  top: "fc6"
+}
+layers {
+  layer {
+    name: "drop6"
+    type: "dropout"
+    dropout_ratio: 0.5
+  }
+  bottom: "fc6"
+  top: "fc6"
+}
+layers {
+  layer {
+    name: "fc7"
+    type: "innerproduct"
+    num_output: 4096
+    weight_filler {
+      type: "gaussian"
+      std: 0.005
+    }
+    bias_filler {
+      type: "constant"
+      value: 0.1
+    }
+    blobs_lr: 1.
+    blobs_lr: 2.
+    weight_decay: 1.
+    weight_decay: 0.
+  }
+  bottom: "fc6"
+  top: "fc7"
+}
+layers {
+  layer {
+    name: "relu7"
+    type: "relu"
+  }
+  bottom: "fc7"
+  top: "fc7"
+}
+layers {
+  layer {
+    name: "drop7"
+    type: "dropout"
+    dropout_ratio: 0.5
+  }
+  bottom: "fc7"
+  top: "fc7"
+}
+layers {
+  layer {
+    name: "fc8"
+    type: "innerproduct"
+    num_output: 1000
+    weight_filler {
+      type: "gaussian"
+      std: 0.01
+    }
+    bias_filler {
+      type: "constant"
+      value: 0.
+    }
+    blobs_lr: 1.
+    blobs_lr: 2.
+    weight_decay: 1.
+    weight_decay: 0.
+  }
+  bottom: "fc7"
+  top: "fc8"
+}
+layers {
+  layer {
+    name: "loss"
+    type: "softmax_loss"
+  }
+  bottom: "fc8"
+  bottom: "label"
+}

diff --git a/examples/imagenet/alexnet_val.prototxt b/examples/imagenet/alexnet_val.prototxt
new file mode 100644 (file)
index 0000000..aff33d0
--- /dev/null
+++ b/examples/imagenet/alexnet_val.prototxt
@@ -0,0 +1,245 @@
+name: "AlexNet"
+layers {
+  layer {
+    name: "data"
+    type: "data"
+    source: "ilsvrc12_val_leveldb"
+    meanfile: "../../data/ilsvrc12/imagenet_mean.binaryproto"
+    batchsize: 50
+    cropsize: 227
+    mirror: false
+  }
+  top: "data"
+  top: "label"
+}
+layers {
+  layer {
+    name: "conv1"
+    type: "conv"
+    num_output: 96
+    kernelsize: 11
+    stride: 4
+  }
+  bottom: "data"
+  top: "conv1"
+}
+layers {
+  layer {
+    name: "relu1"
+    type: "relu"
+  }
+  bottom: "conv1"
+  top: "conv1"
+}
+layers {
+  layer {
+    name: "norm1"
+    type: "lrn"
+    local_size: 5
+    alpha: 0.0001
+    beta: 0.75
+  }
+  bottom: "conv1"
+  top: "norm1"
+}
+layers {
+  layer {
+    name: "pool1"
+    type: "pool"
+    pool: MAX
+    kernelsize: 3
+    stride: 2
+  }
+  bottom: "norm1"
+  top: "pool1"
+}
+layers {
+  layer {
+    name: "conv2"
+    type: "conv"
+    num_output: 256
+    group: 2
+    kernelsize: 5
+    pad: 2
+  }
+  bottom: "pool1"
+  top: "conv2"
+}
+layers {
+  layer {
+    name: "relu2"
+    type: "relu"
+  }
+  bottom: "conv2"
+  top: "conv2"
+}
+layers {
+  layer {
+    name: "norm2"
+    type: "lrn"
+    local_size: 5
+    alpha: 0.0001
+    beta: 0.75
+  }
+  bottom: "conv2"
+  top: "norm2"
+}
+layers {
+  layer {
+    name: "pool2"
+    type: "pool"
+    pool: MAX
+    kernelsize: 3
+    stride: 2
+  }
+  bottom: "norm2"
+  top: "pool2"
+}
+layers {
+  layer {
+    name: "conv3"
+    type: "conv"
+    num_output: 384
+    kernelsize: 3
+    pad: 1
+  }
+  bottom: "pool2"
+  top: "conv3"
+}
+layers {
+  layer {
+    name: "relu3"
+    type: "relu"
+  }
+  bottom: "conv3"
+  top: "conv3"
+}
+layers {
+  layer {
+    name: "conv4"
+    type: "conv"
+    num_output: 384
+    group: 2
+    kernelsize: 3
+    pad: 1
+  }
+  bottom: "conv3"
+  top: "conv4"
+}
+layers {
+  layer {
+    name: "relu4"
+    type: "relu"
+  }
+  bottom: "conv4"
+  top: "conv4"
+}
+layers {
+  layer {
+    name: "conv5"
+    type: "conv"
+    num_output: 256
+    group: 2
+    kernelsize: 3
+    pad: 1
+  }
+  bottom: "conv4"
+  top: "conv5"
+}
+layers {
+  layer {
+    name: "relu5"
+    type: "relu"
+  }
+  bottom: "conv5"
+  top: "conv5"
+}
+layers {
+  layer {
+    name: "pool5"
+    type: "pool"
+    kernelsize: 3
+    pool: MAX
+    stride: 2
+  }
+  bottom: "conv5"
+  top: "pool5"
+}
+layers {
+  layer {
+    name: "fc6"
+    type: "innerproduct"
+    num_output: 4096
+  }
+  bottom: "pool5"
+  top: "fc6"
+}
+layers {
+  layer {
+    name: "relu6"
+    type: "relu"
+  }
+  bottom: "fc6"
+  top: "fc6"
+}
+layers {
+  layer {
+    name: "drop6"
+    type: "dropout"
+    dropout_ratio: 0.5
+  }
+  bottom: "fc6"
+  top: "fc6"
+}
+layers {
+  layer {
+    name: "fc7"
+    type: "innerproduct"
+    num_output: 4096
+  }
+  bottom: "fc6"
+  top: "fc7"
+}
+layers {
+  layer {
+    name: "relu7"
+    type: "relu"
+  }
+  bottom: "fc7"
+  top: "fc7"
+}
+layers {
+  layer {
+    name: "drop7"
+    type: "dropout"
+    dropout_ratio: 0.5
+  }
+  bottom: "fc7"
+  top: "fc7"
+}
+layers {
+  layer {
+    name: "fc8"
+    type: "innerproduct"
+    num_output: 1000
+  }
+  bottom: "fc7"
+  top: "fc8"
+}
+layers {
+  layer {
+    name: "prob"
+    type: "softmax"
+  }
+  bottom: "fc8"
+  top: "prob"
+}
+layers {
+  layer {
+    name: "accuracy"
+    type: "accuracy"
+  }
+  bottom: "prob"
+  bottom: "label"
+  top: "accuracy"
+}

diff --git a/examples/imagenet/train_alexnet.sh b/examples/imagenet/train_alexnet.sh
new file mode 100755 (executable)
index 0000000..c91e910
--- /dev/null
+++ b/examples/imagenet/train_alexnet.sh
@@ -0,0 +1,7 @@
+#!/usr/bin/env sh
+
+TOOLS=../../build/tools
+
+GLOG_logtostderr=1 $TOOLS/train_net.bin alexnet_solver.prototxt
+
+echo "Done."