【caffe】mnist训练日志

@tags caffe

前面根据train_lenet.sh改写了train_lenet.py后，在根目录下执行它，得到一系列输出，内容如下：

I1013 10:05:16.721294  1684 caffe.cpp:218] Using GPUs 0
I1013 10:05:17.525264  1684 caffe.cpp:223] GPU 0: GeForce GTX 970M
I1013 10:05:17.790920  1684 common.cpp:36] System entropy source not available, using fallback algorithm to generate seed instead.
I1013 10:05:17.806543  1684 solver.cpp:48] Initializing solver from parameters:
test_iter: 100
test_interval: 500
base_lr: 0.01
display: 100
max_iter: 10000
lr_policy: "inv"
gamma: 0.0001
power: 0.75
momentum: 0.9
weight_decay: 0.0005
snapshot: 5000
snapshot_prefix: "examples/mnist/lenet"
solver_mode: GPU
device_id: 0
net: "examples/mnist/lenet_train_test.prototxt"
train_state {
  level: 0
  stage: ""
}
I1013 10:05:17.806543  1684 solver.cpp:91] Creating training net from net file: examples/mnist/lenet_train_test.prototxt
I1013 10:05:17.806543  1684 net.cpp:332] The NetState phase (0) differed from the phase (1) specified by a rule in layer mnist
I1013 10:05:17.806543  1684 net.cpp:332] The NetState phase (0) differed from the phase (1) specified by a rule in layer accuracy
I1013 10:05:17.806543  1684 net.cpp:58] Initializing net from parameters:
name: "LeNet"
state {
  phase: TRAIN
  level: 0
  stage: ""
}
layer {
  name: "mnist"
  type: "Data"
  top: "data"
  top: "label"
  include {
    phase: TRAIN
  }
  transform_param {
    scale: 0.00390625
  }
  data_param {
    source: "examples/mnist/mnist_train_lmdb"
    batch_size: 64
    backend: LMDB
  }
}
layer {
  name: "conv1"
  type: "Convolution"
  bottom: "data"
  top: "conv1"
  param {
    lr_mult: 1
  }
  param {
    lr_mult: 2
  }
  convolution_param {
    num_output: 20
    kernel_size: 5
    stride: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}
layer {
  name: "pool1"
  type: "Pooling"
  bottom: "conv1"
  top: "pool1"
  pooling_param {
    pool: MAX
    kernel_size: 2
    stride: 2
  }
}
layer {
  name: "conv2"
  type: "Convolution"
  bottom: "pool1"
  top: "conv2"
  param {
    lr_mult: 1
  }
  param {
    lr_mult: 2
  }
  convolution_param {
    num_output: 50
    kernel_size: 5
    stride: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}
layer {
  name: "pool2"
  type: "Pooling"
  bottom: "conv2"
  top: "pool2"
  pooling_param {
    pool: MAX
    kernel_size: 2
    stride: 2
  }
}
layer {
  name: "ip1"
  type: "InnerProduct"
  bottom: "pool2"
  top: "ip1"
  param {
    lr_mult: 1
  }
  param {
    lr_mult: 2
  }
  inner_product_param {
    num_output: 500
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}
layer {
  name: "relu1"
  type: "ReLU"
  bottom: "ip1"
  top: "ip1"
}
layer {
  name: "ip2"
  type: "InnerProduct"
  bottom: "ip1"
  top: "ip2"
  param {
    lr_mult: 1
  }
  param {
    lr_mult: 2
  }
  inner_product_param {
    num_output: 10
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}
layer {
  name: "loss"
  type: "SoftmaxWithLoss"
  bottom: "ip2"
  bottom: "label"
  top: "loss"
}
I1013 10:05:17.822134  1684 layer_factory.hpp:77] Creating layer mnist
I1013 10:05:17.853427  1684 common.cpp:36] System entropy source not available, using fallback algorithm to generate seed instead.
I1013 10:05:17.853427  1684 net.cpp:100] Creating Layer mnist
I1013 10:05:17.853427  1684 net.cpp:418] mnist -> data
I1013 10:05:17.853427  1684 net.cpp:418] mnist -> label
I1013 10:05:17.853427 10084 common.cpp:36] System entropy source not available, using fallback algorithm to generate seed instead.
I1013 10:05:17.900490 10084 db_lmdb.cpp:40] Opened lmdb examples/mnist/mnist_train_lmdb
I1013 10:05:17.978623  1684 data_layer.cpp:41] output data size: 64,1,28,28
I1013 10:05:17.978623  1684 net.cpp:150] Setting up mnist
I1013 10:05:17.978623  1684 net.cpp:157] Top shape: 64 1 28 28 (50176)
I1013 10:05:17.978623   824 common.cpp:36] System entropy source not available, using fallback algorithm to generate seed instead.
I1013 10:05:17.978623  1684 net.cpp:157] Top shape: 64 (64)
I1013 10:05:17.978623  1684 net.cpp:165] Memory required for data: 200960
I1013 10:05:17.978623  1684 layer_factory.hpp:77] Creating layer conv1
I1013 10:05:17.978623  1684 net.cpp:100] Creating Layer conv1
I1013 10:05:17.978623  1684 net.cpp:444] conv1 <- data
I1013 10:05:17.978623  1684 net.cpp:418] conv1 -> conv1
I1013 10:05:17.994026  1684 net.cpp:150] Setting up conv1
I1013 10:05:17.994026  1684 net.cpp:157] Top shape: 64 20 24 24 (737280)
I1013 10:05:17.994026  1684 net.cpp:165] Memory required for data: 3150080
I1013 10:05:17.994026  1684 layer_factory.hpp:77] Creating layer pool1
I1013 10:05:17.994026  1684 net.cpp:100] Creating Layer pool1
I1013 10:05:17.994026  1684 net.cpp:444] pool1 <- conv1
I1013 10:05:17.994026  1684 net.cpp:418] pool1 -> pool1
I1013 10:05:17.994026  1684 net.cpp:150] Setting up pool1
I1013 10:05:17.994026  1684 net.cpp:157] Top shape: 64 20 12 12 (184320)
I1013 10:05:17.994026  1684 net.cpp:165] Memory required for data: 3887360
I1013 10:05:18.009652  1684 layer_factory.hpp:77] Creating layer conv2
I1013 10:05:18.009652  1684 net.cpp:100] Creating Layer conv2
I1013 10:05:18.009652  1684 net.cpp:444] conv2 <- pool1
I1013 10:05:18.025316  1684 net.cpp:418] conv2 -> conv2
I1013 10:05:18.025316  1684 net.cpp:150] Setting up conv2
I1013 10:05:18.025316  1684 net.cpp:157] Top shape: 64 50 8 8 (204800)
I1013 10:05:18.025316  1684 net.cpp:165] Memory required for data: 4706560
I1013 10:05:18.025316  1684 layer_factory.hpp:77] Creating layer pool2
I1013 10:05:18.040946  1684 net.cpp:100] Creating Layer pool2
I1013 10:05:18.040946  1684 net.cpp:444] pool2 <- conv2
I1013 10:05:18.040946  1684 net.cpp:418] pool2 -> pool2
I1013 10:05:18.040946  1684 net.cpp:150] Setting up pool2
I1013 10:05:18.040946  1684 net.cpp:157] Top shape: 64 50 4 4 (51200)
I1013 10:05:18.040946  1684 net.cpp:165] Memory required for data: 4911360
I1013 10:05:18.056536  1684 layer_factory.hpp:77] Creating layer ip1
I1013 10:05:18.056536  1684 net.cpp:100] Creating Layer ip1
I1013 10:05:18.056536  1684 net.cpp:444] ip1 <- pool2
I1013 10:05:18.056536  1684 net.cpp:418] ip1 -> ip1
I1013 10:05:18.087842  1684 net.cpp:150] Setting up ip1
I1013 10:05:18.087842  1684 net.cpp:157] Top shape: 64 500 (32000)
I1013 10:05:18.087842  1684 net.cpp:165] Memory required for data: 5039360
I1013 10:05:18.087842  1684 layer_factory.hpp:77] Creating layer relu1
I1013 10:05:18.087842  1684 net.cpp:100] Creating Layer relu1
I1013 10:05:18.103415  1684 net.cpp:444] relu1 <- ip1
I1013 10:05:18.103415  1684 net.cpp:405] relu1 -> ip1 (in-place)
I1013 10:05:18.103415  1684 net.cpp:150] Setting up relu1
I1013 10:05:18.103415  1684 net.cpp:157] Top shape: 64 500 (32000)
I1013 10:05:18.103415  1684 net.cpp:165] Memory required for data: 5167360
I1013 10:05:18.119084  1684 layer_factory.hpp:77] Creating layer ip2
I1013 10:05:18.119084  1684 net.cpp:100] Creating Layer ip2
I1013 10:05:18.119084  1684 net.cpp:444] ip2 <- ip1
I1013 10:05:18.119084  1684 net.cpp:418] ip2 -> ip2
I1013 10:05:18.134666  1684 net.cpp:150] Setting up ip2
I1013 10:05:18.134666  1684 net.cpp:157] Top shape: 64 10 (640)
I1013 10:05:18.134666  1684 net.cpp:165] Memory required for data: 5169920
I1013 10:05:18.134666  1684 layer_factory.hpp:77] Creating layer loss
I1013 10:05:18.134666  1684 net.cpp:100] Creating Layer loss
I1013 10:05:18.150292  1684 net.cpp:444] loss <- ip2
I1013 10:05:18.150292  1684 net.cpp:444] loss <- label
I1013 10:05:18.150292  1684 net.cpp:418] loss -> loss
I1013 10:05:18.150292  1684 layer_factory.hpp:77] Creating layer loss
I1013 10:05:18.150292  1684 net.cpp:150] Setting up loss
I1013 10:05:18.165921  1684 net.cpp:157] Top shape: (1)
I1013 10:05:18.165921  1684 net.cpp:160]     with loss weight 1
I1013 10:05:18.165921  1684 net.cpp:165] Memory required for data: 5169924
I1013 10:05:18.165921  1684 net.cpp:226] loss needs backward computation.
I1013 10:05:18.181591  1684 net.cpp:226] ip2 needs backward computation.
I1013 10:05:18.181591  1684 net.cpp:226] relu1 needs backward computation.
I1013 10:05:18.181591  1684 net.cpp:226] ip1 needs backward computation.
I1013 10:05:18.181591  1684 net.cpp:226] pool2 needs backward computation.
I1013 10:05:18.197201  1684 net.cpp:226] conv2 needs backward computation.
I1013 10:05:18.197201  1684 net.cpp:226] pool1 needs backward computation.
I1013 10:05:18.197201  1684 net.cpp:226] conv1 needs backward computation.
I1013 10:05:18.197201  1684 net.cpp:228] mnist does not need backward computation.
I1013 10:05:18.212836  1684 net.cpp:270] This network produces output loss
I1013 10:05:18.212836  1684 net.cpp:283] Network initialization done.
I1013 10:05:18.212836  1684 solver.cpp:181] Creating test net (#0) specified by net file: examples/mnist/lenet_train_test.prototxt
I1013 10:05:18.228471  1684 net.cpp:332] The NetState phase (1) differed from the phase (0) specified by a rule in layer mnist
I1013 10:05:18.228471  1684 net.cpp:58] Initializing net from parameters:
name: "LeNet"
state {
  phase: TEST
}
layer {
  name: "mnist"
  type: "Data"
  top: "data"
  top: "label"
  include {
    phase: TEST
  }
  transform_param {
    scale: 0.00390625
  }
  data_param {
    source: "examples/mnist/mnist_test_lmdb"
    batch_size: 100
    backend: LMDB
  }
}
layer {
  name: "conv1"
  type: "Convolution"
  bottom: "data"
  top: "conv1"
  param {
    lr_mult: 1
  }
  param {
    lr_mult: 2
  }
  convolution_param {
    num_output: 20
    kernel_size: 5
    stride: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}
layer {
  name: "pool1"
  type: "Pooling"
  bottom: "conv1"
  top: "pool1"
  pooling_param {
    pool: MAX
    kernel_size: 2
    stride: 2
  }
}
layer {
  name: "conv2"
  type: "Convolution"
  bottom: "pool1"
  top: "conv2"
  param {
    lr_mult: 1
  }
  param {
    lr_mult: 2
  }
  convolution_param {
    num_output: 50
    kernel_size: 5
    stride: 1
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}
layer {
  name: "pool2"
  type: "Pooling"
  bottom: "conv2"
  top: "pool2"
  pooling_param {
    pool: MAX
    kernel_size: 2
    stride: 2
  }
}
layer {
  name: "ip1"
  type: "InnerProduct"
  bottom: "pool2"
  top: "ip1"
  param {
    lr_mult: 1
  }
  param {
    lr_mult: 2
  }
  inner_product_param {
    num_output: 500
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}
layer {
  name: "relu1"
  type: "ReLU"
  bottom: "ip1"
  top: "ip1"
}
layer {
  name: "ip2"
  type: "InnerProduct"
  bottom: "ip1"
  top: "ip2"
  param {
    lr_mult: 1
  }
  param {
    lr_mult: 2
  }
  inner_product_param {
    num_output: 10
    weight_filler {
      type: "xavier"
    }
    bias_filler {
      type: "constant"
    }
  }
}
layer {
  name: "accuracy"
  type: "Accuracy"
  bottom: "ip2"
  bottom: "label"
  top: "accuracy"
  include {
    phase: TEST
  }
}
layer {
  name: "loss"
  type: "SoftmaxWithLoss"
  bottom: "ip2"
  bottom: "label"
  top: "loss"
}
I1013 10:05:18.275310  1684 layer_factory.hpp:77] Creating layer mnist
I1013 10:05:18.291010  1684 net.cpp:100] Creating Layer mnist
I1013 10:05:18.291010  1684 net.cpp:418] mnist -> data
I1013 10:05:18.291010  1684 net.cpp:418] mnist -> label
I1013 10:05:18.291010  7500 common.cpp:36] System entropy source not available, using fallback algorithm to generate seed instead.
I1013 10:05:18.369072  7500 db_lmdb.cpp:40] Opened lmdb examples/mnist/mnist_test_lmdb
I1013 10:05:18.369072  1684 data_layer.cpp:41] output data size: 100,1,28,28
I1013 10:05:18.384691  1684 net.cpp:150] Setting up mnist
I1013 10:05:18.384691  1684 net.cpp:157] Top shape: 100 1 28 28 (78400)
I1013 10:05:18.384691  1684 net.cpp:157] Top shape: 100 (100)
I1013 10:05:18.384691  1684 net.cpp:165] Memory required for data: 314000
I1013 10:05:18.384691  1684 layer_factory.hpp:77] Creating layer label_mnist_1_split
I1013 10:05:18.384691  2420 common.cpp:36] System entropy source not available, using fallback algorithm to generate seed instead.
I1013 10:05:18.384691  1684 net.cpp:100] Creating Layer label_mnist_1_split
I1013 10:05:18.384691  1684 net.cpp:444] label_mnist_1_split <- label
I1013 10:05:18.400321  1684 net.cpp:418] label_mnist_1_split -> label_mnist_1_split_0
I1013 10:05:18.400321  1684 net.cpp:418] label_mnist_1_split -> label_mnist_1_split_1
I1013 10:05:18.400321  1684 net.cpp:150] Setting up label_mnist_1_split
I1013 10:05:18.400321  1684 net.cpp:157] Top shape: 100 (100)
I1013 10:05:18.400321  1684 net.cpp:157] Top shape: 100 (100)
I1013 10:05:18.400321  1684 net.cpp:165] Memory required for data: 314800
I1013 10:05:18.400321  1684 layer_factory.hpp:77] Creating layer conv1
I1013 10:05:18.400321  1684 net.cpp:100] Creating Layer conv1
I1013 10:05:18.400321  1684 net.cpp:444] conv1 <- data
I1013 10:05:18.415946  1684 net.cpp:418] conv1 -> conv1
I1013 10:05:18.415946  1684 net.cpp:150] Setting up conv1
I1013 10:05:18.415946  1684 net.cpp:157] Top shape: 100 20 24 24 (1152000)
I1013 10:05:18.415946  1684 net.cpp:165] Memory required for data: 4922800
I1013 10:05:18.415946  1684 layer_factory.hpp:77] Creating layer pool1
I1013 10:05:18.415946  1684 net.cpp:100] Creating Layer pool1
I1013 10:05:18.415946  1684 net.cpp:444] pool1 <- conv1
I1013 10:05:18.415946  1684 net.cpp:418] pool1 -> pool1
I1013 10:05:18.415946  1684 net.cpp:150] Setting up pool1
I1013 10:05:18.415946  1684 net.cpp:157] Top shape: 100 20 12 12 (288000)
I1013 10:05:18.431571  1684 net.cpp:165] Memory required for data: 6074800
I1013 10:05:18.431571  1684 layer_factory.hpp:77] Creating layer conv2
I1013 10:05:18.431571  1684 net.cpp:100] Creating Layer conv2
I1013 10:05:18.431571  1684 net.cpp:444] conv2 <- pool1
I1013 10:05:18.431571  1684 net.cpp:418] conv2 -> conv2
I1013 10:05:18.431571  1684 net.cpp:150] Setting up conv2
I1013 10:05:18.431571  1684 net.cpp:157] Top shape: 100 50 8 8 (320000)
I1013 10:05:18.431571  1684 net.cpp:165] Memory required for data: 7354800
I1013 10:05:18.431571  1684 layer_factory.hpp:77] Creating layer pool2
I1013 10:05:18.431571  1684 net.cpp:100] Creating Layer pool2
I1013 10:05:18.431571  1684 net.cpp:444] pool2 <- conv2
I1013 10:05:18.447198  1684 net.cpp:418] pool2 -> pool2
I1013 10:05:18.447198  1684 net.cpp:150] Setting up pool2
I1013 10:05:18.447198  1684 net.cpp:157] Top shape: 100 50 4 4 (80000)
I1013 10:05:18.447198  1684 net.cpp:165] Memory required for data: 7674800
I1013 10:05:18.447198  1684 layer_factory.hpp:77] Creating layer ip1
I1013 10:05:18.447198  1684 net.cpp:100] Creating Layer ip1
I1013 10:05:18.447198  1684 net.cpp:444] ip1 <- pool2
I1013 10:05:18.447198  1684 net.cpp:418] ip1 -> ip1
I1013 10:05:18.462826  1684 net.cpp:150] Setting up ip1
I1013 10:05:18.462826  1684 net.cpp:157] Top shape: 100 500 (50000)
I1013 10:05:18.462826  1684 net.cpp:165] Memory required for data: 7874800
I1013 10:05:18.462826  1684 layer_factory.hpp:77] Creating layer relu1
I1013 10:05:18.462826  1684 net.cpp:100] Creating Layer relu1
I1013 10:05:18.462826  1684 net.cpp:444] relu1 <- ip1
I1013 10:05:18.462826  1684 net.cpp:405] relu1 -> ip1 (in-place)
I1013 10:05:18.462826  1684 net.cpp:150] Setting up relu1
I1013 10:05:18.462826  1684 net.cpp:157] Top shape: 100 500 (50000)
I1013 10:05:18.462826  1684 net.cpp:165] Memory required for data: 8074800
I1013 10:05:18.462826  1684 layer_factory.hpp:77] Creating layer ip2
I1013 10:05:18.478452  1684 net.cpp:100] Creating Layer ip2
I1013 10:05:18.478452  1684 net.cpp:444] ip2 <- ip1
I1013 10:05:18.478452  1684 net.cpp:418] ip2 -> ip2
I1013 10:05:18.478452  1684 net.cpp:150] Setting up ip2
I1013 10:05:18.478452  1684 net.cpp:157] Top shape: 100 10 (1000)
I1013 10:05:18.478452  1684 net.cpp:165] Memory required for data: 8078800
I1013 10:05:18.478452  1684 layer_factory.hpp:77] Creating layer ip2_ip2_0_split
I1013 10:05:18.494081  1684 net.cpp:100] Creating Layer ip2_ip2_0_split
I1013 10:05:18.494081  1684 net.cpp:444] ip2_ip2_0_split <- ip2
I1013 10:05:18.494081  1684 net.cpp:418] ip2_ip2_0_split -> ip2_ip2_0_split_0
I1013 10:05:18.494081  1684 net.cpp:418] ip2_ip2_0_split -> ip2_ip2_0_split_1
I1013 10:05:18.494081  1684 net.cpp:150] Setting up ip2_ip2_0_split
I1013 10:05:18.494081  1684 net.cpp:157] Top shape: 100 10 (1000)
I1013 10:05:18.494081  1684 net.cpp:157] Top shape: 100 10 (1000)
I1013 10:05:18.494081  1684 net.cpp:165] Memory required for data: 8086800
I1013 10:05:18.509729  1684 layer_factory.hpp:77] Creating layer accuracy
I1013 10:05:18.509729  1684 net.cpp:100] Creating Layer accuracy
I1013 10:05:18.509729  1684 net.cpp:444] accuracy <- ip2_ip2_0_split_0
I1013 10:05:18.509729  1684 net.cpp:444] accuracy <- label_mnist_1_split_0
I1013 10:05:18.509729  1684 net.cpp:418] accuracy -> accuracy
I1013 10:05:18.509729  1684 net.cpp:150] Setting up accuracy
I1013 10:05:18.509729  1684 net.cpp:157] Top shape: (1)
I1013 10:05:18.509729  1684 net.cpp:165] Memory required for data: 8086804
I1013 10:05:18.509729  1684 layer_factory.hpp:77] Creating layer loss
I1013 10:05:18.509729  1684 net.cpp:100] Creating Layer loss
I1013 10:05:18.509729  1684 net.cpp:444] loss <- ip2_ip2_0_split_1
I1013 10:05:18.525331  1684 net.cpp:444] loss <- label_mnist_1_split_1
I1013 10:05:18.525331  1684 net.cpp:418] loss -> loss
I1013 10:05:18.525331  1684 layer_factory.hpp:77] Creating layer loss
I1013 10:05:18.525331  1684 net.cpp:150] Setting up loss
I1013 10:05:18.525331  1684 net.cpp:157] Top shape: (1)
I1013 10:05:18.525331  1684 net.cpp:160]     with loss weight 1
I1013 10:05:18.525331  1684 net.cpp:165] Memory required for data: 8086808
I1013 10:05:18.525331  1684 net.cpp:226] loss needs backward computation.
I1013 10:05:18.525331  1684 net.cpp:228] accuracy does not need backward computation.
I1013 10:05:18.525331  1684 net.cpp:226] ip2_ip2_0_split needs backward computation.
I1013 10:05:18.540958  1684 net.cpp:226] ip2 needs backward computation.
I1013 10:05:18.540958  1684 net.cpp:226] relu1 needs backward computation.
I1013 10:05:18.540958  1684 net.cpp:226] ip1 needs backward computation.
I1013 10:05:18.540958  1684 net.cpp:226] pool2 needs backward computation.
I1013 10:05:18.540958  1684 net.cpp:226] conv2 needs backward computation.
I1013 10:05:18.540958  1684 net.cpp:226] pool1 needs backward computation.
I1013 10:05:18.540958  1684 net.cpp:226] conv1 needs backward computation.
I1013 10:05:18.556589  1684 net.cpp:228] label_mnist_1_split does not need backward computation.
I1013 10:05:18.556589  1684 net.cpp:228] mnist does not need backward computation.
I1013 10:05:18.556589  1684 net.cpp:270] This network produces output accuracy
I1013 10:05:18.556589  1684 net.cpp:270] This network produces output loss
I1013 10:05:18.556589  1684 net.cpp:283] Network initialization done.
I1013 10:05:18.572244  1684 solver.cpp:60] Solver scaffolding done.
I1013 10:05:18.572244  1684 caffe.cpp:252] Starting Optimization
I1013 10:05:18.572244  1684 solver.cpp:279] Solving LeNet
I1013 10:05:18.572244  1684 solver.cpp:280] Learning Rate Policy: inv
I1013 10:05:18.572244  1684 solver.cpp:337] Iteration 0, Testing net (#0)
I1013 10:05:19.978624  1684 solver.cpp:404]     Test net output #0: accuracy = 0.0789
I1013 10:05:19.978624  1684 solver.cpp:404]     Test net output #1: loss = 2.36376 (* 1 = 2.36376 loss)
I1013 10:05:20.009863  1684 solver.cpp:228] Iteration 0, loss = 2.34559
I1013 10:05:20.009863  1684 solver.cpp:244]     Train net output #0: loss = 2.34559 (* 1 = 2.34559 loss)
I1013 10:05:20.009863  1684 sgd_solver.cpp:106] Iteration 0, lr = 0.01
I1013 10:05:22.134766  1684 solver.cpp:228] Iteration 100, loss = 0.226693
I1013 10:05:22.136801  1684 solver.cpp:244]     Train net output #0: loss = 0.226693 (* 1 = 0.226693 loss)
I1013 10:05:22.137765  1684 sgd_solver.cpp:106] Iteration 100, lr = 0.00992565
I1013 10:05:24.268718  1684 solver.cpp:228] Iteration 200, loss = 0.142792
I1013 10:05:24.270691  1684 solver.cpp:244]     Train net output #0: loss = 0.142792 (* 1 = 0.142792 loss)
I1013 10:05:24.272729  1684 sgd_solver.cpp:106] Iteration 200, lr = 0.00985258
I1013 10:05:26.396376  1684 solver.cpp:228] Iteration 300, loss = 0.192766
I1013 10:05:26.399351  1684 solver.cpp:244]     Train net output #0: loss = 0.192766 (* 1 = 0.192766 loss)
I1013 10:05:26.400354  1684 sgd_solver.cpp:106] Iteration 300, lr = 0.00978075
I1013 10:05:28.526006  1684 solver.cpp:228] Iteration 400, loss = 0.0834785
I1013 10:05:28.528012  1684 solver.cpp:244]     Train net output #0: loss = 0.0834785 (* 1 = 0.0834785 loss)
I1013 10:05:28.531019  1684 sgd_solver.cpp:106] Iteration 400, lr = 0.00971013
I1013 10:05:30.658334  1684 solver.cpp:337] Iteration 500, Testing net (#0)
I1013 10:05:32.030649  1684 solver.cpp:404]     Test net output #0: accuracy = 0.9678
I1013 10:05:32.031683  1684 solver.cpp:404]     Test net output #1: loss = 0.0990599 (* 1 = 0.0990599 loss)
I1013 10:05:32.044724  1684 solver.cpp:228] Iteration 500, loss = 0.112297
I1013 10:05:32.045688  1684 solver.cpp:244]     Train net output #0: loss = 0.112297 (* 1 = 0.112297 loss)
I1013 10:05:32.049700  1684 sgd_solver.cpp:106] Iteration 500, lr = 0.00964069
I1013 10:05:34.181881  1684 solver.cpp:228] Iteration 600, loss = 0.101184
I1013 10:05:34.182885  1684 solver.cpp:244]     Train net output #0: loss = 0.101184 (* 1 = 0.101184 loss)
I1013 10:05:34.183862  1684 sgd_solver.cpp:106] Iteration 600, lr = 0.0095724
I1013 10:05:36.311400  1684 solver.cpp:228] Iteration 700, loss = 0.179369
I1013 10:05:36.312403  1684 solver.cpp:244]     Train net output #0: loss = 0.179369 (* 1 = 0.179369 loss)
I1013 10:05:36.314407  1684 sgd_solver.cpp:106] Iteration 700, lr = 0.00950522
I1013 10:05:38.447108  1684 solver.cpp:228] Iteration 800, loss = 0.209864
I1013 10:05:38.449084  1684 solver.cpp:244]     Train net output #0: loss = 0.209864 (* 1 = 0.209864 loss)
I1013 10:05:38.450114  1684 sgd_solver.cpp:106] Iteration 800, lr = 0.00943913
I1013 10:05:40.575814  1684 solver.cpp:228] Iteration 900, loss = 0.142768
I1013 10:05:40.575814  1684 solver.cpp:244]     Train net output #0: loss = 0.142768 (* 1 = 0.142768 loss)
I1013 10:05:40.575814  1684 sgd_solver.cpp:106] Iteration 900, lr = 0.00937411
I1013 10:05:42.700186  1684 solver.cpp:337] Iteration 1000, Testing net (#0)
I1013 10:05:44.075335  1684 solver.cpp:404]     Test net output #0: accuracy = 0.9808
I1013 10:05:44.075335  1684 solver.cpp:404]     Test net output #1: loss = 0.0613375 (* 1 = 0.0613375 loss)
I1013 10:05:44.090960  1684 solver.cpp:228] Iteration 1000, loss = 0.0704594
I1013 10:05:44.090960  1684 solver.cpp:244]     Train net output #0: loss = 0.0704594 (* 1 = 0.0704594 loss)
I1013 10:05:44.090960  1684 sgd_solver.cpp:106] Iteration 1000, lr = 0.00931012
I1013 10:05:46.231811  1684 solver.cpp:228] Iteration 1100, loss = 0.00886345
I1013 10:05:46.231811  1684 solver.cpp:244]     Train net output #0: loss = 0.00886345 (* 1 = 0.00886345 loss)
I1013 10:05:46.231811  1684 sgd_solver.cpp:106] Iteration 1100, lr = 0.00924715
I1013 10:05:48.372705  1684 solver.cpp:228] Iteration 1200, loss = 0.0159409
I1013 10:05:48.372705  1684 solver.cpp:244]     Train net output #0: loss = 0.0159409 (* 1 = 0.0159409 loss)
I1013 10:05:48.372705  1684 sgd_solver.cpp:106] Iteration 1200, lr = 0.00918515
I1013 10:05:50.513516  1684 solver.cpp:228] Iteration 1300, loss = 0.0102466
I1013 10:05:50.513516  1684 solver.cpp:244]     Train net output #0: loss = 0.0102465 (* 1 = 0.0102465 loss)
I1013 10:05:50.513516  1684 sgd_solver.cpp:106] Iteration 1300, lr = 0.00912412
I1013 10:05:52.639024  1684 solver.cpp:228] Iteration 1400, loss = 0.00691616
I1013 10:05:52.639024  1684 solver.cpp:244]     Train net output #0: loss = 0.00691615 (* 1 = 0.00691615 loss)
I1013 10:05:52.639024  1684 sgd_solver.cpp:106] Iteration 1400, lr = 0.00906403
I1013 10:05:54.748378  1684 solver.cpp:337] Iteration 1500, Testing net (#0)
I1013 10:05:56.123487  1684 solver.cpp:404]     Test net output #0: accuracy = 0.9824
I1013 10:05:56.123487  1684 solver.cpp:404]     Test net output #1: loss = 0.0558028 (* 1 = 0.0558028 loss)
I1013 10:05:56.139156  1684 solver.cpp:228] Iteration 1500, loss = 0.0770894
I1013 10:05:56.139156  1684 solver.cpp:244]     Train net output #0: loss = 0.0770894 (* 1 = 0.0770894 loss)
I1013 10:05:56.139156  1684 sgd_solver.cpp:106] Iteration 1500, lr = 0.00900485
I1013 10:05:58.279999  1684 solver.cpp:228] Iteration 1600, loss = 0.08424
I1013 10:05:58.279999  1684 solver.cpp:244]     Train net output #0: loss = 0.0842399 (* 1 = 0.0842399 loss)
I1013 10:05:58.279999  1684 sgd_solver.cpp:106] Iteration 1600, lr = 0.00894657
I1013 10:06:00.405194  1684 solver.cpp:228] Iteration 1700, loss = 0.0452077
I1013 10:06:00.405194  1684 solver.cpp:244]     Train net output #0: loss = 0.0452077 (* 1 = 0.0452077 loss)
I1013 10:06:00.405194  1684 sgd_solver.cpp:106] Iteration 1700, lr = 0.00888916
I1013 10:06:02.546080  1684 solver.cpp:228] Iteration 1800, loss = 0.0248114
I1013 10:06:02.546080  1684 solver.cpp:244]     Train net output #0: loss = 0.0248114 (* 1 = 0.0248114 loss)
I1013 10:06:02.546080  1684 sgd_solver.cpp:106] Iteration 1800, lr = 0.0088326
I1013 10:06:04.671310  1684 solver.cpp:228] Iteration 1900, loss = 0.114547
I1013 10:06:04.671310  1684 solver.cpp:244]     Train net output #0: loss = 0.114547 (* 1 = 0.114547 loss)
I1013 10:06:04.686897  1684 sgd_solver.cpp:106] Iteration 1900, lr = 0.00877687
I1013 10:06:06.796535  1684 solver.cpp:337] Iteration 2000, Testing net (#0)
I1013 10:06:08.171643  1684 solver.cpp:404]     Test net output #0: accuracy = 0.9841
I1013 10:06:08.187270  1684 solver.cpp:404]     Test net output #1: loss = 0.0490052 (* 1 = 0.0490052 loss)
I1013 10:06:08.187270  1684 solver.cpp:228] Iteration 2000, loss = 0.00911095
I1013 10:06:08.202911  1684 solver.cpp:244]     Train net output #0: loss = 0.0091109 (* 1 = 0.0091109 loss)
I1013 10:06:08.202911  1684 sgd_solver.cpp:106] Iteration 2000, lr = 0.00872196
I1013 10:06:10.328163  1684 solver.cpp:228] Iteration 2100, loss = 0.0175512
I1013 10:06:10.328163  1684 solver.cpp:244]     Train net output #0: loss = 0.0175512 (* 1 = 0.0175512 loss)
I1013 10:06:10.328163  1684 sgd_solver.cpp:106] Iteration 2100, lr = 0.00866784
I1013 10:06:12.456619  1684 solver.cpp:228] Iteration 2200, loss = 0.0182508
I1013 10:06:12.456619  1684 solver.cpp:244]     Train net output #0: loss = 0.0182508 (* 1 = 0.0182508 loss)
I1013 10:06:12.472260  1684 sgd_solver.cpp:106] Iteration 2200, lr = 0.0086145
I1013 10:06:14.597468  1684 solver.cpp:228] Iteration 2300, loss = 0.0929874
I1013 10:06:14.597468  1684 solver.cpp:244]     Train net output #0: loss = 0.0929874 (* 1 = 0.0929874 loss)
I1013 10:06:14.597468  1684 sgd_solver.cpp:106] Iteration 2300, lr = 0.00856192
I1013 10:06:16.738363  1684 solver.cpp:228] Iteration 2400, loss = 0.0156817
I1013 10:06:16.738363  1684 solver.cpp:244]     Train net output #0: loss = 0.0156816 (* 1 = 0.0156816 loss)
I1013 10:06:16.738363  1684 sgd_solver.cpp:106] Iteration 2400, lr = 0.00851008
I1013 10:06:18.847921  1684 solver.cpp:337] Iteration 2500, Testing net (#0)
I1013 10:06:20.223072  1684 solver.cpp:404]     Test net output #0: accuracy = 0.9853
I1013 10:06:20.223072  1684 solver.cpp:404]     Test net output #1: loss = 0.0476141 (* 1 = 0.0476141 loss)
I1013 10:06:20.238706  1684 solver.cpp:228] Iteration 2500, loss = 0.0254326
I1013 10:06:20.238706  1684 solver.cpp:244]     Train net output #0: loss = 0.0254326 (* 1 = 0.0254326 loss)
I1013 10:06:20.238706  1684 sgd_solver.cpp:106] Iteration 2500, lr = 0.00845897
I1013 10:06:22.379546  1684 solver.cpp:228] Iteration 2600, loss = 0.0614191
I1013 10:06:22.379546  1684 solver.cpp:244]     Train net output #0: loss = 0.061419 (* 1 = 0.061419 loss)
I1013 10:06:22.379546  1684 sgd_solver.cpp:106] Iteration 2600, lr = 0.00840857
I1013 10:06:24.520401  1684 solver.cpp:228] Iteration 2700, loss = 0.0625541
I1013 10:06:24.520401  1684 solver.cpp:244]     Train net output #0: loss = 0.062554 (* 1 = 0.062554 loss)
I1013 10:06:24.520401  1684 sgd_solver.cpp:106] Iteration 2700, lr = 0.00835886
I1013 10:06:26.645644  1684 solver.cpp:228] Iteration 2800, loss = 0.00305949
I1013 10:06:26.645644  1684 solver.cpp:244]     Train net output #0: loss = 0.00305946 (* 1 = 0.00305946 loss)
I1013 10:06:26.645644  1684 sgd_solver.cpp:106] Iteration 2800, lr = 0.00830984
I1013 10:06:28.786510  1684 solver.cpp:228] Iteration 2900, loss = 0.0252702
I1013 10:06:28.786510  1684 solver.cpp:244]     Train net output #0: loss = 0.0252702 (* 1 = 0.0252702 loss)
I1013 10:06:28.786510  1684 sgd_solver.cpp:106] Iteration 2900, lr = 0.00826148
I1013 10:06:30.896109  1684 solver.cpp:337] Iteration 3000, Testing net (#0)
I1013 10:06:32.271224  1684 solver.cpp:404]     Test net output #0: accuracy = 0.9861
I1013 10:06:32.271224  1684 solver.cpp:404]     Test net output #1: loss = 0.0419692 (* 1 = 0.0419692 loss)
I1013 10:06:32.286850  1684 solver.cpp:228] Iteration 3000, loss = 0.00504212
I1013 10:06:32.286850  1684 solver.cpp:244]     Train net output #0: loss = 0.00504212 (* 1 = 0.00504212 loss)
I1013 10:06:32.286850  1684 sgd_solver.cpp:106] Iteration 3000, lr = 0.00821377
I1013 10:06:34.412075  1684 solver.cpp:228] Iteration 3100, loss = 0.0165952
I1013 10:06:34.412075  1684 solver.cpp:244]     Train net output #0: loss = 0.0165953 (* 1 = 0.0165953 loss)
I1013 10:06:34.427702  1684 sgd_solver.cpp:106] Iteration 3100, lr = 0.0081667
I1013 10:06:36.552963  1684 solver.cpp:228] Iteration 3200, loss = 0.0144548
I1013 10:06:36.552963  1684 solver.cpp:244]     Train net output #0: loss = 0.0144548 (* 1 = 0.0144548 loss)
I1013 10:06:36.552963  1684 sgd_solver.cpp:106] Iteration 3200, lr = 0.00812025
I1013 10:06:38.693781  1684 solver.cpp:228] Iteration 3300, loss = 0.0481921
I1013 10:06:38.693781  1684 solver.cpp:244]     Train net output #0: loss = 0.0481921 (* 1 = 0.0481921 loss)
I1013 10:06:38.693781  1684 sgd_solver.cpp:106] Iteration 3300, lr = 0.00807442
I1013 10:06:40.834671  1684 solver.cpp:228] Iteration 3400, loss = 0.0168258
I1013 10:06:40.834671  1684 solver.cpp:244]     Train net output #0: loss = 0.0168259 (* 1 = 0.0168259 loss)
I1013 10:06:40.834671  1684 sgd_solver.cpp:106] Iteration 3400, lr = 0.00802918
I1013 10:06:42.944270  1684 solver.cpp:337] Iteration 3500, Testing net (#0)
I1013 10:06:44.303789  1684 solver.cpp:404]     Test net output #0: accuracy = 0.9863
I1013 10:06:44.319380  1684 solver.cpp:404]     Test net output #1: loss = 0.043148 (* 1 = 0.043148 loss)
I1013 10:06:44.335008  1684 solver.cpp:228] Iteration 3500, loss = 0.00682415
I1013 10:06:44.335008  1684 solver.cpp:244]     Train net output #0: loss = 0.00682418 (* 1 = 0.00682418 loss)
I1013 10:06:44.335008  1684 sgd_solver.cpp:106] Iteration 3500, lr = 0.00798454
I1013 10:06:46.460263  1684 solver.cpp:228] Iteration 3600, loss = 0.0317525
I1013 10:06:46.460263  1684 solver.cpp:244]     Train net output #0: loss = 0.0317526 (* 1 = 0.0317526 loss)
I1013 10:06:46.460263  1684 sgd_solver.cpp:106] Iteration 3600, lr = 0.00794046
I1013 10:06:48.601121  1684 solver.cpp:228] Iteration 3700, loss = 0.0246315
I1013 10:06:48.601121  1684 solver.cpp:244]     Train net output #0: loss = 0.0246315 (* 1 = 0.0246315 loss)
I1013 10:06:48.601121  1684 sgd_solver.cpp:106] Iteration 3700, lr = 0.00789695
I1013 10:06:50.726347  1684 solver.cpp:228] Iteration 3800, loss = 0.00837651
I1013 10:06:50.726347  1684 solver.cpp:244]     Train net output #0: loss = 0.00837653 (* 1 = 0.00837653 loss)
I1013 10:06:50.726347  1684 sgd_solver.cpp:106] Iteration 3800, lr = 0.007854
I1013 10:06:52.871928  1684 solver.cpp:228] Iteration 3900, loss = 0.0320845
I1013 10:06:52.874935  1684 solver.cpp:244]     Train net output #0: loss = 0.0320845 (* 1 = 0.0320845 loss)
I1013 10:06:52.876941  1684 sgd_solver.cpp:106] Iteration 3900, lr = 0.00781158
I1013 10:06:54.979713  1684 solver.cpp:337] Iteration 4000, Testing net (#0)
I1013 10:06:56.354836  1684 solver.cpp:404]     Test net output #0: accuracy = 0.9875
I1013 10:06:56.354836  1684 solver.cpp:404]     Test net output #1: loss = 0.0353671 (* 1 = 0.0353671 loss)
I1013 10:06:56.370452  1684 solver.cpp:228] Iteration 4000, loss = 0.0140691
I1013 10:06:56.370452  1684 solver.cpp:244]     Train net output #0: loss = 0.0140691 (* 1 = 0.0140691 loss)
I1013 10:06:56.370452  1684 sgd_solver.cpp:106] Iteration 4000, lr = 0.00776969
I1013 10:06:58.511303  1684 solver.cpp:228] Iteration 4100, loss = 0.0263123
I1013 10:06:58.511303  1684 solver.cpp:244]     Train net output #0: loss = 0.0263123 (* 1 = 0.0263123 loss)
I1013 10:06:58.511303  1684 sgd_solver.cpp:106] Iteration 4100, lr = 0.00772833
I1013 10:07:00.652200  1684 solver.cpp:228] Iteration 4200, loss = 0.0117368
I1013 10:07:00.652200  1684 solver.cpp:244]     Train net output #0: loss = 0.0117368 (* 1 = 0.0117368 loss)
I1013 10:07:00.652200  1684 sgd_solver.cpp:106] Iteration 4200, lr = 0.00768748
I1013 10:07:02.793052  1684 solver.cpp:228] Iteration 4300, loss = 0.0490961
I1013 10:07:02.793052  1684 solver.cpp:244]     Train net output #0: loss = 0.0490961 (* 1 = 0.0490961 loss)
I1013 10:07:02.793052  1684 sgd_solver.cpp:106] Iteration 4300, lr = 0.00764712
I1013 10:07:04.933894  1684 solver.cpp:228] Iteration 4400, loss = 0.0143547
I1013 10:07:04.933894  1684 solver.cpp:244]     Train net output #0: loss = 0.0143547 (* 1 = 0.0143547 loss)
I1013 10:07:04.933894  1684 sgd_solver.cpp:106] Iteration 4400, lr = 0.00760726
I1013 10:07:07.043498  1684 solver.cpp:337] Iteration 4500, Testing net (#0)
I1013 10:07:08.418617  1684 solver.cpp:404]     Test net output #0: accuracy = 0.9875
I1013 10:07:08.418617  1684 solver.cpp:404]     Test net output #1: loss = 0.039773 (* 1 = 0.039773 loss)
I1013 10:07:08.434267  1684 solver.cpp:228] Iteration 4500, loss = 0.00660795
I1013 10:07:08.434267  1684 solver.cpp:244]     Train net output #0: loss = 0.00660791 (* 1 = 0.00660791 loss)
I1013 10:07:08.434267  1684 sgd_solver.cpp:106] Iteration 4500, lr = 0.00756788
I1013 10:07:10.575119  1684 solver.cpp:228] Iteration 4600, loss = 0.0135348
I1013 10:07:10.575119  1684 solver.cpp:244]     Train net output #0: loss = 0.0135347 (* 1 = 0.0135347 loss)
I1013 10:07:10.575119  1684 sgd_solver.cpp:106] Iteration 4600, lr = 0.00752897
I1013 10:07:12.715939  1684 solver.cpp:228] Iteration 4700, loss = 0.00858051
I1013 10:07:12.715939  1684 solver.cpp:244]     Train net output #0: loss = 0.00858048 (* 1 = 0.00858048 loss)
I1013 10:07:12.715939  1684 sgd_solver.cpp:106] Iteration 4700, lr = 0.00749052
I1013 10:07:14.856828  1684 solver.cpp:228] Iteration 4800, loss = 0.013837
I1013 10:07:14.856828  1684 solver.cpp:244]     Train net output #0: loss = 0.013837 (* 1 = 0.013837 loss)
I1013 10:07:14.856828  1684 sgd_solver.cpp:106] Iteration 4800, lr = 0.00745253
I1013 10:07:16.997676  1684 solver.cpp:228] Iteration 4900, loss = 0.00716435
I1013 10:07:16.997676  1684 solver.cpp:244]     Train net output #0: loss = 0.00716432 (* 1 = 0.00716432 loss)
I1013 10:07:16.997676  1684 sgd_solver.cpp:106] Iteration 4900, lr = 0.00741498
I1013 10:07:19.107244  1684 solver.cpp:454] Snapshotting to binary proto file examples/mnist/lenet_iter_5000.caffemodel
I1013 10:07:19.138531  1684 sgd_solver.cpp:273] Snapshotting solver state to binary proto file examples/mnist/lenet_iter_5000.solverstate
I1013 10:07:19.154156  1684 solver.cpp:337] Iteration 5000, Testing net (#0)
I1013 10:07:20.529271  1684 solver.cpp:404]     Test net output #0: accuracy = 0.9886
I1013 10:07:20.529271  1684 solver.cpp:404]     Test net output #1: loss = 0.0343976 (* 1 = 0.0343976 loss)
I1013 10:07:20.544936  1684 solver.cpp:228] Iteration 5000, loss = 0.046033
I1013 10:07:20.544936  1684 solver.cpp:244]     Train net output #0: loss = 0.0460329 (* 1 = 0.0460329 loss)
I1013 10:07:20.544936  1684 sgd_solver.cpp:106] Iteration 5000, lr = 0.00737788
I1013 10:07:22.670121  1684 solver.cpp:228] Iteration 5100, loss = 0.0231957
I1013 10:07:22.670121  1684 solver.cpp:244]     Train net output #0: loss = 0.0231957 (* 1 = 0.0231957 loss)
I1013 10:07:22.670121  1684 sgd_solver.cpp:106] Iteration 5100, lr = 0.0073412
I1013 10:07:24.810972  1684 solver.cpp:228] Iteration 5200, loss = 0.00935967
I1013 10:07:24.810972  1684 solver.cpp:244]     Train net output #0: loss = 0.00935963 (* 1 = 0.00935963 loss)
I1013 10:07:24.826604  1684 sgd_solver.cpp:106] Iteration 5200, lr = 0.00730495
I1013 10:07:26.951828  1684 solver.cpp:228] Iteration 5300, loss = 0.00283169
I1013 10:07:26.951828  1684 solver.cpp:244]     Train net output #0: loss = 0.00283165 (* 1 = 0.00283165 loss)
I1013 10:07:26.951828  1684 sgd_solver.cpp:106] Iteration 5300, lr = 0.00726911
I1013 10:07:29.092718  1684 solver.cpp:228] Iteration 5400, loss = 0.00842249
I1013 10:07:29.092718  1684 solver.cpp:244]     Train net output #0: loss = 0.00842245 (* 1 = 0.00842245 loss)
I1013 10:07:29.092718  1684 sgd_solver.cpp:106] Iteration 5400, lr = 0.00723368
I1013 10:07:31.202320  1684 solver.cpp:337] Iteration 5500, Testing net (#0)
I1013 10:07:32.577424  1684 solver.cpp:404]     Test net output #0: accuracy = 0.9883
I1013 10:07:32.577424  1684 solver.cpp:404]     Test net output #1: loss = 0.0350875 (* 1 = 0.0350875 loss)
I1013 10:07:32.593072  1684 solver.cpp:228] Iteration 5500, loss = 0.00971781
I1013 10:07:32.593072  1684 solver.cpp:244]     Train net output #0: loss = 0.00971777 (* 1 = 0.00971777 loss)
I1013 10:07:32.593072  1684 sgd_solver.cpp:106] Iteration 5500, lr = 0.00719865
I1013 10:07:34.733940  1684 solver.cpp:228] Iteration 5600, loss = 0.000905203
I1013 10:07:34.733940  1684 solver.cpp:244]     Train net output #0: loss = 0.000905167 (* 1 = 0.000905167 loss)
I1013 10:07:34.733940  1684 sgd_solver.cpp:106] Iteration 5600, lr = 0.00716402
I1013 10:07:36.874794  1684 solver.cpp:228] Iteration 5700, loss = 0.00458089
I1013 10:07:36.874794  1684 solver.cpp:244]     Train net output #0: loss = 0.00458086 (* 1 = 0.00458086 loss)
I1013 10:07:36.874794  1684 sgd_solver.cpp:106] Iteration 5700, lr = 0.00712977
I1013 10:07:39.000007  1684 solver.cpp:228] Iteration 5800, loss = 0.0429197
I1013 10:07:39.015626  1684 solver.cpp:244]     Train net output #0: loss = 0.0429196 (* 1 = 0.0429196 loss)
I1013 10:07:39.015626  1684 sgd_solver.cpp:106] Iteration 5800, lr = 0.0070959
I1013 10:07:41.140871  1684 solver.cpp:228] Iteration 5900, loss = 0.00847424
I1013 10:07:41.140871  1684 solver.cpp:244]     Train net output #0: loss = 0.0084742 (* 1 = 0.0084742 loss)
I1013 10:07:41.140871  1684 sgd_solver.cpp:106] Iteration 5900, lr = 0.0070624
I1013 10:07:43.280727  1684 solver.cpp:337] Iteration 6000, Testing net (#0)
I1013 10:07:44.657387  1684 solver.cpp:404]     Test net output #0: accuracy = 0.9892
I1013 10:07:44.658390  1684 solver.cpp:404]     Test net output #1: loss = 0.0333308 (* 1 = 0.0333308 loss)
I1013 10:07:44.672427  1684 solver.cpp:228] Iteration 6000, loss = 0.00297941
I1013 10:07:44.673431  1684 solver.cpp:244]     Train net output #0: loss = 0.00297938 (* 1 = 0.00297938 loss)
I1013 10:07:44.675467  1684 sgd_solver.cpp:106] Iteration 6000, lr = 0.00702927
I1013 10:07:46.812116  1684 solver.cpp:228] Iteration 6100, loss = 0.00404553
I1013 10:07:46.814121  1684 solver.cpp:244]     Train net output #0: loss = 0.0040455 (* 1 = 0.0040455 loss)
I1013 10:07:46.815125  1684 sgd_solver.cpp:106] Iteration 6100, lr = 0.0069965
I1013 10:07:48.949837  1684 solver.cpp:228] Iteration 6200, loss = 0.00796121
I1013 10:07:48.951807  1684 solver.cpp:244]     Train net output #0: loss = 0.00796118 (* 1 = 0.00796118 loss)
I1013 10:07:48.953860  1684 sgd_solver.cpp:106] Iteration 6200, lr = 0.00696408
I1013 10:07:51.083505  1684 solver.cpp:228] Iteration 6300, loss = 0.00927992
I1013 10:07:51.085481  1684 solver.cpp:244]     Train net output #0: loss = 0.0092799 (* 1 = 0.0092799 loss)
I1013 10:07:51.086510  1684 sgd_solver.cpp:106] Iteration 6300, lr = 0.00693201
I1013 10:07:53.220190  1684 solver.cpp:228] Iteration 6400, loss = 0.00616177
I1013 10:07:53.222162  1684 solver.cpp:244]     Train net output #0: loss = 0.00616174 (* 1 = 0.00616174 loss)
I1013 10:07:53.224195  1684 sgd_solver.cpp:106] Iteration 6400, lr = 0.00690029
I1013 10:07:55.335819  1684 solver.cpp:337] Iteration 6500, Testing net (#0)
I1013 10:07:56.705461  1684 solver.cpp:404]     Test net output #0: accuracy = 0.9892
I1013 10:07:56.707448  1684 solver.cpp:404]     Test net output #1: loss = 0.0342351 (* 1 = 0.0342351 loss)
I1013 10:07:56.721467  1684 solver.cpp:228] Iteration 6500, loss = 0.00857477
I1013 10:07:56.722470  1684 solver.cpp:244]     Train net output #0: loss = 0.00857473 (* 1 = 0.00857473 loss)
I1013 10:07:56.723474  1684 sgd_solver.cpp:106] Iteration 6500, lr = 0.0068689
I1013 10:07:58.860191  1684 solver.cpp:228] Iteration 6600, loss = 0.0264124
I1013 10:07:58.861191  1684 solver.cpp:244]     Train net output #0: loss = 0.0264124 (* 1 = 0.0264124 loss)
I1013 10:07:58.863162  1684 sgd_solver.cpp:106] Iteration 6600, lr = 0.00683784
I1013 10:08:00.991823  1684 solver.cpp:228] Iteration 6700, loss = 0.00683724
I1013 10:08:00.993829  1684 solver.cpp:244]     Train net output #0: loss = 0.00683721 (* 1 = 0.00683721 loss)
I1013 10:08:00.995842  1684 sgd_solver.cpp:106] Iteration 6700, lr = 0.00680711
I1013 10:08:03.131726  1684 solver.cpp:228] Iteration 6800, loss = 0.00408112
I1013 10:08:03.133730  1684 solver.cpp:244]     Train net output #0: loss = 0.0040811 (* 1 = 0.0040811 loss)
I1013 10:08:03.135735  1684 sgd_solver.cpp:106] Iteration 6800, lr = 0.0067767
I1013 10:08:05.266402  1684 solver.cpp:228] Iteration 6900, loss = 0.00522403
I1013 10:08:05.268406  1684 solver.cpp:244]     Train net output #0: loss = 0.00522401 (* 1 = 0.00522401 loss)
I1013 10:08:05.269409  1684 sgd_solver.cpp:106] Iteration 6900, lr = 0.0067466
I1013 10:08:07.395082  1684 solver.cpp:337] Iteration 7000, Testing net (#0)
I1013 10:08:08.769718  1684 solver.cpp:404]     Test net output #0: accuracy = 0.9888
I1013 10:08:08.771723  1684 solver.cpp:404]     Test net output #1: loss = 0.035025 (* 1 = 0.035025 loss)
I1013 10:08:08.784770  1684 solver.cpp:228] Iteration 7000, loss = 0.00657448
I1013 10:08:08.785768  1684 solver.cpp:244]     Train net output #0: loss = 0.00657445 (* 1 = 0.00657445 loss)
I1013 10:08:08.787765  1684 sgd_solver.cpp:106] Iteration 7000, lr = 0.00671681
I1013 10:08:10.924448  1684 solver.cpp:228] Iteration 7100, loss = 0.0121463
I1013 10:08:10.926453  1684 solver.cpp:244]     Train net output #0: loss = 0.0121463 (* 1 = 0.0121463 loss)
I1013 10:08:10.928458  1684 sgd_solver.cpp:106] Iteration 7100, lr = 0.00668733
I1013 10:08:13.061159  1684 solver.cpp:228] Iteration 7200, loss = 0.00267776
I1013 10:08:13.063134  1684 solver.cpp:244]     Train net output #0: loss = 0.00267773 (* 1 = 0.00267773 loss)
I1013 10:08:13.064137  1684 sgd_solver.cpp:106] Iteration 7200, lr = 0.00665815
I1013 10:08:15.199861  1684 solver.cpp:228] Iteration 7300, loss = 0.0185436
I1013 10:08:15.201831  1684 solver.cpp:244]     Train net output #0: loss = 0.0185435 (* 1 = 0.0185435 loss)
I1013 10:08:15.203866  1684 sgd_solver.cpp:106] Iteration 7300, lr = 0.00662927
I1013 10:08:17.338510  1684 solver.cpp:228] Iteration 7400, loss = 0.0036527
I1013 10:08:17.341522  1684 solver.cpp:244]     Train net output #0: loss = 0.00365268 (* 1 = 0.00365268 loss)
I1013 10:08:17.343523  1684 sgd_solver.cpp:106] Iteration 7400, lr = 0.00660067
I1013 10:08:19.459148  1684 solver.cpp:337] Iteration 7500, Testing net (#0)
I1013 10:08:20.831836  1684 solver.cpp:404]     Test net output #0: accuracy = 0.9892
I1013 10:08:20.832801  1684 solver.cpp:404]     Test net output #1: loss = 0.0364178 (* 1 = 0.0364178 loss)
I1013 10:08:20.845861  1684 solver.cpp:228] Iteration 7500, loss = 0.00223585
I1013 10:08:20.846865  1684 solver.cpp:244]     Train net output #0: loss = 0.00223582 (* 1 = 0.00223582 loss)
I1013 10:08:20.847841  1684 sgd_solver.cpp:106] Iteration 7500, lr = 0.00657236
I1013 10:08:22.981528  1684 solver.cpp:228] Iteration 7600, loss = 0.00394381
I1013 10:08:22.983520  1684 solver.cpp:244]     Train net output #0: loss = 0.00394378 (* 1 = 0.00394378 loss)
I1013 10:08:22.985527  1684 sgd_solver.cpp:106] Iteration 7600, lr = 0.00654433
I1013 10:08:25.115223  1684 solver.cpp:228] Iteration 7700, loss = 0.0196834
I1013 10:08:25.117230  1684 solver.cpp:244]     Train net output #0: loss = 0.0196834 (* 1 = 0.0196834 loss)
I1013 10:08:25.118197  1684 sgd_solver.cpp:106] Iteration 7700, lr = 0.00651658
I1013 10:08:27.252872  1684 solver.cpp:228] Iteration 7800, loss = 0.00327404
I1013 10:08:27.254878  1684 solver.cpp:244]     Train net output #0: loss = 0.00327401 (* 1 = 0.00327401 loss)
I1013 10:08:27.255897  1684 sgd_solver.cpp:106] Iteration 7800, lr = 0.00648911
I1013 10:08:29.388586  1684 solver.cpp:228] Iteration 7900, loss = 0.00185404
I1013 10:08:29.390593  1684 solver.cpp:244]     Train net output #0: loss = 0.001854 (* 1 = 0.001854 loss)
I1013 10:08:29.392597  1684 sgd_solver.cpp:106] Iteration 7900, lr = 0.0064619
I1013 10:08:31.501201  1684 solver.cpp:337] Iteration 8000, Testing net (#0)
I1013 10:08:32.873819  1684 solver.cpp:404]     Test net output #0: accuracy = 0.9892
I1013 10:08:32.875864  1684 solver.cpp:404]     Test net output #1: loss = 0.0335527 (* 1 = 0.0335527 loss)
I1013 10:08:32.889863  1684 solver.cpp:228] Iteration 8000, loss = 0.00614705
I1013 10:08:32.889863  1684 solver.cpp:244]     Train net output #0: loss = 0.00614701 (* 1 = 0.00614701 loss)
I1013 10:08:32.890882  1684 sgd_solver.cpp:106] Iteration 8000, lr = 0.00643496
I1013 10:08:35.023572  1684 solver.cpp:228] Iteration 8100, loss = 0.0192059
I1013 10:08:35.024581  1684 solver.cpp:244]     Train net output #0: loss = 0.0192059 (* 1 = 0.0192059 loss)
I1013 10:08:35.025543  1684 sgd_solver.cpp:106] Iteration 8100, lr = 0.00640827
I1013 10:08:37.159255  1684 solver.cpp:228] Iteration 8200, loss = 0.00787218
I1013 10:08:37.161262  1684 solver.cpp:244]     Train net output #0: loss = 0.00787215 (* 1 = 0.00787215 loss)
I1013 10:08:37.162261  1684 sgd_solver.cpp:106] Iteration 8200, lr = 0.00638185
I1013 10:08:39.295898  1684 solver.cpp:228] Iteration 8300, loss = 0.0265738
I1013 10:08:39.296929  1684 solver.cpp:244]     Train net output #0: loss = 0.0265737 (* 1 = 0.0265737 loss)
I1013 10:08:39.299909  1684 sgd_solver.cpp:106] Iteration 8300, lr = 0.00635568
I1013 10:08:41.430843  1684 solver.cpp:228] Iteration 8400, loss = 0.00670668
I1013 10:08:41.434828  1684 solver.cpp:244]     Train net output #0: loss = 0.00670665 (* 1 = 0.00670665 loss)
I1013 10:08:41.436877  1684 sgd_solver.cpp:106] Iteration 8400, lr = 0.00632975
I1013 10:08:43.540426  1684 solver.cpp:337] Iteration 8500, Testing net (#0)
I1013 10:08:44.918090  1684 solver.cpp:404]     Test net output #0: accuracy = 0.99
I1013 10:08:44.919092  1684 solver.cpp:404]     Test net output #1: loss = 0.0330528 (* 1 = 0.0330528 loss)
I1013 10:08:44.933130  1684 solver.cpp:228] Iteration 8500, loss = 0.00646596
I1013 10:08:44.934134  1684 solver.cpp:244]     Train net output #0: loss = 0.00646593 (* 1 = 0.00646593 loss)
I1013 10:08:44.936137  1684 sgd_solver.cpp:106] Iteration 8500, lr = 0.00630407
I1013 10:08:47.070852  1684 solver.cpp:228] Iteration 8600, loss = 0.000641635
I1013 10:08:47.072856  1684 solver.cpp:244]     Train net output #0: loss = 0.000641601 (* 1 = 0.000641601 loss)
I1013 10:08:47.074916  1684 sgd_solver.cpp:106] Iteration 8600, lr = 0.00627864
I1013 10:08:49.204524  1684 solver.cpp:228] Iteration 8700, loss = 0.00248919
I1013 10:08:49.206532  1684 solver.cpp:244]     Train net output #0: loss = 0.00248916 (* 1 = 0.00248916 loss)
I1013 10:08:49.207542  1684 sgd_solver.cpp:106] Iteration 8700, lr = 0.00625344
I1013 10:08:51.339238  1684 solver.cpp:228] Iteration 8800, loss = 0.00115433
I1013 10:08:51.341305  1684 solver.cpp:244]     Train net output #0: loss = 0.0011543 (* 1 = 0.0011543 loss)
I1013 10:08:51.343212  1684 sgd_solver.cpp:106] Iteration 8800, lr = 0.00622847
I1013 10:08:53.474915  1684 solver.cpp:228] Iteration 8900, loss = 0.00148415
I1013 10:08:53.475916  1684 solver.cpp:244]     Train net output #0: loss = 0.00148413 (* 1 = 0.00148413 loss)
I1013 10:08:53.477922  1684 sgd_solver.cpp:106] Iteration 8900, lr = 0.00620374
I1013 10:08:55.589537  1684 solver.cpp:337] Iteration 9000, Testing net (#0)
I1013 10:08:56.966166  1684 solver.cpp:404]     Test net output #0: accuracy = 0.9886
I1013 10:08:56.968189  1684 solver.cpp:404]     Test net output #1: loss = 0.0339002 (* 1 = 0.0339002 loss)
I1013 10:08:56.981241  1684 solver.cpp:228] Iteration 9000, loss = 0.0147503
I1013 10:08:56.982228  1684 solver.cpp:244]     Train net output #0: loss = 0.0147502 (* 1 = 0.0147502 loss)
I1013 10:08:56.983245  1684 sgd_solver.cpp:106] Iteration 9000, lr = 0.00617924
I1013 10:08:59.115936  1684 solver.cpp:228] Iteration 9100, loss = 0.00737076
I1013 10:08:59.117923  1684 solver.cpp:244]     Train net output #0: loss = 0.00737073 (* 1 = 0.00737073 loss)
I1013 10:08:59.119928  1684 sgd_solver.cpp:106] Iteration 9100, lr = 0.00615496
I1013 10:09:01.251560  1684 solver.cpp:228] Iteration 9200, loss = 0.00446405
I1013 10:09:01.252562  1684 solver.cpp:244]     Train net output #0: loss = 0.00446402 (* 1 = 0.00446402 loss)
I1013 10:09:01.253566  1684 sgd_solver.cpp:106] Iteration 9200, lr = 0.0061309
I1013 10:09:03.386270  1684 solver.cpp:228] Iteration 9300, loss = 0.00824475
I1013 10:09:03.388242  1684 solver.cpp:244]     Train net output #0: loss = 0.00824472 (* 1 = 0.00824472 loss)
I1013 10:09:03.389245  1684 sgd_solver.cpp:106] Iteration 9300, lr = 0.00610706
I1013 10:09:05.521981  1684 solver.cpp:228] Iteration 9400, loss = 0.0200841
I1013 10:09:05.523952  1684 solver.cpp:244]     Train net output #0: loss = 0.020084 (* 1 = 0.020084 loss)
I1013 10:09:05.525956  1684 sgd_solver.cpp:106] Iteration 9400, lr = 0.00608343
I1013 10:09:07.638610  1684 solver.cpp:337] Iteration 9500, Testing net (#0)
I1013 10:09:09.004231  1684 solver.cpp:404]     Test net output #0: accuracy = 0.9877
I1013 10:09:09.007218  1684 solver.cpp:404]     Test net output #1: loss = 0.0394568 (* 1 = 0.0394568 loss)
I1013 10:09:09.021282  1684 solver.cpp:228] Iteration 9500, loss = 0.00323504
I1013 10:09:09.021282  1684 solver.cpp:244]     Train net output #0: loss = 0.00323501 (* 1 = 0.00323501 loss)
I1013 10:09:09.023258  1684 sgd_solver.cpp:106] Iteration 9500, lr = 0.00606002
I1013 10:09:11.161974  1684 solver.cpp:228] Iteration 9600, loss = 0.00335854
I1013 10:09:11.163949  1684 solver.cpp:244]     Train net output #0: loss = 0.00335851 (* 1 = 0.00335851 loss)
I1013 10:09:11.164966  1684 sgd_solver.cpp:106] Iteration 9600, lr = 0.00603682
I1013 10:09:13.296618  1684 solver.cpp:228] Iteration 9700, loss = 0.0024854
I1013 10:09:13.298666  1684 solver.cpp:244]     Train net output #0: loss = 0.00248537 (* 1 = 0.00248537 loss)
I1013 10:09:13.300631  1684 sgd_solver.cpp:106] Iteration 9700, lr = 0.00601382
I1013 10:09:15.431372  1684 solver.cpp:228] Iteration 9800, loss = 0.0139184
I1013 10:09:15.433302  1684 solver.cpp:244]     Train net output #0: loss = 0.0139184 (* 1 = 0.0139184 loss)
I1013 10:09:15.435307  1684 sgd_solver.cpp:106] Iteration 9800, lr = 0.00599102
I1013 10:09:17.568011  1684 solver.cpp:228] Iteration 9900, loss = 0.00603178
I1013 10:09:17.569984  1684 solver.cpp:244]     Train net output #0: loss = 0.00603175 (* 1 = 0.00603175 loss)
I1013 10:09:17.570989  1684 sgd_solver.cpp:106] Iteration 9900, lr = 0.00596843
I1013 10:09:19.683639  1684 solver.cpp:454] Snapshotting to binary proto file examples/mnist/lenet_iter_10000.caffemodel
I1013 10:09:19.745410  1684 sgd_solver.cpp:273] Snapshotting solver state to binary proto file examples/mnist/lenet_iter_10000.solverstate
I1013 10:09:19.767470  1684 solver.cpp:317] Iteration 10000, loss = 0.00483315
I1013 10:09:19.768472  1684 solver.cpp:337] Iteration 10000, Testing net (#0)
I1013 10:09:21.132349  1684 solver.cpp:404]     Test net output #0: accuracy = 0.9899
I1013 10:09:21.134388  1684 solver.cpp:404]     Test net output #1: loss = 0.0316015 (* 1 = 0.0316015 loss)
I1013 10:09:21.136361  1684 solver.cpp:322] Optimization Done.
I1013 10:09:21.137379  1684 caffe.cpp:255] Optimization Done.