Caffe 下 示例程序 mnist 日志输出

共分3个部分：

1. Caffe命令 usage

root@ip-172-30-0-251:/caffe# 
root@ip-172-30-0-251:/caffe# build/tools/caffe 
caffe: command line brew
usage: caffe <command> <args>


commands:
  train           train or finetune a model
  test            score a model
  device_query    show GPU diagnostic information
  time            benchmark model execution time


  Flags from tools/caffe.cpp:
    -gpu (Optional; run in GPU mode on given device IDs separated by ','.Use
      '-gpu all' to run on all available GPUs. The effective training batch
      size is multiplied by the number of devices.) type: string default: ""
    -iterations (The number of iterations to run.) type: int32 default: 50
      currently: 100
    -level (Optional; network level.) type: int32 default: 0
    -model (The model definition protocol buffer text file.) type: string
      default: "" currently: "examples/mnist/lenet_train_test.prototxt"
    -phase (Optional; network phase (TRAIN or TEST). Only used for 'time'.)
      type: string default: ""
    -sighup_effect (Optional; action to take when a SIGHUP signal is received:
      snapshot, stop or none.) type: string default: "snapshot"
    -sigint_effect (Optional; action to take when a SIGINT signal is received:
      snapshot, stop or none.) type: string default: "stop"
    -snapshot (Optional; the snapshot solver state to resume training.)
      type: string default: ""
    -solver (The solver definition protocol buffer text file.) type: string
      default: ""
    -stage (Optional; network stages (not to be confused with phase), separated
      by ','.) type: string default: ""
    -weights (Optional; the pretrained weights to initialize finetuning,
      separated by ','. Cannot be set simultaneously with snapshot.)
      type: string default: ""
      currently: "examples/mnist/lenet_iter_10000.caffemodel"

2. 训练过程输出

I1228 12:42:22.826128  1478 layer_factory.hpp:77] Creating layer mnist
I1228 12:42:22.829514  1478 net.cpp:100] Creating Layer mnist
I1228 12:42:22.829562  1478 net.cpp:408] mnist -> data
I1228 12:42:22.829612  1478 net.cpp:408] mnist -> label
I1228 12:42:22.829762  1479 db_lmdb.cpp:35] Opened lmdb examples/mnist/mnist_train_lmdb
I1228 12:42:22.840992  1478 data_layer.cpp:41] output data size: 64,1,28,28
I1228 12:42:22.841316  1478 net.cpp:150] Setting up mnist
I1228 12:42:22.841362  1478 net.cpp:157] Top shape: 64 1 28 28 (50176)
I1228 12:42:22.841388  1478 net.cpp:157] Top shape: 64 (64)
I1228 12:42:22.841411  1478 net.cpp:165] Memory required for data: 200960
I1228 12:42:22.841439  1478 layer_factory.hpp:77] Creating layer conv1
I1228 12:42:22.841478  1478 net.cpp:100] Creating Layer conv1
I1228 12:42:22.841506  1478 net.cpp:434] conv1 <- data
I1228 12:42:22.841538  1478 net.cpp:408] conv1 -> conv1
I1228 12:42:22.842957  1478 net.cpp:150] Setting up conv1
I1228 12:42:22.842998  1478 net.cpp:157] Top shape: 64 20 24 24 (737280)
I1228 12:42:22.843020  1478 net.cpp:165] Memory required for data: 3150080
I1228 12:42:22.843055  1478 layer_factory.hpp:77] Creating layer pool1
I1228 12:42:22.843087  1478 net.cpp:100] Creating Layer pool1
I1228 12:42:22.843112  1478 net.cpp:434] pool1 <- conv1
I1228 12:42:22.843163  1478 net.cpp:408] pool1 -> pool1
I1228 12:42:22.843204  1478 net.cpp:150] Setting up pool1
I1228 12:42:22.843231  1478 net.cpp:157] Top shape: 64 20 12 12 (184320)
I1228 12:42:22.843253  1478 net.cpp:165] Memory required for data: 3887360
I1228 12:42:22.843276  1478 layer_factory.hpp:77] Creating layer conv2
I1228 12:42:22.843304  1478 net.cpp:100] Creating Layer conv2
I1228 12:42:22.843328  1478 net.cpp:434] conv2 <- pool1
I1228 12:42:22.843353  1478 net.cpp:408] conv2 -> conv2
I1228 12:42:22.843607  1478 net.cpp:150] Setting up conv2
I1228 12:42:22.843641  1478 net.cpp:157] Top shape: 64 50 8 8 (204800)
I1228 12:42:22.843663  1478 net.cpp:165] Memory required for data: 4706560
I1228 12:42:22.843691  1478 layer_factory.hpp:77] Creating layer pool2
I1228 12:42:22.843719  1478 net.cpp:100] Creating Layer pool2
I1228 12:42:22.843742  1478 net.cpp:434] pool2 <- conv2
I1228 12:42:22.843767  1478 net.cpp:408] pool2 -> pool2
I1228 12:42:22.843796  1478 net.cpp:150] Setting up pool2
I1228 12:42:22.843822  1478 net.cpp:157] Top shape: 64 50 4 4 (51200)
I1228 12:42:22.843844  1478 net.cpp:165] Memory required for data: 4911360
I1228 12:42:22.843868  1478 layer_factory.hpp:77] Creating layer ip1
I1228 12:42:22.843897  1478 net.cpp:100] Creating Layer ip1
I1228 12:42:22.843921  1478 net.cpp:434] ip1 <- pool2
I1228 12:42:22.843946  1478 net.cpp:408] ip1 -> ip1
I1228 12:42:22.847321  1478 net.cpp:150] Setting up ip1
I1228 12:42:22.847625  1478 net.cpp:157] Top shape: 64 500 (32000)
I1228 12:42:22.847650  1478 net.cpp:165] Memory required for data: 5039360
I1228 12:42:22.847679  1478 layer_factory.hpp:77] Creating layer relu1
I1228 12:42:22.847707  1478 net.cpp:100] Creating Layer relu1
I1228 12:42:22.847731  1478 net.cpp:434] relu1 <- ip1
I1228 12:42:22.847756  1478 net.cpp:395] relu1 -> ip1 (in-place)
I1228 12:42:22.847787  1478 net.cpp:150] Setting up relu1
I1228 12:42:22.847815  1478 net.cpp:157] Top shape: 64 500 (32000)
I1228 12:42:22.847836  1478 net.cpp:165] Memory required for data: 5167360
I1228 12:42:22.847859  1478 layer_factory.hpp:77] Creating layer ip2
I1228 12:42:22.847885  1478 net.cpp:100] Creating Layer ip2
I1228 12:42:22.847909  1478 net.cpp:434] ip2 <- ip1
I1228 12:42:22.847935  1478 net.cpp:408] ip2 -> ip2
I1228 12:42:22.848016  1478 net.cpp:150] Setting up ip2
I1228 12:42:22.848045  1478 net.cpp:157] Top shape: 64 10 (640)
I1228 12:42:22.848068  1478 net.cpp:165] Memory required for data: 5169920
I1228 12:42:22.848093  1478 layer_factory.hpp:77] Creating layer loss
I1228 12:42:22.848122  1478 net.cpp:100] Creating Layer loss
I1228 12:42:22.848145  1478 net.cpp:434] loss <- ip2
I1228 12:42:22.848168  1478 net.cpp:434] loss <- label
I1228 12:42:22.848194  1478 net.cpp:408] loss -> loss
I1228 12:42:22.848229  1478 layer_factory.hpp:77] Creating layer loss
I1228 12:42:22.848270  1478 net.cpp:150] Setting up loss
I1228 12:42:22.848297  1478 net.cpp:157] Top shape: (1)
I1228 12:42:22.848318  1478 net.cpp:160]     with loss weight 1
I1228 12:42:22.848358  1478 net.cpp:165] Memory required for data: 5169924
I1228 12:42:22.848381  1478 net.cpp:226] loss needs backward computation.
I1228 12:42:22.848404  1478 net.cpp:226] ip2 needs backward computation.
I1228 12:42:22.848426  1478 net.cpp:226] relu1 needs backward computation.
I1228 12:42:22.848448  1478 net.cpp:226] ip1 needs backward computation.
I1228 12:42:22.848470  1478 net.cpp:226] pool2 needs backward computation.
I1228 12:42:22.848492  1478 net.cpp:226] conv2 needs backward computation.
I1228 12:42:22.848515  1478 net.cpp:226] pool1 needs backward computation.
I1228 12:42:22.848541  1478 net.cpp:226] conv1 needs backward computation.
I1228 12:42:22.848563  1478 net.cpp:228] mnist does not need backward computation.
I1228 12:42:22.848585  1478 net.cpp:270] This network produces output loss
I1228 12:42:22.848613  1478 net.cpp:283] Network initialization done.
I1228 12:42:22.850889  1478 solver.cpp:181] Creating test net (#0) specified by net file: examples/mnist/lenet_train_test.prototxt
I1228 12:42:22.850950  1478 net.cpp:322] The NetState phase (1) differed from the phase (0) specified by a rule in layer mnist
I1228 12:42:22.851075  1478 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"
}
I1228 12:42:22.852957  1478 layer_factory.hpp:77] Creating layer mnist
I1228 12:42:22.853076  1478 net.cpp:100] Creating Layer mnist
I1228 12:42:22.853106  1478 net.cpp:408] mnist -> data
I1228 12:42:22.853135  1478 net.cpp:408] mnist -> label
I1228 12:42:22.853212  1481 db_lmdb.cpp:35] Opened lmdb examples/mnist/mnist_test_lmdb
I1228 12:42:22.853286  1478 data_layer.cpp:41] output data size: 100,1,28,28
I1228 12:42:22.853688  1478 net.cpp:150] Setting up mnist
I1228 12:42:22.853718  1478 net.cpp:157] Top shape: 100 1 28 28 (78400)
I1228 12:42:22.853739  1478 net.cpp:157] Top shape: 100 (100)
I1228 12:42:22.853757  1478 net.cpp:165] Memory required for data: 314000
I1228 12:42:22.853777  1478 layer_factory.hpp:77] Creating layer label_mnist_1_split
I1228 12:42:22.853802  1478 net.cpp:100] Creating Layer label_mnist_1_split
I1228 12:42:22.853821  1478 net.cpp:434] label_mnist_1_split <- label
I1228 12:42:22.853842  1478 net.cpp:408] label_mnist_1_split -> label_mnist_1_split_0
I1228 12:42:22.853866  1478 net.cpp:408] label_mnist_1_split -> label_mnist_1_split_1
I1228 12:42:22.853889  1478 net.cpp:150] Setting up label_mnist_1_split
I1228 12:42:22.853909  1478 net.cpp:157] Top shape: 100 (100)
I1228 12:42:22.853925  1478 net.cpp:157] Top shape: 100 (100)
I1228 12:42:22.853940  1478 net.cpp:165] Memory required for data: 314800
I1228 12:42:22.853955  1478 layer_factory.hpp:77] Creating layer conv1
I1228 12:42:22.853977  1478 net.cpp:100] Creating Layer conv1
I1228 12:42:22.853994  1478 net.cpp:434] conv1 <- data
I1228 12:42:22.854012  1478 net.cpp:408] conv1 -> conv1
I1228 12:42:22.854054  1478 net.cpp:150] Setting up conv1
I1228 12:42:22.854075  1478 net.cpp:157] Top shape: 100 20 24 24 (1152000)
I1228 12:42:22.854090  1478 net.cpp:165] Memory required for data: 4922800
I1228 12:42:22.854111  1478 layer_factory.hpp:77] Creating layer pool1
I1228 12:42:22.854132  1478 net.cpp:100] Creating Layer pool1
I1228 12:42:22.854161  1478 net.cpp:434] pool1 <- conv1
I1228 12:42:22.854178  1478 net.cpp:408] pool1 -> pool1
I1228 12:42:22.854200  1478 net.cpp:150] Setting up pool1
I1228 12:42:22.854220  1478 net.cpp:157] Top shape: 100 20 12 12 (288000)
I1228 12:42:22.854235  1478 net.cpp:165] Memory required for data: 6074800
I1228 12:42:22.854250  1478 layer_factory.hpp:77] Creating layer conv2
I1228 12:42:22.854270  1478 net.cpp:100] Creating Layer conv2
I1228 12:42:22.854286  1478 net.cpp:434] conv2 <- pool1
I1228 12:42:22.854306  1478 net.cpp:408] conv2 -> conv2
I1228 12:42:22.854542  1478 net.cpp:150] Setting up conv2
I1228 12:42:22.854564  1478 net.cpp:157] Top shape: 100 50 8 8 (320000)
I1228 12:42:22.854579  1478 net.cpp:165] Memory required for data: 7354800
I1228 12:42:22.854599  1478 layer_factory.hpp:77] Creating layer pool2
I1228 12:42:22.854617  1478 net.cpp:100] Creating Layer pool2
I1228 12:42:22.854636  1478 net.cpp:434] pool2 <- conv2
I1228 12:42:22.854655  1478 net.cpp:408] pool2 -> pool2
I1228 12:42:22.854677  1478 net.cpp:150] Setting up pool2
I1228 12:42:22.854694  1478 net.cpp:157] Top shape: 100 50 4 4 (80000)
I1228 12:42:22.854709  1478 net.cpp:165] Memory required for data: 7674800
I1228 12:42:22.854723  1478 layer_factory.hpp:77] Creating layer ip1
I1228 12:42:22.854742  1478 net.cpp:100] Creating Layer ip1
I1228 12:42:22.854759  1478 net.cpp:434] ip1 <- pool2
I1228 12:42:22.854776  1478 net.cpp:408] ip1 -> ip1
I1228 12:42:22.858683  1478 net.cpp:150] Setting up ip1
I1228 12:42:22.858729  1478 net.cpp:157] Top shape: 100 500 (50000)
I1228 12:42:22.858747  1478 net.cpp:165] Memory required for data: 7874800
I1228 12:42:22.858768  1478 layer_factory.hpp:77] Creating layer relu1
I1228 12:42:22.858789  1478 net.cpp:100] Creating Layer relu1
I1228 12:42:22.858805  1478 net.cpp:434] relu1 <- ip1
I1228 12:42:22.858822  1478 net.cpp:395] relu1 -> ip1 (in-place)
I1228 12:42:22.858842  1478 net.cpp:150] Setting up relu1
I1228 12:42:22.858860  1478 net.cpp:157] Top shape: 100 500 (50000)
I1228 12:42:22.858873  1478 net.cpp:165] Memory required for data: 8074800
I1228 12:42:22.858888  1478 layer_factory.hpp:77] Creating layer ip2
I1228 12:42:22.858908  1478 net.cpp:100] Creating Layer ip2
I1228 12:42:22.858924  1478 net.cpp:434] ip2 <- ip1
I1228 12:42:22.858942  1478 net.cpp:408] ip2 -> ip2
I1228 12:42:22.859010  1478 net.cpp:150] Setting up ip2
I1228 12:42:22.859030  1478 net.cpp:157] Top shape: 100 10 (1000)
I1228 12:42:22.859045  1478 net.cpp:165] Memory required for data: 8078800
I1228 12:42:22.859061  1478 layer_factory.hpp:77] Creating layer ip2_ip2_0_split
I1228 12:42:22.859079  1478 net.cpp:100] Creating Layer ip2_ip2_0_split
I1228 12:42:22.859094  1478 net.cpp:434] ip2_ip2_0_split <- ip2
I1228 12:42:22.859110  1478 net.cpp:408] ip2_ip2_0_split -> ip2_ip2_0_split_0
I1228 12:42:22.859129  1478 net.cpp:408] ip2_ip2_0_split -> ip2_ip2_0_split_1
I1228 12:42:22.859148  1478 net.cpp:150] Setting up ip2_ip2_0_split
I1228 12:42:22.859164  1478 net.cpp:157] Top shape: 100 10 (1000)
I1228 12:42:22.859186  1478 net.cpp:157] Top shape: 100 10 (1000)
I1228 12:42:22.859201  1478 net.cpp:165] Memory required for data: 8086800
I1228 12:42:22.859216  1478 layer_factory.hpp:77] Creating layer accuracy
I1228 12:42:22.859237  1478 net.cpp:100] Creating Layer accuracy
I1228 12:42:22.859253  1478 net.cpp:434] accuracy <- ip2_ip2_0_split_0
I1228 12:42:22.859268  1478 net.cpp:434] accuracy <- label_mnist_1_split_0
I1228 12:42:22.859288  1478 net.cpp:408] accuracy -> accuracy
I1228 12:42:22.888463  1478 net.cpp:150] Setting up accuracy
I1228 12:42:22.888525  1478 net.cpp:157] Top shape: (1)
I1228 12:42:22.888541  1478 net.cpp:165] Memory required for data: 8086804
I1228 12:42:22.888559  1478 layer_factory.hpp:77] Creating layer loss
I1228 12:42:22.888581  1478 net.cpp:100] Creating Layer loss
I1228 12:42:22.888599  1478 net.cpp:434] loss <- ip2_ip2_0_split_1
I1228 12:42:22.888617  1478 net.cpp:434] loss <- label_mnist_1_split_1
I1228 12:42:22.888635  1478 net.cpp:408] loss -> loss
I1228 12:42:22.888660  1478 layer_factory.hpp:77] Creating layer loss
I1228 12:42:22.888725  1478 net.cpp:150] Setting up loss
I1228 12:42:22.888746  1478 net.cpp:157] Top shape: (1)
I1228 12:42:22.888761  1478 net.cpp:160]     with loss weight 1
I1228 12:42:22.888783  1478 net.cpp:165] Memory required for data: 8086808
I1228 12:42:22.888799  1478 net.cpp:226] loss needs backward computation.
I1228 12:42:22.888815  1478 net.cpp:228] accuracy does not need backward computation.
I1228 12:42:22.888831  1478 net.cpp:226] ip2_ip2_0_split needs backward computation.
I1228 12:42:22.888846  1478 net.cpp:226] ip2 needs backward computation.
I1228 12:42:22.888862  1478 net.cpp:226] relu1 needs backward computation.
I1228 12:42:22.888877  1478 net.cpp:226] ip1 needs backward computation.
I1228 12:42:22.888892  1478 net.cpp:226] pool2 needs backward computation.
I1228 12:42:22.888909  1478 net.cpp:226] conv2 needs backward computation.
I1228 12:42:22.888926  1478 net.cpp:226] pool1 needs backward computation.
I1228 12:42:22.888941  1478 net.cpp:226] conv1 needs backward computation.
I1228 12:42:22.888957  1478 net.cpp:228] label_mnist_1_split does not need backward computation.
I1228 12:42:22.888972  1478 net.cpp:228] mnist does not need backward computation.
I1228 12:42:22.888988  1478 net.cpp:270] This network produces output accuracy
I1228 12:42:22.889003  1478 net.cpp:270] This network produces output loss
I1228 12:42:22.889026  1478 net.cpp:283] Network initialization done.
I1228 12:42:22.889114  1478 solver.cpp:60] Solver scaffolding done.
I1228 12:42:22.889156  1478 caffe.cpp:251] Starting Optimization
I1228 12:42:22.889174  1478 solver.cpp:279] Solving LeNet
I1228 12:42:22.889189  1478 solver.cpp:280] Learning Rate Policy: inv
I1228 12:42:22.889972  1478 solver.cpp:337] Iteration 0, Testing net (#0)
I1228 12:42:28.686501  1478 solver.cpp:404]     Test net output #0: accuracy = 0.0946
I1228 12:42:28.686633  1478 solver.cpp:404]     Test net output #1: loss = 2.33963 (* 1 = 2.33963 loss)
I1228 12:42:28.778022  1478 solver.cpp:228] Iteration 0, loss = 2.30006
I1228 12:42:28.778156  1478 solver.cpp:244]     Train net output #0: loss = 2.30006 (* 1 = 2.30006 loss)
I1228 12:42:28.778203  1478 sgd_solver.cpp:106] Iteration 0, lr = 0.01
I1228 12:42:37.817469  1478 solver.cpp:228] Iteration 100, loss = 0.18433
I1228 12:42:37.817602  1478 solver.cpp:244]     Train net output #0: loss = 0.184329 (* 1 = 0.184329 loss)
I1228 12:42:37.817647  1478 sgd_solver.cpp:106] Iteration 100, lr = 0.00992565
I1228 12:42:46.837074  1478 solver.cpp:228] Iteration 200, loss = 0.128627
I1228 12:42:46.837229  1478 solver.cpp:244]     Train net output #0: loss = 0.128627 (* 1 = 0.128627 loss)
I1228 12:42:46.837281  1478 sgd_solver.cpp:106] Iteration 200, lr = 0.00985258
I1228 12:42:55.860499  1478 solver.cpp:228] Iteration 300, loss = 0.184737
I1228 12:42:55.860676  1478 solver.cpp:244]     Train net output #0: loss = 0.184737 (* 1 = 0.184737 loss)
I1228 12:42:55.860715  1478 sgd_solver.cpp:106] Iteration 300, lr = 0.00978075
I1228 12:43:04.895009  1478 solver.cpp:228] Iteration 400, loss = 0.072573
I1228 12:43:04.895154  1478 solver.cpp:244]     Train net output #0: loss = 0.072573 (* 1 = 0.072573 loss)
I1228 12:43:04.895196  1478 sgd_solver.cpp:106] Iteration 400, lr = 0.00971013
I1228 12:43:13.807948  1478 solver.cpp:337] Iteration 500, Testing net (#0)
I1228 12:43:19.521517  1478 solver.cpp:404]     Test net output #0: accuracy = 0.9725
I1228 12:43:19.521663  1478 solver.cpp:404]     Test net output #1: loss = 0.0851369 (* 1 = 0.0851369 loss)
I1228 12:43:19.610584  1478 solver.cpp:228] Iteration 500, loss = 0.0924923
I1228 12:43:19.610713  1478 solver.cpp:244]     Train net output #0: loss = 0.0924922 (* 1 = 0.0924922 loss)
I1228 12:43:19.610759  1478 sgd_solver.cpp:106] Iteration 500, lr = 0.00964069
I1228 12:43:28.614856  1478 solver.cpp:228] Iteration 600, loss = 0.0846298
I1228 12:43:28.615057  1478 solver.cpp:244]     Train net output #0: loss = 0.0846298 (* 1 = 0.0846298 loss)
I1228 12:43:28.615099  1478 sgd_solver.cpp:106] Iteration 600, lr = 0.0095724
I1228 12:43:37.611666  1478 solver.cpp:228] Iteration 700, loss = 0.152674
I1228 12:43:37.611802  1478 solver.cpp:244]     Train net output #0: loss = 0.152674 (* 1 = 0.152674 loss)
I1228 12:43:37.611852  1478 sgd_solver.cpp:106] Iteration 700, lr = 0.00950522
I1228 12:43:46.585642  1478 solver.cpp:228] Iteration 800, loss = 0.197292
I1228 12:43:46.585780  1478 solver.cpp:244]     Train net output #0: loss = 0.197292 (* 1 = 0.197292 loss)
I1228 12:43:46.585820  1478 sgd_solver.cpp:106] Iteration 800, lr = 0.00943913
I1228 12:43:55.564452  1478 solver.cpp:228] Iteration 900, loss = 0.229118
I1228 12:43:55.564601  1478 solver.cpp:244]     Train net output #0: loss = 0.229118 (* 1 = 0.229118 loss)
I1228 12:43:55.564649  1478 sgd_solver.cpp:106] Iteration 900, lr = 0.00937411
I1228 12:44:04.491788  1478 solver.cpp:337] Iteration 1000, Testing net (#0)
I1228 12:44:10.195562  1478 solver.cpp:404]     Test net output #0: accuracy = 0.9813
I1228 12:44:10.195693  1478 solver.cpp:404]     Test net output #1: loss = 0.0588604 (* 1 = 0.0588604 loss)
I1228 12:44:10.284278  1478 solver.cpp:228] Iteration 1000, loss = 0.122648
I1228 12:44:10.284402  1478 solver.cpp:244]     Train net output #0: loss = 0.122648 (* 1 = 0.122648 loss)
I1228 12:44:10.284446  1478 sgd_solver.cpp:106] Iteration 1000, lr = 0.00931012
I1228 12:44:19.278929  1478 solver.cpp:228] Iteration 1100, loss = 0.00652424
I1228 12:44:19.279067  1478 solver.cpp:244]     Train net output #0: loss = 0.00652422 (* 1 = 0.00652422 loss)
I1228 12:44:19.279112  1478 sgd_solver.cpp:106] Iteration 1100, lr = 0.00924715
I1228 12:44:28.262037  1478 solver.cpp:228] Iteration 1200, loss = 0.0227062
I1228 12:44:28.262176  1478 solver.cpp:244]     Train net output #0: loss = 0.0227062 (* 1 = 0.0227062 loss)
I1228 12:44:28.262217  1478 sgd_solver.cpp:106] Iteration 1200, lr = 0.00918515
I1228 12:44:37.261036  1478 solver.cpp:228] Iteration 1300, loss = 0.0162803
I1228 12:44:37.261204  1478 solver.cpp:244]     Train net output #0: loss = 0.0162802 (* 1 = 0.0162802 loss)
I1228 12:44:37.261263  1478 sgd_solver.cpp:106] Iteration 1300, lr = 0.00912412
I1228 12:44:46.274796  1478 solver.cpp:228] Iteration 1400, loss = 0.0083404
I1228 12:44:46.274945  1478 solver.cpp:244]     Train net output #0: loss = 0.0083403 (* 1 = 0.0083403 loss)
I1228 12:44:46.274991  1478 sgd_solver.cpp:106] Iteration 1400, lr = 0.00906403
I1228 12:44:55.182224  1478 solver.cpp:337] Iteration 1500, Testing net (#0)
I1228 12:45:00.875882  1478 solver.cpp:404]     Test net output #0: accuracy = 0.9858
I1228 12:45:00.876024  1478 solver.cpp:404]     Test net output #1: loss = 0.0460209 (* 1 = 0.0460209 loss)
I1228 12:45:00.964514  1478 solver.cpp:228] Iteration 1500, loss = 0.0671758
I1228 12:45:00.964635  1478 solver.cpp:244]     Train net output #0: loss = 0.0671757 (* 1 = 0.0671757 loss)
I1228 12:45:00.964679  1478 sgd_solver.cpp:106] Iteration 1500, lr = 0.00900485
I1228 12:45:10.080106  1478 solver.cpp:228] Iteration 1600, loss = 0.106931
I1228 12:45:10.080278  1478 solver.cpp:244]     Train net output #0: loss = 0.106931 (* 1 = 0.106931 loss)
I1228 12:45:10.080320  1478 sgd_solver.cpp:106] Iteration 1600, lr = 0.00894657
I1228 12:45:19.073309  1478 solver.cpp:228] Iteration 1700, loss = 0.0179893
I1228 12:45:19.073454  1478 solver.cpp:244]     Train net output #0: loss = 0.0179892 (* 1 = 0.0179892 loss)
I1228 12:45:19.073495  1478 sgd_solver.cpp:106] Iteration 1700, lr = 0.00888916
I1228 12:45:28.069286  1478 solver.cpp:228] Iteration 1800, loss = 0.0228243
I1228 12:45:28.069427  1478 solver.cpp:244]     Train net output #0: loss = 0.0228242 (* 1 = 0.0228242 loss)
I1228 12:45:28.069473  1478 sgd_solver.cpp:106] Iteration 1800, lr = 0.0088326
I1228 12:45:37.070579  1478 solver.cpp:228] Iteration 1900, loss = 0.124677
I1228 12:45:37.070720  1478 solver.cpp:244]     Train net output #0: loss = 0.124677 (* 1 = 0.124677 loss)
I1228 12:45:37.070765  1478 sgd_solver.cpp:106] Iteration 1900, lr = 0.00877687
I1228 12:45:45.984365  1478 solver.cpp:337] Iteration 2000, Testing net (#0)
I1228 12:45:51.697198  1478 solver.cpp:404]     Test net output #0: accuracy = 0.9848
I1228 12:45:51.697345  1478 solver.cpp:404]     Test net output #1: loss = 0.0456813 (* 1 = 0.0456813 loss)
I1228 12:45:51.786160  1478 solver.cpp:228] Iteration 2000, loss = 0.0130174
I1228 12:45:51.786289  1478 solver.cpp:244]     Train net output #0: loss = 0.0130173 (* 1 = 0.0130173 loss)
I1228 12:45:51.786335  1478 sgd_solver.cpp:106] Iteration 2000, lr = 0.00872196
I1228 12:46:00.796350  1478 solver.cpp:228] Iteration 2100, loss = 0.0242315
I1228 12:46:00.796484  1478 solver.cpp:244]     Train net output #0: loss = 0.0242314 (* 1 = 0.0242314 loss)
I1228 12:46:00.796530  1478 sgd_solver.cpp:106] Iteration 2100, lr = 0.00866784
I1228 12:46:09.774925  1478 solver.cpp:228] Iteration 2200, loss = 0.0239079
I1228 12:46:09.775059  1478 solver.cpp:244]     Train net output #0: loss = 0.0239079 (* 1 = 0.0239079 loss)
I1228 12:46:09.775106  1478 sgd_solver.cpp:106] Iteration 2200, lr = 0.0086145
I1228 12:46:18.773983  1478 solver.cpp:228] Iteration 2300, loss = 0.0938903
I1228 12:46:18.774143  1478 solver.cpp:244]     Train net output #0: loss = 0.0938903 (* 1 = 0.0938903 loss)
I1228 12:46:18.774189  1478 sgd_solver.cpp:106] Iteration 2300, lr = 0.00856192
I1228 12:46:27.773385  1478 solver.cpp:228] Iteration 2400, loss = 0.0137466
I1228 12:46:27.773525  1478 solver.cpp:244]     Train net output #0: loss = 0.0137466 (* 1 = 0.0137466 loss)
I1228 12:46:27.773576  1478 sgd_solver.cpp:106] Iteration 2400, lr = 0.00851008


    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"
}
I1228 12:42:22.852957  1478 layer_factory.hpp:77] Creating layer mnist
I1228 12:42:22.853076  1478 net.cpp:100] Creating Layer mnist
I1228 12:42:22.853106  1478 net.cpp:408] mnist -> data
I1228 12:42:22.853135  1478 net.cpp:408] mnist -> label
I1228 12:42:22.853212  1481 db_lmdb.cpp:35] Opened lmdb examples/mnist/mnist_test_lmdb
I1228 12:42:22.853286  1478 data_layer.cpp:41] output data size: 100,1,28,28
I1228 12:42:22.853688  1478 net.cpp:150] Setting up mnist
I1228 12:42:22.853718  1478 net.cpp:157] Top shape: 100 1 28 28 (78400)
I1228 12:42:22.853739  1478 net.cpp:157] Top shape: 100 (100)
I1228 12:42:22.853757  1478 net.cpp:165] Memory required for data: 314000
I1228 12:42:22.853777  1478 layer_factory.hpp:77] Creating layer label_mnist_1_split
I1228 12:42:22.853802  1478 net.cpp:100] Creating Layer label_mnist_1_split
I1228 12:42:22.853821  1478 net.cpp:434] label_mnist_1_split <- label
I1228 12:42:22.853842  1478 net.cpp:408] label_mnist_1_split -> label_mnist_1_split_0
I1228 12:42:22.853866  1478 net.cpp:408] label_mnist_1_split -> label_mnist_1_split_1
I1228 12:42:22.853889  1478 net.cpp:150] Setting up label_mnist_1_split
I1228 12:42:22.853909  1478 net.cpp:157] Top shape: 100 (100)
I1228 12:42:22.853925  1478 net.cpp:157] Top shape: 100 (100)
I1228 12:42:22.853940  1478 net.cpp:165] Memory required for data: 314800
I1228 12:42:22.853955  1478 layer_factory.hpp:77] Creating layer conv1
I1228 12:42:22.853977  1478 net.cpp:100] Creating Layer conv1
I1228 12:42:22.853994  1478 net.cpp:434] conv1 <- data
I1228 12:42:22.854012  1478 net.cpp:408] conv1 -> conv1
I1228 12:42:22.854054  1478 net.cpp:150] Setting up conv1
I1228 12:42:22.854075  1478 net.cpp:157] Top shape: 100 20 24 24 (1152000)
I1228 12:42:22.854090  1478 net.cpp:165] Memory required for data: 4922800
I1228 12:42:22.854111  1478 layer_factory.hpp:77] Creating layer pool1
I1228 12:42:22.854132  1478 net.cpp:100] Creating Layer pool1
I1228 12:42:22.854161  1478 net.cpp:434] pool1 <- conv1
I1228 12:42:22.854178  1478 net.cpp:408] pool1 -> pool1
I1228 12:42:22.854200  1478 net.cpp:150] Setting up pool1
I1228 12:42:22.854220  1478 net.cpp:157] Top shape: 100 20 12 12 (288000)
I1228 12:42:22.854235  1478 net.cpp:165] Memory required for data: 6074800
I1228 12:42:22.854250  1478 layer_factory.hpp:77] Creating layer conv2
I1228 12:42:22.854270  1478 net.cpp:100] Creating Layer conv2
I1228 12:42:22.854286  1478 net.cpp:434] conv2 <- pool1
I1228 12:42:22.854306  1478 net.cpp:408] conv2 -> conv2
I1228 12:42:22.854542  1478 net.cpp:150] Setting up conv2
I1228 12:42:22.854564  1478 net.cpp:157] Top shape: 100 50 8 8 (320000)
I1228 12:42:22.854579  1478 net.cpp:165] Memory required for data: 7354800
I1228 12:42:22.854599  1478 layer_factory.hpp:77] Creating layer pool2
I1228 12:42:22.854617  1478 net.cpp:100] Creating Layer pool2
I1228 12:42:22.854636  1478 net.cpp:434] pool2 <- conv2
I1228 12:42:22.854655  1478 net.cpp:408] pool2 -> pool2
I1228 12:42:22.854677  1478 net.cpp:150] Setting up pool2
I1228 12:42:22.854694  1478 net.cpp:157] Top shape: 100 50 4 4 (80000)
I1228 12:42:22.854709  1478 net.cpp:165] Memory required for data: 7674800
I1228 12:42:22.854723  1478 layer_factory.hpp:77] Creating layer ip1
I1228 12:42:22.854742  1478 net.cpp:100] Creating Layer ip1
I1228 12:42:22.854759  1478 net.cpp:434] ip1 <- pool2
I1228 12:42:22.854776  1478 net.cpp:408] ip1 -> ip1
I1228 12:42:22.858683  1478 net.cpp:150] Setting up ip1
I1228 12:42:22.858729  1478 net.cpp:157] Top shape: 100 500 (50000)
I1228 12:42:22.858747  1478 net.cpp:165] Memory required for data: 7874800
I1228 12:42:22.858768  1478 layer_factory.hpp:77] Creating layer relu1
I1228 12:42:22.858789  1478 net.cpp:100] Creating Layer relu1
I1228 12:42:22.858805  1478 net.cpp:434] relu1 <- ip1
I1228 12:42:22.858822  1478 net.cpp:395] relu1 -> ip1 (in-place)
I1228 12:42:22.858842  1478 net.cpp:150] Setting up relu1
I1228 12:42:22.858860  1478 net.cpp:157] Top shape: 100 500 (50000)
I1228 12:42:22.858873  1478 net.cpp:165] Memory required for data: 8074800
I1228 12:42:22.858888  1478 layer_factory.hpp:77] Creating layer ip2
I1228 12:42:22.858908  1478 net.cpp:100] Creating Layer ip2
I1228 12:42:22.858924  1478 net.cpp:434] ip2 <- ip1
I1228 12:42:22.858942  1478 net.cpp:408] ip2 -> ip2
I1228 12:42:22.859010  1478 net.cpp:150] Setting up ip2
I1228 12:42:22.859030  1478 net.cpp:157] Top shape: 100 10 (1000)
I1228 12:42:22.859045  1478 net.cpp:165] Memory required for data: 8078800
I1228 12:42:22.859061  1478 layer_factory.hpp:77] Creating layer ip2_ip2_0_split
I1228 12:42:22.859079  1478 net.cpp:100] Creating Layer ip2_ip2_0_split
I1228 12:42:22.859094  1478 net.cpp:434] ip2_ip2_0_split <- ip2
I1228 12:42:22.859110  1478 net.cpp:408] ip2_ip2_0_split -> ip2_ip2_0_split_0
I1228 12:42:22.859129  1478 net.cpp:408] ip2_ip2_0_split -> ip2_ip2_0_split_1
I1228 12:42:22.859148  1478 net.cpp:150] Setting up ip2_ip2_0_split
I1228 12:42:22.859164  1478 net.cpp:157] Top shape: 100 10 (1000)
I1228 12:42:22.859186  1478 net.cpp:157] Top shape: 100 10 (1000)
I1228 12:42:22.859201  1478 net.cpp:165] Memory required for data: 8086800
I1228 12:42:22.859216  1478 layer_factory.hpp:77] Creating layer accuracy
I1228 12:42:22.859237  1478 net.cpp:100] Creating Layer accuracy
I1228 12:42:22.859253  1478 net.cpp:434] accuracy <- ip2_ip2_0_split_0
I1228 12:42:22.859268  1478 net.cpp:434] accuracy <- label_mnist_1_split_0
I1228 12:42:22.859288  1478 net.cpp:408] accuracy -> accuracy
I1228 12:42:22.888463  1478 net.cpp:150] Setting up accuracy
I1228 12:42:22.888525  1478 net.cpp:157] Top shape: (1)
I1228 12:42:22.888541  1478 net.cpp:165] Memory required for data: 8086804
I1228 12:42:22.888559  1478 layer_factory.hpp:77] Creating layer loss
I1228 12:42:22.888581  1478 net.cpp:100] Creating Layer loss
I1228 12:42:22.888599  1478 net.cpp:434] loss <- ip2_ip2_0_split_1
I1228 12:42:22.888617  1478 net.cpp:434] loss <- label_mnist_1_split_1
I1228 12:42:22.888635  1478 net.cpp:408] loss -> loss
I1228 12:42:22.888660  1478 layer_factory.hpp:77] Creating layer loss
I1228 12:42:22.888725  1478 net.cpp:150] Setting up loss
I1228 12:42:22.888746  1478 net.cpp:157] Top shape: (1)
I1228 12:42:22.888761  1478 net.cpp:160]     with loss weight 1
I1228 12:42:22.888783  1478 net.cpp:165] Memory required for data: 8086808
I1228 12:42:22.888799  1478 net.cpp:226] loss needs backward computation.
I1228 12:42:22.888815  1478 net.cpp:228] accuracy does not need backward computation.
I1228 12:42:22.888831  1478 net.cpp:226] ip2_ip2_0_split needs backward computation.
I1228 12:42:22.888846  1478 net.cpp:226] ip2 needs backward computation.
I1228 12:42:22.888862  1478 net.cpp:226] relu1 needs backward computation.
I1228 12:42:22.888877  1478 net.cpp:226] ip1 needs backward computation.
I1228 12:42:22.888892  1478 net.cpp:226] pool2 needs backward computation.
I1228 12:42:22.888909  1478 net.cpp:226] conv2 needs backward computation.
I1228 12:42:22.888926  1478 net.cpp:226] pool1 needs backward computation.
I1228 12:42:22.888941  1478 net.cpp:226] conv1 needs backward computation.
I1228 12:42:22.888957  1478 net.cpp:228] label_mnist_1_split does not need backward computation.
I1228 12:42:22.888972  1478 net.cpp:228] mnist does not need backward computation.
I1228 12:42:22.888988  1478 net.cpp:270] This network produces output accuracy
I1228 12:42:22.889003  1478 net.cpp:270] This network produces output loss
I1228 12:42:22.889026  1478 net.cpp:283] Network initialization done.
I1228 12:42:22.889114  1478 solver.cpp:60] Solver scaffolding done.
I1228 12:42:22.889156  1478 caffe.cpp:251] Starting Optimization
I1228 12:42:22.889174  1478 solver.cpp:279] Solving LeNet
I1228 12:42:22.889189  1478 solver.cpp:280] Learning Rate Policy: inv
I1228 12:42:22.889972  1478 solver.cpp:337] Iteration 0, Testing net (#0)
I1228 12:42:28.686501  1478 solver.cpp:404]     Test net output #0: accuracy = 0.0946
I1228 12:42:28.686633  1478 solver.cpp:404]     Test net output #1: loss = 2.33963 (* 1 = 2.33963 loss)
I1228 12:42:28.778022  1478 solver.cpp:228] Iteration 0, loss = 2.30006
I1228 12:42:28.778156  1478 solver.cpp:244]     Train net output #0: loss = 2.30006 (* 1 = 2.30006 loss)
I1228 12:42:28.778203  1478 sgd_solver.cpp:106] Iteration 0, lr = 0.01
I1228 12:42:37.817469  1478 solver.cpp:228] Iteration 100, loss = 0.18433
I1228 12:42:37.817602  1478 solver.cpp:244]     Train net output #0: loss = 0.184329 (* 1 = 0.184329 loss)
I1228 12:42:37.817647  1478 sgd_solver.cpp:106] Iteration 100, lr = 0.00992565
I1228 12:42:46.837074  1478 solver.cpp:228] Iteration 200, loss = 0.128627
I1228 12:42:46.837229  1478 solver.cpp:244]     Train net output #0: loss = 0.128627 (* 1 = 0.128627 loss)
I1228 12:42:46.837281  1478 sgd_solver.cpp:106] Iteration 200, lr = 0.00985258
I1228 12:42:55.860499  1478 solver.cpp:228] Iteration 300, loss = 0.184737
I1228 12:42:55.860676  1478 solver.cpp:244]     Train net output #0: loss = 0.184737 (* 1 = 0.184737 loss)
I1228 12:42:55.860715  1478 sgd_solver.cpp:106] Iteration 300, lr = 0.00978075
I1228 12:43:04.895009  1478 solver.cpp:228] Iteration 400, loss = 0.072573
I1228 12:43:04.895154  1478 solver.cpp:244]     Train net output #0: loss = 0.072573 (* 1 = 0.072573 loss)
I1228 12:43:04.895196  1478 sgd_solver.cpp:106] Iteration 400, lr = 0.00971013
I1228 12:43:13.807948  1478 solver.cpp:337] Iteration 500, Testing net (#0)
I1228 12:43:19.521517  1478 solver.cpp:404]     Test net output #0: accuracy = 0.9725
I1228 12:43:19.521663  1478 solver.cpp:404]     Test net output #1: loss = 0.0851369 (* 1 = 0.0851369 loss)
I1228 12:43:19.610584  1478 solver.cpp:228] Iteration 500, loss = 0.0924923
I1228 12:43:19.610713  1478 solver.cpp:244]     Train net output #0: loss = 0.0924922 (* 1 = 0.0924922 loss)
I1228 12:43:19.610759  1478 sgd_solver.cpp:106] Iteration 500, lr = 0.00964069
I1228 12:43:28.614856  1478 solver.cpp:228] Iteration 600, loss = 0.0846298
I1228 12:43:28.615057  1478 solver.cpp:244]     Train net output #0: loss = 0.0846298 (* 1 = 0.0846298 loss)
I1228 12:43:28.615099  1478 sgd_solver.cpp:106] Iteration 600, lr = 0.0095724
I1228 12:43:37.611666  1478 solver.cpp:228] Iteration 700, loss = 0.152674
I1228 12:43:37.611802  1478 solver.cpp:244]     Train net output #0: loss = 0.152674 (* 1 = 0.152674 loss)
I1228 12:43:37.611852  1478 sgd_solver.cpp:106] Iteration 700, lr = 0.00950522
I1228 12:43:46.585642  1478 solver.cpp:228] Iteration 800, loss = 0.197292
I1228 12:43:46.585780  1478 solver.cpp:244]     Train net output #0: loss = 0.197292 (* 1 = 0.197292 loss)
I1228 12:43:46.585820  1478 sgd_solver.cpp:106] Iteration 800, lr = 0.00943913
I1228 12:43:55.564452  1478 solver.cpp:228] Iteration 900, loss = 0.229118
I1228 12:43:55.564601  1478 solver.cpp:244]     Train net output #0: loss = 0.229118 (* 1 = 0.229118 loss)
I1228 12:43:55.564649  1478 sgd_solver.cpp:106] Iteration 900, lr = 0.00937411
I1228 12:44:04.491788  1478 solver.cpp:337] Iteration 1000, Testing net (#0)
I1228 12:44:10.195562  1478 solver.cpp:404]     Test net output #0: accuracy = 0.9813
I1228 12:44:10.195693  1478 solver.cpp:404]     Test net output #1: loss = 0.0588604 (* 1 = 0.0588604 loss)
I1228 12:44:10.284278  1478 solver.cpp:228] Iteration 1000, loss = 0.122648
I1228 12:44:10.284402  1478 solver.cpp:244]     Train net output #0: loss = 0.122648 (* 1 = 0.122648 loss)
I1228 12:44:10.284446  1478 sgd_solver.cpp:106] Iteration 1000, lr = 0.00931012
I1228 12:44:19.278929  1478 solver.cpp:228] Iteration 1100, loss = 0.00652424
I1228 12:44:19.279067  1478 solver.cpp:244]     Train net output #0: loss = 0.00652422 (* 1 = 0.00652422 loss)
I1228 12:44:19.279112  1478 sgd_solver.cpp:106] Iteration 1100, lr = 0.00924715
I1228 12:44:28.262037  1478 solver.cpp:228] Iteration 1200, loss = 0.0227062
I1228 12:44:28.262176  1478 solver.cpp:244]     Train net output #0: loss = 0.0227062 (* 1 = 0.0227062 loss)
I1228 12:44:28.262217  1478 sgd_solver.cpp:106] Iteration 1200, lr = 0.00918515
I1228 12:44:37.261036  1478 solver.cpp:228] Iteration 1300, loss = 0.0162803
I1228 12:44:37.261204  1478 solver.cpp:244]     Train net output #0: loss = 0.0162802 (* 1 = 0.0162802 loss)
I1228 12:44:37.261263  1478 sgd_solver.cpp:106] Iteration 1300, lr = 0.00912412
I1228 12:44:46.274796  1478 solver.cpp:228] Iteration 1400, loss = 0.0083404
I1228 12:44:46.274945  1478 solver.cpp:244]     Train net output #0: loss = 0.0083403 (* 1 = 0.0083403 loss)
I1228 12:44:46.274991  1478 sgd_solver.cpp:106] Iteration 1400, lr = 0.00906403
I1228 12:44:55.182224  1478 solver.cpp:337] Iteration 1500, Testing net (#0)
I1228 12:45:00.875882  1478 solver.cpp:404]     Test net output #0: accuracy = 0.9858
I1228 12:45:00.876024  1478 solver.cpp:404]     Test net output #1: loss = 0.0460209 (* 1 = 0.0460209 loss)
I1228 12:45:00.964514  1478 solver.cpp:228] Iteration 1500, loss = 0.0671758
I1228 12:45:00.964635  1478 solver.cpp:244]     Train net output #0: loss = 0.0671757 (* 1 = 0.0671757 loss)
I1228 12:45:00.964679  1478 sgd_solver.cpp:106] Iteration 1500, lr = 0.00900485
I1228 12:45:10.080106  1478 solver.cpp:228] Iteration 1600, loss = 0.106931
I1228 12:45:10.080278  1478 solver.cpp:244]     Train net output #0: loss = 0.106931 (* 1 = 0.106931 loss)
I1228 12:45:10.080320  1478 sgd_solver.cpp:106] Iteration 1600, lr = 0.00894657
I1228 12:45:19.073309  1478 solver.cpp:228] Iteration 1700, loss = 0.0179893
I1228 12:45:19.073454  1478 solver.cpp:244]     Train net output #0: loss = 0.0179892 (* 1 = 0.0179892 loss)
I1228 12:45:19.073495  1478 sgd_solver.cpp:106] Iteration 1700, lr = 0.00888916
I1228 12:45:28.069286  1478 solver.cpp:228] Iteration 1800, loss = 0.0228243
I1228 12:45:28.069427  1478 solver.cpp:244]     Train net output #0: loss = 0.0228242 (* 1 = 0.0228242 loss)
I1228 12:45:28.069473  1478 sgd_solver.cpp:106] Iteration 1800, lr = 0.0088326
I1228 12:45:37.070579  1478 solver.cpp:228] Iteration 1900, loss = 0.124677
I1228 12:45:37.070720  1478 solver.cpp:244]     Train net output #0: loss = 0.124677 (* 1 = 0.124677 loss)
I1228 12:45:37.070765  1478 sgd_solver.cpp:106] Iteration 1900, lr = 0.00877687
I1228 12:45:45.984365  1478 solver.cpp:337] Iteration 2000, Testing net (#0)
I1228 12:45:51.697198  1478 solver.cpp:404]     Test net output #0: accuracy = 0.9848
I1228 12:45:51.697345  1478 solver.cpp:404]     Test net output #1: loss = 0.0456813 (* 1 = 0.0456813 loss)
I1228 12:45:51.786160  1478 solver.cpp:228] Iteration 2000, loss = 0.0130174
I1228 12:45:51.786289  1478 solver.cpp:244]     Train net output #0: loss = 0.0130173 (* 1 = 0.0130173 loss)
I1228 12:45:51.786335  1478 sgd_solver.cpp:106] Iteration 2000, lr = 0.00872196
I1228 12:46:00.796350  1478 solver.cpp:228] Iteration 2100, loss = 0.0242315
I1228 12:46:00.796484  1478 solver.cpp:244]     Train net output #0: loss = 0.0242314 (* 1 = 0.0242314 loss)
I1228 12:46:00.796530  1478 sgd_solver.cpp:106] Iteration 2100, lr = 0.00866784
I1228 12:46:09.774925  1478 solver.cpp:228] Iteration 2200, loss = 0.0239079
I1228 12:46:09.775059  1478 solver.cpp:244]     Train net output #0: loss = 0.0239079 (* 1 = 0.0239079 loss)
I1228 12:46:09.775106  1478 sgd_solver.cpp:106] Iteration 2200, lr = 0.0086145
I1228 12:46:18.773983  1478 solver.cpp:228] Iteration 2300, loss = 0.0938903
I1228 12:46:18.774143  1478 solver.cpp:244]     Train net output #0: loss = 0.0938903 (* 1 = 0.0938903 loss)
I1228 12:46:18.774189  1478 sgd_solver.cpp:106] Iteration 2300, lr = 0.00856192
I1228 12:46:27.773385  1478 solver.cpp:228] Iteration 2400, loss = 0.0137466
I1228 12:46:27.773525  1478 solver.cpp:244]     Train net output #0: loss = 0.0137466 (* 1 = 0.0137466 loss)
I1228 12:46:27.773576  1478 sgd_solver.cpp:106] Iteration 2400, lr = 0.00851008
I1228 12:46:36.668257  1478 solver.cpp:337] Iteration 2500, Testing net (#0)
I1228 12:46:42.371877  1478 solver.cpp:404]     Test net output #0: accuracy = 0.9871
I1228 12:46:42.372016  1478 solver.cpp:404]     Test net output #1: loss = 0.0411888 (* 1 = 0.0411888 loss)
I1228 12:46:42.460232  1478 solver.cpp:228] Iteration 2500, loss = 0.026106
I1228 12:46:42.460366  1478 solver.cpp:244]     Train net output #0: loss = 0.0261059 (* 1 = 0.0261059 loss)
I1228 12:46:42.460404  1478 sgd_solver.cpp:106] Iteration 2500, lr = 0.00845897
I1228 12:46:51.466620  1478 solver.cpp:228] Iteration 2600, loss = 0.0685898
I1228 12:46:51.466787  1478 solver.cpp:244]     Train net output #0: loss = 0.0685898 (* 1 = 0.0685898 loss)
I1228 12:46:51.466833  1478 sgd_solver.cpp:106] Iteration 2600, lr = 0.00840857
I1228 12:47:00.456198  1478 solver.cpp:228] Iteration 2700, loss = 0.0783101
I1228 12:47:00.456333  1478 solver.cpp:244]     Train net output #0: loss = 0.0783101 (* 1 = 0.0783101 loss)
I1228 12:47:00.456378  1478 sgd_solver.cpp:106] Iteration 2700, lr = 0.00835886
I1228 12:47:09.458505  1478 solver.cpp:228] Iteration 2800, loss = 0.00134948
I1228 12:47:09.458650  1478 solver.cpp:244]     Train net output #0: loss = 0.00134948 (* 1 = 0.00134948 loss)
I1228 12:47:09.458689  1478 sgd_solver.cpp:106] Iteration 2800, lr = 0.00830984
I1228 12:47:18.460925  1478 solver.cpp:228] Iteration 2900, loss = 0.0173515
I1228 12:47:18.461069  1478 solver.cpp:244]     Train net output #0: loss = 0.0173515 (* 1 = 0.0173515 loss)
I1228 12:47:18.461120  1478 sgd_solver.cpp:106] Iteration 2900, lr = 0.00826148
I1228 12:47:27.370965  1478 solver.cpp:337] Iteration 3000, Testing net (#0)
I1228 12:47:33.058574  1478 solver.cpp:404]     Test net output #0: accuracy = 0.9878
I1228 12:47:33.058715  1478 solver.cpp:404]     Test net output #1: loss = 0.0368021 (* 1 = 0.0368021 loss)
I1228 12:47:33.146149  1478 solver.cpp:228] Iteration 3000, loss = 0.0130952
I1228 12:47:33.146278  1478 solver.cpp:244]     Train net output #0: loss = 0.0130951 (* 1 = 0.0130951 loss)
I1228 12:47:33.146327  1478 sgd_solver.cpp:106] Iteration 3000, lr = 0.00821377
I1228 12:47:42.119889  1478 solver.cpp:228] Iteration 3100, loss = 0.0163137
I1228 12:47:42.120026  1478 solver.cpp:244]     Train net output #0: loss = 0.0163136 (* 1 = 0.0163136 loss)
I1228 12:47:42.120066  1478 sgd_solver.cpp:106] Iteration 3100, lr = 0.0081667
I1228 12:47:51.101455  1478 solver.cpp:228] Iteration 3200, loss = 0.00634108
I1228 12:47:51.101596  1478 solver.cpp:244]     Train net output #0: loss = 0.00634104 (* 1 = 0.00634104 loss)
I1228 12:47:51.101644  1478 sgd_solver.cpp:106] Iteration 3200, lr = 0.00812025
I1228 12:48:00.102923  1478 solver.cpp:228] Iteration 3300, loss = 0.00822882
I1228 12:48:00.103108  1478 solver.cpp:244]     Train net output #0: loss = 0.00822874 (* 1 = 0.00822874 loss)
I1228 12:48:00.103154  1478 sgd_solver.cpp:106] Iteration 3300, lr = 0.00807442
I1228 12:48:09.112653  1478 solver.cpp:228] Iteration 3400, loss = 0.00865752
I1228 12:48:09.112803  1478 solver.cpp:244]     Train net output #0: loss = 0.00865745 (* 1 = 0.00865745 loss)
I1228 12:48:09.112845  1478 sgd_solver.cpp:106] Iteration 3400, lr = 0.00802918
I1228 12:48:18.008867  1478 solver.cpp:337] Iteration 3500, Testing net (#0)
I1228 12:48:23.709991  1478 solver.cpp:404]     Test net output #0: accuracy = 0.9855
I1228 12:48:23.710139  1478 solver.cpp:404]     Test net output #1: loss = 0.0418341 (* 1 = 0.0418341 loss)
I1228 12:48:23.797904  1478 solver.cpp:228] Iteration 3500, loss = 0.00539367
I1228 12:48:23.798038  1478 solver.cpp:244]     Train net output #0: loss = 0.00539362 (* 1 = 0.00539362 loss)
I1228 12:48:23.798079  1478 sgd_solver.cpp:106] Iteration 3500, lr = 0.00798454
I1228 12:48:32.793352  1478 solver.cpp:228] Iteration 3600, loss = 0.0325999
I1228 12:48:32.793519  1478 solver.cpp:244]     Train net output #0: loss = 0.0325999 (* 1 = 0.0325999 loss)
I1228 12:48:32.793565  1478 sgd_solver.cpp:106] Iteration 3600, lr = 0.00794046
I1228 12:48:41.778705  1478 solver.cpp:228] Iteration 3700, loss = 0.0312646
I1228 12:48:41.778842  1478 solver.cpp:244]     Train net output #0: loss = 0.0312646 (* 1 = 0.0312646 loss)
I1228 12:48:41.778890  1478 sgd_solver.cpp:106] Iteration 3700, lr = 0.00789695
I1228 12:48:50.761559  1478 solver.cpp:228] Iteration 3800, loss = 0.0199443
I1228 12:48:50.761703  1478 solver.cpp:244]     Train net output #0: loss = 0.0199443 (* 1 = 0.0199443 loss)
I1228 12:48:50.761754  1478 sgd_solver.cpp:106] Iteration 3800, lr = 0.007854
I1228 12:48:59.751957  1478 solver.cpp:228] Iteration 3900, loss = 0.0247828
I1228 12:48:59.752090  1478 solver.cpp:244]     Train net output #0: loss = 0.0247827 (* 1 = 0.0247827 loss)
I1228 12:48:59.752137  1478 sgd_solver.cpp:106] Iteration 3900, lr = 0.00781158
I1228 12:49:08.663188  1478 solver.cpp:337] Iteration 4000, Testing net (#0)
I1228 12:49:14.386026  1478 solver.cpp:404]     Test net output #0: accuracy = 0.9901
I1228 12:49:14.386165  1478 solver.cpp:404]     Test net output #1: loss = 0.031397 (* 1 = 0.031397 loss)
I1228 12:49:14.474560  1478 solver.cpp:228] Iteration 4000, loss = 0.0101118
I1228 12:49:14.474681  1478 solver.cpp:244]     Train net output #0: loss = 0.0101117 (* 1 = 0.0101117 loss)
I1228 12:49:14.474720  1478 sgd_solver.cpp:106] Iteration 4000, lr = 0.0077697
I1228 12:49:23.466630  1478 solver.cpp:228] Iteration 4100, loss = 0.0178342
I1228 12:49:23.466775  1478 solver.cpp:244]     Train net output #0: loss = 0.0178341 (* 1 = 0.0178341 loss)
I1228 12:49:23.466816  1478 sgd_solver.cpp:106] Iteration 4100, lr = 0.00772833
I1228 12:49:32.453449  1478 solver.cpp:228] Iteration 4200, loss = 0.00712245
I1228 12:49:32.453586  1478 solver.cpp:244]     Train net output #0: loss = 0.00712239 (* 1 = 0.00712239 loss)
I1228 12:49:32.453632  1478 sgd_solver.cpp:106] Iteration 4200, lr = 0.00768748
I1228 12:49:41.431674  1478 solver.cpp:228] Iteration 4300, loss = 0.0362492
I1228 12:49:41.431838  1478 solver.cpp:244]     Train net output #0: loss = 0.0362491 (* 1 = 0.0362491 loss)
I1228 12:49:41.431886  1478 sgd_solver.cpp:106] Iteration 4300, lr = 0.00764712
I1228 12:49:50.409802  1478 solver.cpp:228] Iteration 4400, loss = 0.0190512
I1228 12:49:50.409947  1478 solver.cpp:244]     Train net output #0: loss = 0.0190512 (* 1 = 0.0190512 loss)
I1228 12:49:50.409997  1478 sgd_solver.cpp:106] Iteration 4400, lr = 0.00760726
I1228 12:49:59.300992  1478 solver.cpp:337] Iteration 4500, Testing net (#0)
I1228 12:50:05.038296  1478 solver.cpp:404]     Test net output #0: accuracy = 0.9882
I1228 12:50:05.038434  1478 solver.cpp:404]     Test net output #1: loss = 0.0357055 (* 1 = 0.0357055 loss)
I1228 12:50:05.127156  1478 solver.cpp:228] Iteration 4500, loss = 0.00624241
I1228 12:50:05.127286  1478 solver.cpp:244]     Train net output #0: loss = 0.00624236 (* 1 = 0.00624236 loss)
I1228 12:50:05.127332  1478 sgd_solver.cpp:106] Iteration 4500, lr = 0.00756788
I1228 12:50:14.118090  1478 solver.cpp:228] Iteration 4600, loss = 0.0156433
I1228 12:50:14.118289  1478 solver.cpp:244]     Train net output #0: loss = 0.0156433 (* 1 = 0.0156433 loss)
I1228 12:50:14.118346  1478 sgd_solver.cpp:106] Iteration 4600, lr = 0.00752897
I1228 12:50:23.106927  1478 solver.cpp:228] Iteration 4700, loss = 0.00575534
I1228 12:50:23.107070  1478 solver.cpp:244]     Train net output #0: loss = 0.00575529 (* 1 = 0.00575529 loss)
I1228 12:50:23.107112  1478 sgd_solver.cpp:106] Iteration 4700, lr = 0.00749052
I1228 12:50:32.103515  1478 solver.cpp:228] Iteration 4800, loss = 0.0172193
I1228 12:50:32.103652  1478 solver.cpp:244]     Train net output #0: loss = 0.0172193 (* 1 = 0.0172193 loss)
I1228 12:50:32.103698  1478 sgd_solver.cpp:106] Iteration 4800, lr = 0.00745253
I1228 12:50:41.092864  1478 solver.cpp:228] Iteration 4900, loss = 0.0088059
I1228 12:50:41.092998  1478 solver.cpp:244]     Train net output #0: loss = 0.00880586 (* 1 = 0.00880586 loss)
I1228 12:50:41.093045  1478 sgd_solver.cpp:106] Iteration 4900, lr = 0.00741498
I1228 12:50:50.007995  1478 solver.cpp:454] Snapshotting to binary proto file examples/mnist/lenet_iter_5000.caffemodel
I1228 12:50:50.013608  1478 sgd_solver.cpp:273] Snapshotting solver state to binary proto file examples/mnist/lenet_iter_5000.solverstate
I1228 12:50:50.016250  1478 solver.cpp:337] Iteration 5000, Testing net (#0)
I1228 12:50:55.714702  1478 solver.cpp:404]     Test net output #0: accuracy = 0.99
I1228 12:50:55.714843  1478 solver.cpp:404]     Test net output #1: loss = 0.0308458 (* 1 = 0.0308458 loss)
I1228 12:50:55.802348  1478 solver.cpp:228] Iteration 5000, loss = 0.0337758
I1228 12:50:55.802469  1478 solver.cpp:244]     Train net output #0: loss = 0.0337758 (* 1 = 0.0337758 loss)
I1228 12:50:55.802513  1478 sgd_solver.cpp:106] Iteration 5000, lr = 0.00737788
I1228 12:51:04.774809  1478 solver.cpp:228] Iteration 5100, loss = 0.0166655
I1228 12:51:04.774946  1478 solver.cpp:244]     Train net output #0: loss = 0.0166655 (* 1 = 0.0166655 loss)
I1228 12:51:04.774987  1478 sgd_solver.cpp:106] Iteration 5100, lr = 0.0073412
I1228 12:51:13.770524  1478 solver.cpp:228] Iteration 5200, loss = 0.00703603
I1228 12:51:13.770663  1478 solver.cpp:244]     Train net output #0: loss = 0.00703599 (* 1 = 0.00703599 loss)
I1228 12:51:13.770704  1478 sgd_solver.cpp:106] Iteration 5200, lr = 0.00730495
I1228 12:51:22.773340  1478 solver.cpp:228] Iteration 5300, loss = 0.00212657
I1228 12:51:22.773501  1478 solver.cpp:244]     Train net output #0: loss = 0.00212655 (* 1 = 0.00212655 loss)
I1228 12:51:22.773547  1478 sgd_solver.cpp:106] Iteration 5300, lr = 0.00726911
I1228 12:51:31.756722  1478 solver.cpp:228] Iteration 5400, loss = 0.00780362
I1228 12:51:31.756865  1478 solver.cpp:244]     Train net output #0: loss = 0.00780359 (* 1 = 0.00780359 loss)
I1228 12:51:31.756906  1478 sgd_solver.cpp:106] Iteration 5400, lr = 0.00723368
I1228 12:51:40.641762  1478 solver.cpp:337] Iteration 5500, Testing net (#0)
I1228 12:51:46.328320  1478 solver.cpp:404]     Test net output #0: accuracy = 0.9892
I1228 12:51:46.328452  1478 solver.cpp:404]     Test net output #1: loss = 0.0326295 (* 1 = 0.0326295 loss)
I1228 12:51:46.417279  1478 solver.cpp:228] Iteration 5500, loss = 0.0109335
I1228 12:51:46.417410  1478 solver.cpp:244]     Train net output #0: loss = 0.0109334 (* 1 = 0.0109334 loss)
I1228 12:51:46.417450  1478 sgd_solver.cpp:106] Iteration 5500, lr = 0.00719865
I1228 12:51:55.415185  1478 solver.cpp:228] Iteration 5600, loss = 0.000549599
I1228 12:51:55.415355  1478 solver.cpp:244]     Train net output #0: loss = 0.000549559 (* 1 = 0.000549559 loss)
I1228 12:51:55.415397  1478 sgd_solver.cpp:106] Iteration 5600, lr = 0.00716402
I1228 12:52:04.402909  1478 solver.cpp:228] Iteration 5700, loss = 0.00133583
I1228 12:52:04.403048  1478 solver.cpp:244]     Train net output #0: loss = 0.00133578 (* 1 = 0.00133578 loss)
I1228 12:52:04.403093  1478 sgd_solver.cpp:106] Iteration 5700, lr = 0.00712977
I1228 12:52:13.359767  1478 solver.cpp:228] Iteration 5800, loss = 0.0199287
I1228 12:52:13.359917  1478 solver.cpp:244]     Train net output #0: loss = 0.0199286 (* 1 = 0.0199286 loss)
I1228 12:52:13.359971  1478 sgd_solver.cpp:106] Iteration 5800, lr = 0.0070959
I1228 12:52:22.356218  1478 solver.cpp:228] Iteration 5900, loss = 0.00755135
I1228 12:52:22.356365  1478 solver.cpp:244]     Train net output #0: loss = 0.00755131 (* 1 = 0.00755131 loss)
I1228 12:52:22.356408  1478 sgd_solver.cpp:106] Iteration 5900, lr = 0.0070624
I1228 12:52:31.254874  1478 solver.cpp:337] Iteration 6000, Testing net (#0)
I1228 12:52:36.958190  1478 solver.cpp:404]     Test net output #0: accuracy = 0.9903
I1228 12:52:36.958329  1478 solver.cpp:404]     Test net output #1: loss = 0.0295362 (* 1 = 0.0295362 loss)
I1228 12:52:37.046069  1478 solver.cpp:228] Iteration 6000, loss = 0.00353343
I1228 12:52:37.046198  1478 solver.cpp:244]     Train net output #0: loss = 0.0035334 (* 1 = 0.0035334 loss)
I1228 12:52:37.046244  1478 sgd_solver.cpp:106] Iteration 6000, lr = 0.00702927
I1228 12:52:46.023576  1478 solver.cpp:228] Iteration 6100, loss = 0.00124939
I1228 12:52:46.023715  1478 solver.cpp:244]     Train net output #0: loss = 0.00124934 (* 1 = 0.00124934 loss)
I1228 12:52:46.023766  1478 sgd_solver.cpp:106] Iteration 6100, lr = 0.0069965
I1228 12:52:55.006597  1478 solver.cpp:228] Iteration 6200, loss = 0.0103351
I1228 12:52:55.006745  1478 solver.cpp:244]     Train net output #0: loss = 0.010335 (* 1 = 0.010335 loss)
I1228 12:52:55.006791  1478 sgd_solver.cpp:106] Iteration 6200, lr = 0.00696408
I1228 12:53:04.008935  1478 solver.cpp:228] Iteration 6300, loss = 0.00820768
I1228 12:53:04.009102  1478 solver.cpp:244]     Train net output #0: loss = 0.00820762 (* 1 = 0.00820762 loss)
I1228 12:53:04.009150  1478 sgd_solver.cpp:106] Iteration 6300, lr = 0.00693201
I1228 12:53:13.010681  1478 solver.cpp:228] Iteration 6400, loss = 0.00429668
I1228 12:53:13.010815  1478 solver.cpp:244]     Train net output #0: loss = 0.0042966 (* 1 = 0.0042966 loss)
I1228 12:53:13.010862  1478 sgd_solver.cpp:106] Iteration 6400, lr = 0.00690029
I1228 12:53:21.929111  1478 solver.cpp:337] Iteration 6500, Testing net (#0)
I1228 12:53:27.614408  1478 solver.cpp:404]     Test net output #0: accuracy = 0.9904
I1228 12:53:27.614542  1478 solver.cpp:404]     Test net output #1: loss = 0.0322547 (* 1 = 0.0322547 loss)
I1228 12:53:27.703261  1478 solver.cpp:228] Iteration 6500, loss = 0.0182292
I1228 12:53:27.703389  1478 solver.cpp:244]     Train net output #0: loss = 0.0182291 (* 1 = 0.0182291 loss)
I1228 12:53:27.703429  1478 sgd_solver.cpp:106] Iteration 6500, lr = 0.0068689
I1228 12:53:36.712234  1478 solver.cpp:228] Iteration 6600, loss = 0.0449371
I1228 12:53:36.712409  1478 solver.cpp:244]     Train net output #0: loss = 0.044937 (* 1 = 0.044937 loss)
I1228 12:53:36.712451  1478 sgd_solver.cpp:106] Iteration 6600, lr = 0.00683784
I1228 12:53:45.693488  1478 solver.cpp:228] Iteration 6700, loss = 0.0062013
I1228 12:53:45.693635  1478 solver.cpp:244]     Train net output #0: loss = 0.00620122 (* 1 = 0.00620122 loss)
I1228 12:53:45.693684  1478 sgd_solver.cpp:106] Iteration 6700, lr = 0.00680711
I1228 12:53:54.686039  1478 solver.cpp:228] Iteration 6800, loss = 0.00341845
I1228 12:53:54.686187  1478 solver.cpp:244]     Train net output #0: loss = 0.00341837 (* 1 = 0.00341837 loss)
I1228 12:53:54.686236  1478 sgd_solver.cpp:106] Iteration 6800, lr = 0.0067767
I1228 12:54:03.675786  1478 solver.cpp:228] Iteration 6900, loss = 0.00433415
I1228 12:54:03.675925  1478 solver.cpp:244]     Train net output #0: loss = 0.00433407 (* 1 = 0.00433407 loss)
I1228 12:54:03.675966  1478 sgd_solver.cpp:106] Iteration 6900, lr = 0.0067466
I1228 12:54:12.580624  1478 solver.cpp:337] Iteration 7000, Testing net (#0)
I1228 12:54:18.301859  1478 solver.cpp:404]     Test net output #0: accuracy = 0.9903
I1228 12:54:18.302001  1478 solver.cpp:404]     Test net output #1: loss = 0.0303683 (* 1 = 0.0303683 loss)
I1228 12:54:18.389706  1478 solver.cpp:228] Iteration 7000, loss = 0.00686864
I1228 12:54:18.389844  1478 solver.cpp:244]     Train net output #0: loss = 0.00686856 (* 1 = 0.00686856 loss)
I1228 12:54:18.389890  1478 sgd_solver.cpp:106] Iteration 7000, lr = 0.00671681
I1228 12:54:27.372422  1478 solver.cpp:228] Iteration 7100, loss = 0.00886238
I1228 12:54:27.372568  1478 solver.cpp:244]     Train net output #0: loss = 0.00886229 (* 1 = 0.00886229 loss)
I1228 12:54:27.372619  1478 sgd_solver.cpp:106] Iteration 7100, lr = 0.00668733
I1228 12:54:36.339958  1478 solver.cpp:228] Iteration 7200, loss = 0.00408799
I1228 12:54:36.340101  1478 solver.cpp:244]     Train net output #0: loss = 0.00408792 (* 1 = 0.00408792 loss)
I1228 12:54:36.340149  1478 sgd_solver.cpp:106] Iteration 7200, lr = 0.00665815
I1228 12:54:45.325973  1478 solver.cpp:228] Iteration 7300, loss = 0.0171202
I1228 12:54:45.326149  1478 solver.cpp:244]     Train net output #0: loss = 0.0171201 (* 1 = 0.0171201 loss)
I1228 12:54:45.326191  1478 sgd_solver.cpp:106] Iteration 7300, lr = 0.00662927
I1228 12:54:54.329740  1478 solver.cpp:228] Iteration 7400, loss = 0.00486002
I1228 12:54:54.329880  1478 solver.cpp:244]     Train net output #0: loss = 0.00485994 (* 1 = 0.00485994 loss)
I1228 12:54:54.329922  1478 sgd_solver.cpp:106] Iteration 7400, lr = 0.00660067
I1228 12:55:03.241461  1478 solver.cpp:337] Iteration 7500, Testing net (#0)
I1228 12:55:08.933132  1478 solver.cpp:404]     Test net output #0: accuracy = 0.99
I1228 12:55:08.933291  1478 solver.cpp:404]     Test net output #1: loss = 0.0325328 (* 1 = 0.0325328 loss)
I1228 12:55:09.021898  1478 solver.cpp:228] Iteration 7500, loss = 0.00268578
I1228 12:55:09.022032  1478 solver.cpp:244]     Train net output #0: loss = 0.0026857 (* 1 = 0.0026857 loss)
I1228 12:55:09.022073  1478 sgd_solver.cpp:106] Iteration 7500, lr = 0.00657236
I1228 12:55:18.020838  1478 solver.cpp:228] Iteration 7600, loss = 0.0053811
I1228 12:55:18.021009  1478 solver.cpp:244]     Train net output #0: loss = 0.00538101 (* 1 = 0.00538101 loss)
I1228 12:55:18.021050  1478 sgd_solver.cpp:106] Iteration 7600, lr = 0.00654433
I1228 12:55:26.999490  1478 solver.cpp:228] Iteration 7700, loss = 0.0290023
I1228 12:55:26.999635  1478 solver.cpp:244]     Train net output #0: loss = 0.0290022 (* 1 = 0.0290022 loss)
I1228 12:55:26.999687  1478 sgd_solver.cpp:106] Iteration 7700, lr = 0.00651658
I1228 12:55:35.992398  1478 solver.cpp:228] Iteration 7800, loss = 0.00495144
I1228 12:55:35.992539  1478 solver.cpp:244]     Train net output #0: loss = 0.00495136 (* 1 = 0.00495136 loss)
I1228 12:55:35.992589  1478 sgd_solver.cpp:106] Iteration 7800, lr = 0.00648911
I1228 12:55:44.981251  1478 solver.cpp:228] Iteration 7900, loss = 0.00728552
I1228 12:55:44.981393  1478 solver.cpp:244]     Train net output #0: loss = 0.00728544 (* 1 = 0.00728544 loss)
I1228 12:55:44.981439  1478 sgd_solver.cpp:106] Iteration 7900, lr = 0.0064619
I1228 12:55:53.876646  1478 solver.cpp:337] Iteration 8000, Testing net (#0)
I1228 12:55:59.564527  1478 solver.cpp:404]     Test net output #0: accuracy = 0.9903
I1228 12:55:59.564671  1478 solver.cpp:404]     Test net output #1: loss = 0.0303976 (* 1 = 0.0303976 loss)
I1228 12:55:59.653228  1478 solver.cpp:228] Iteration 8000, loss = 0.00487312
I1228 12:55:59.653352  1478 solver.cpp:244]     Train net output #0: loss = 0.00487304 (* 1 = 0.00487304 loss)
I1228 12:55:59.653396  1478 sgd_solver.cpp:106] Iteration 8000, lr = 0.00643496
I1228 12:56:08.654803  1478 solver.cpp:228] Iteration 8100, loss = 0.00874484
I1228 12:56:08.654949  1478 solver.cpp:244]     Train net output #0: loss = 0.00874476 (* 1 = 0.00874476 loss)
I1228 12:56:08.654991  1478 sgd_solver.cpp:106] Iteration 8100, lr = 0.00640827
I1228 12:56:17.630285  1478 solver.cpp:228] Iteration 8200, loss = 0.00797972
I1228 12:56:17.630429  1478 solver.cpp:244]     Train net output #0: loss = 0.00797963 (* 1 = 0.00797963 loss)
I1228 12:56:17.630470  1478 sgd_solver.cpp:106] Iteration 8200, lr = 0.00638185
I1228 12:56:26.607856  1478 solver.cpp:228] Iteration 8300, loss = 0.0234358
I1228 12:56:26.608052  1478 solver.cpp:244]     Train net output #0: loss = 0.0234357 (* 1 = 0.0234357 loss)
I1228 12:56:26.608094  1478 sgd_solver.cpp:106] Iteration 8300, lr = 0.00635567
I1228 12:56:35.597169  1478 solver.cpp:228] Iteration 8400, loss = 0.00902765
I1228 12:56:35.597326  1478 solver.cpp:244]     Train net output #0: loss = 0.00902755 (* 1 = 0.00902755 loss)
I1228 12:56:35.597374  1478 sgd_solver.cpp:106] Iteration 8400, lr = 0.00632975
I1228 12:56:44.483781  1478 solver.cpp:337] Iteration 8500, Testing net (#0)
I1228 12:56:50.179303  1478 solver.cpp:404]     Test net output #0: accuracy = 0.9906
I1228 12:56:50.179446  1478 solver.cpp:404]     Test net output #1: loss = 0.0297353 (* 1 = 0.0297353 loss)
I1228 12:56:50.268074  1478 solver.cpp:228] Iteration 8500, loss = 0.00623272
I1228 12:56:50.268204  1478 solver.cpp:244]     Train net output #0: loss = 0.00623262 (* 1 = 0.00623262 loss)
I1228 12:56:50.268244  1478 sgd_solver.cpp:106] Iteration 8500, lr = 0.00630407
I1228 12:56:59.266109  1478 solver.cpp:228] Iteration 8600, loss = 0.000967596
I1228 12:56:59.266285  1478 solver.cpp:244]     Train net output #0: loss = 0.000967495 (* 1 = 0.000967495 loss)
I1228 12:56:59.266335  1478 sgd_solver.cpp:106] Iteration 8600, lr = 0.00627864
I1228 12:57:08.254199  1478 solver.cpp:228] Iteration 8700, loss = 0.00239873
I1228 12:57:08.254345  1478 solver.cpp:244]     Train net output #0: loss = 0.00239862 (* 1 = 0.00239862 loss)
I1228 12:57:08.254386  1478 sgd_solver.cpp:106] Iteration 8700, lr = 0.00625344
I1228 12:57:17.251121  1478 solver.cpp:228] Iteration 8800, loss = 0.00144548
I1228 12:57:17.251268  1478 solver.cpp:244]     Train net output #0: loss = 0.00144538 (* 1 = 0.00144538 loss)
I1228 12:57:17.251318  1478 sgd_solver.cpp:106] Iteration 8800, lr = 0.00622847
I1228 12:57:26.228776  1478 solver.cpp:228] Iteration 8900, loss = 0.000345843
I1228 12:57:26.228924  1478 solver.cpp:244]     Train net output #0: loss = 0.000345736 (* 1 = 0.000345736 loss)
I1228 12:57:26.228965  1478 sgd_solver.cpp:106] Iteration 8900, lr = 0.00620374
I1228 12:57:35.132206  1478 solver.cpp:337] Iteration 9000, Testing net (#0)
I1228 12:57:40.816022  1478 solver.cpp:404]     Test net output #0: accuracy = 0.9909
I1228 12:57:40.816154  1478 solver.cpp:404]     Test net output #1: loss = 0.0278701 (* 1 = 0.0278701 loss)
I1228 12:57:40.904713  1478 solver.cpp:228] Iteration 9000, loss = 0.0132574
I1228 12:57:40.904846  1478 solver.cpp:244]     Train net output #0: loss = 0.0132573 (* 1 = 0.0132573 loss)
I1228 12:57:40.904887  1478 sgd_solver.cpp:106] Iteration 9000, lr = 0.00617924
I1228 12:57:49.887485  1478 solver.cpp:228] Iteration 9100, loss = 0.00660216
I1228 12:57:49.887621  1478 solver.cpp:244]     Train net output #0: loss = 0.00660205 (* 1 = 0.00660205 loss)
I1228 12:57:49.887666  1478 sgd_solver.cpp:106] Iteration 9100, lr = 0.00615496
I1228 12:57:58.880218  1478 solver.cpp:228] Iteration 9200, loss = 0.00292292
I1228 12:57:58.880364  1478 solver.cpp:244]     Train net output #0: loss = 0.00292282 (* 1 = 0.00292282 loss)
I1228 12:57:58.880406  1478 sgd_solver.cpp:106] Iteration 9200, lr = 0.0061309
I1228 12:58:07.862749  1478 solver.cpp:228] Iteration 9300, loss = 0.00458688
I1228 12:58:07.862910  1478 solver.cpp:244]     Train net output #0: loss = 0.00458676 (* 1 = 0.00458676 loss)
I1228 12:58:07.862957  1478 sgd_solver.cpp:106] Iteration 9300, lr = 0.00610706
I1228 12:58:16.836311  1478 solver.cpp:228] Iteration 9400, loss = 0.0341082
I1228 12:58:16.836454  1478 solver.cpp:244]     Train net output #0: loss = 0.0341081 (* 1 = 0.0341081 loss)
I1228 12:58:16.836496  1478 sgd_solver.cpp:106] Iteration 9400, lr = 0.00608343
I1228 12:58:25.749003  1478 solver.cpp:337] Iteration 9500, Testing net (#0)
I1228 12:58:31.442245  1478 solver.cpp:404]     Test net output #0: accuracy = 0.989
I1228 12:58:31.442381  1478 solver.cpp:404]     Test net output #1: loss = 0.0340806 (* 1 = 0.0340806 loss)
I1228 12:58:31.531092  1478 solver.cpp:228] Iteration 9500, loss = 0.00251447
I1228 12:58:31.531222  1478 solver.cpp:244]     Train net output #0: loss = 0.00251435 (* 1 = 0.00251435 loss)
I1228 12:58:31.531267  1478 sgd_solver.cpp:106] Iteration 9500, lr = 0.00606002
I1228 12:58:40.530058  1478 solver.cpp:228] Iteration 9600, loss = 0.0033457
I1228 12:58:40.530247  1478 solver.cpp:244]     Train net output #0: loss = 0.00334557 (* 1 = 0.00334557 loss)
I1228 12:58:40.530294  1478 sgd_solver.cpp:106] Iteration 9600, lr = 0.00603682
I1228 12:58:49.502943  1478 solver.cpp:228] Iteration 9700, loss = 0.00308436
I1228 12:58:49.503093  1478 solver.cpp:244]     Train net output #0: loss = 0.00308424 (* 1 = 0.00308424 loss)
I1228 12:58:49.503145  1478 sgd_solver.cpp:106] Iteration 9700, lr = 0.00601382
I1228 12:58:58.485988  1478 solver.cpp:228] Iteration 9800, loss = 0.00986355
I1228 12:58:58.486138  1478 solver.cpp:244]     Train net output #0: loss = 0.00986342 (* 1 = 0.00986342 loss)
I1228 12:58:58.486186  1478 sgd_solver.cpp:106] Iteration 9800, lr = 0.00599102
I1228 12:59:07.481578  1478 solver.cpp:228] Iteration 9900, loss = 0.00657413
I1228 12:59:07.481719  1478 solver.cpp:244]     Train net output #0: loss = 0.006574 (* 1 = 0.006574 loss)
I1228 12:59:07.481765  1478 sgd_solver.cpp:106] Iteration 9900, lr = 0.00596843
I1228 12:59:16.380491  1478 solver.cpp:454] Snapshotting to binary proto file examples/mnist/lenet_iter_10000.caffemodel
I1228 12:59:16.386615  1478 sgd_solver.cpp:273] Snapshotting solver state to binary proto file examples/mnist/lenet_iter_10000.solverstate
I1228 12:59:16.425662  1478 solver.cpp:317] Iteration 10000, loss = 0.00357283
I1228 12:59:16.425776  1478 solver.cpp:337] Iteration 10000, Testing net (#0)
I1228 12:59:22.108726  1478 solver.cpp:404]     Test net output #0: accuracy = 0.9913
I1228 12:59:22.108870  1478 solver.cpp:404]     Test net output #1: loss = 0.0273851 (* 1 = 0.0273851 loss)
I1228 12:59:22.108907  1478 solver.cpp:322] Optimization Done.
I1228 12:59:22.108942  1478 caffe.cpp:254] Optimization Done.
root@ip-172-30-0-251:/caffe# Write failed: Broken pipe
[root@xuyongshi aws.rds]# 








3. 预测过程部分输出



root@ip-172-30-0-251:/caffe# build/tools/caffe test  --model=examples/mnist/lenet_train_test.prototxt  --weights examples/mnist/lenet_iter_10000.caffemodel  -iterations 100I1229 08:30:11.945741  4159 caffe.cpp:279] Use CPU.
I1229 08:30:11.947610  4159 net.cpp:322] The NetState phase (1) differed from the phase (0) specified by a rule in layer mnist
I1229 08:30:11.947794  4159 net.cpp:58] Initializing net from parameters: 
name: "LeNet"
state {
  phase: TEST
  level: 0
  stage: ""
}
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"
}
I1229 08:30:11.950266  4159 layer_factory.hpp:77] Creating layer mnist
I1229 08:30:11.950832  4159 net.cpp:100] Creating Layer mnist
I1229 08:30:11.950876  4159 net.cpp:408] mnist -> data
I1229 08:30:11.950924  4159 net.cpp:408] mnist -> label
I1229 08:30:11.951063  4160 db_lmdb.cpp:35] Opened lmdb examples/mnist/mnist_test_lmdb
I1229 08:30:11.951145  4159 data_layer.cpp:41] output data size: 100,1,28,28
I1229 08:30:11.951928  4159 net.cpp:150] Setting up mnist
I1229 08:30:11.951970  4159 net.cpp:157] Top shape: 100 1 28 28 (78400)
I1229 08:30:11.951997  4159 net.cpp:157] Top shape: 100 (100)
I1229 08:30:11.952021  4159 net.cpp:165] Memory required for data: 314000
I1229 08:30:11.952049  4159 layer_factory.hpp:77] Creating layer label_mnist_1_split
I1229 08:30:11.952078  4159 net.cpp:100] Creating Layer label_mnist_1_split
I1229 08:30:11.952105  4159 net.cpp:434] label_mnist_1_split <- label
I1229 08:30:11.952137  4159 net.cpp:408] label_mnist_1_split -> label_mnist_1_split_0
I1229 08:30:11.952167  4159 net.cpp:408] label_mnist_1_split -> label_mnist_1_split_1
I1229 08:30:11.952200  4159 net.cpp:150] Setting up label_mnist_1_split
I1229 08:30:11.952227  4159 net.cpp:157] Top shape: 100 (100)
I1229 08:30:11.952252  4159 net.cpp:157] Top shape: 100 (100)
I1229 08:30:11.952275  4159 net.cpp:165] Memory required for data: 314800
I1229 08:30:11.952298  4159 layer_factory.hpp:77] Creating layer conv1
I1229 08:30:11.952334  4159 net.cpp:100] Creating Layer conv1
I1229 08:30:11.952360  4159 net.cpp:434] conv1 <- data
I1229 08:30:11.952386  4159 net.cpp:408] conv1 -> conv1
I1229 08:30:11.952474  4159 net.cpp:150] Setting up conv1
I1229 08:30:11.952507  4159 net.cpp:157] Top shape: 100 20 24 24 (1152000)
I1229 08:30:11.952530  4159 net.cpp:165] Memory required for data: 4922800
I1229 08:30:11.952565  4159 layer_factory.hpp:77] Creating layer pool1
I1229 08:30:11.952615  4159 net.cpp:100] Creating Layer pool1
I1229 08:30:11.952641  4159 net.cpp:434] pool1 <- conv1
I1229 08:30:11.952667  4159 net.cpp:408] pool1 -> pool1
I1229 08:30:11.952705  4159 net.cpp:150] Setting up pool1
I1229 08:30:11.952733  4159 net.cpp:157] Top shape: 100 20 12 12 (288000)
I1229 08:30:11.952756  4159 net.cpp:165] Memory required for data: 6074800
I1229 08:30:11.952780  4159 layer_factory.hpp:77] Creating layer conv2
I1229 08:30:11.952808  4159 net.cpp:100] Creating Layer conv2
I1229 08:30:11.952833  4159 net.cpp:434] conv2 <- pool1
I1229 08:30:11.952859  4159 net.cpp:408] conv2 -> conv2
I1229 08:30:11.953111  4159 net.cpp:150] Setting up conv2
I1229 08:30:11.953145  4159 net.cpp:157] Top shape: 100 50 8 8 (320000)
I1229 08:30:11.953168  4159 net.cpp:165] Memory required for data: 7354800
I1229 08:30:11.953197  4159 layer_factory.hpp:77] Creating layer pool2
I1229 08:30:11.953263  4159 net.cpp:100] Creating Layer pool2
I1229 08:30:11.953294  4159 net.cpp:434] pool2 <- conv2
I1229 08:30:11.953320  4159 net.cpp:408] pool2 -> pool2
I1229 08:30:11.953351  4159 net.cpp:150] Setting up pool2
I1229 08:30:11.953377  4159 net.cpp:157] Top shape: 100 50 4 4 (80000)
I1229 08:30:11.953400  4159 net.cpp:165] Memory required for data: 7674800
I1229 08:30:11.953423  4159 layer_factory.hpp:77] Creating layer ip1
I1229 08:30:11.953454  4159 net.cpp:100] Creating Layer ip1
I1229 08:30:11.953480  4159 net.cpp:434] ip1 <- pool2
I1229 08:30:11.953505  4159 net.cpp:408] ip1 -> ip1
I1229 08:30:11.956861  4159 net.cpp:150] Setting up ip1
I1229 08:30:11.957449  4159 net.cpp:157] Top shape: 100 500 (50000)
I1229 08:30:11.957475  4159 net.cpp:165] Memory required for data: 7874800
I1229 08:30:11.957505  4159 layer_factory.hpp:77] Creating layer relu1
I1229 08:30:11.957592  4159 net.cpp:100] Creating Layer relu1
I1229 08:30:11.957619  4159 net.cpp:434] relu1 <- ip1
I1229 08:30:11.957644  4159 net.cpp:395] relu1 -> ip1 (in-place)
I1229 08:30:11.957675  4159 net.cpp:150] Setting up relu1
I1229 08:30:11.957725  4159 net.cpp:157] Top shape: 100 500 (50000)
I1229 08:30:11.957777  4159 net.cpp:165] Memory required for data: 8074800
I1229 08:30:11.957818  4159 layer_factory.hpp:77] Creating layer ip2
I1229 08:30:11.957871  4159 net.cpp:100] Creating Layer ip2
I1229 08:30:11.957929  4159 net.cpp:434] ip2 <- ip1
I1229 08:30:11.957976  4159 net.cpp:408] ip2 -> ip2
I1229 08:30:11.958083  4159 net.cpp:150] Setting up ip2
I1229 08:30:11.958230  4159 net.cpp:157] Top shape: 100 10 (1000)
I1229 08:30:11.958276  4159 net.cpp:165] Memory required for data: 8078800
I1229 08:30:11.958319  4159 layer_factory.hpp:77] Creating layer ip2_ip2_0_split
I1229 08:30:11.958362  4159 net.cpp:100] Creating Layer ip2_ip2_0_split
I1229 08:30:11.958406  4159 net.cpp:434] ip2_ip2_0_split <- ip2
I1229 08:30:11.958449  4159 net.cpp:408] ip2_ip2_0_split -> ip2_ip2_0_split_0
I1229 08:30:11.958498  4159 net.cpp:408] ip2_ip2_0_split -> ip2_ip2_0_split_1
I1229 08:30:11.958545  4159 net.cpp:150] Setting up ip2_ip2_0_split
I1229 08:30:11.958600  4159 net.cpp:157] Top shape: 100 10 (1000)
I1229 08:30:11.958642  4159 net.cpp:157] Top shape: 100 10 (1000)
I1229 08:30:11.958678  4159 net.cpp:165] Memory required for data: 8086800
I1229 08:30:11.958710  4159 layer_factory.hpp:77] Creating layer accuracy
I1229 08:30:11.958757  4159 net.cpp:100] Creating Layer accuracy
I1229 08:30:11.958796  4159 net.cpp:434] accuracy <- ip2_ip2_0_split_0
I1229 08:30:11.958844  4159 net.cpp:434] accuracy <- label_mnist_1_split_0
I1229 08:30:11.958890  4159 net.cpp:408] accuracy -> accuracy
I1229 08:30:11.958942  4159 net.cpp:150] Setting up accuracy
I1229 08:30:11.958997  4159 net.cpp:157] Top shape: (1)
I1229 08:30:11.959038  4159 net.cpp:165] Memory required for data: 8086804
I1229 08:30:11.959071  4159 layer_factory.hpp:77] Creating layer loss
I1229 08:30:11.959116  4159 net.cpp:100] Creating Layer loss
I1229 08:30:11.959161  4159 net.cpp:434] loss <- ip2_ip2_0_split_1
I1229 08:30:11.959203  4159 net.cpp:434] loss <- label_mnist_1_split_1
I1229 08:30:11.959287  4159 net.cpp:408] loss -> loss
I1229 08:30:11.959345  4159 layer_factory.hpp:77] Creating layer loss
I1229 08:30:11.959434  4159 net.cpp:150] Setting up loss
I1229 08:30:11.959507  4159 net.cpp:157] Top shape: (1)
I1229 08:30:11.959543  4159 net.cpp:160]     with loss weight 1
I1229 08:30:11.959638  4159 net.cpp:165] Memory required for data: 8086808
I1229 08:30:11.959679  4159 net.cpp:226] loss needs backward computation.
I1229 08:30:11.959713  4159 net.cpp:228] accuracy does not need backward computation.
I1229 08:30:11.959748  4159 net.cpp:226] ip2_ip2_0_split needs backward computation.
I1229 08:30:11.959791  4159 net.cpp:226] ip2 needs backward computation.
I1229 08:30:11.959825  4159 net.cpp:226] relu1 needs backward computation.
I1229 08:30:11.959856  4159 net.cpp:226] ip1 needs backward computation.
I1229 08:30:11.959890  4159 net.cpp:226] pool2 needs backward computation.
I1229 08:30:11.959930  4159 net.cpp:226] conv2 needs backward computation.
I1229 08:30:11.959962  4159 net.cpp:226] pool1 needs backward computation.
I1229 08:30:11.960003  4159 net.cpp:226] conv1 needs backward computation.
I1229 08:30:11.960037  4159 net.cpp:228] label_mnist_1_split does not need backward computation.
I1229 08:30:11.960079  4159 net.cpp:228] mnist does not need backward computation.
I1229 08:30:11.960114  4159 net.cpp:270] This network produces output accuracy
I1229 08:30:11.960155  4159 net.cpp:270] This network produces output loss
I1229 08:30:11.960197  4159 net.cpp:283] Network initialization done.
I1229 08:30:11.965471  4159 caffe.cpp:285] Running for 100 iterations.
I1229 08:30:12.025677  4159 caffe.cpp:308] Batch 0, accuracy = 1
I1229 08:30:12.025766  4159 caffe.cpp:308] Batch 0, loss = 0.00755897
I1229 08:30:12.083183  4159 caffe.cpp:308] Batch 1, accuracy = 0.99
I1229 08:30:12.083266  4159 caffe.cpp:308] Batch 1, loss = 0.0125238
I1229 08:30:12.140435  4159 caffe.cpp:308] Batch 2, accuracy = 0.99
I1229 08:30:12.140517  4159 caffe.cpp:308] Batch 2, loss = 0.0366887
I1229 08:30:12.197609  4159 caffe.cpp:308] Batch 3, accuracy = 0.99
I1229 08:30:12.197691  4159 caffe.cpp:308] Batch 3, loss = 0.0179997
I1229 08:30:12.255146  4159 caffe.cpp:308] Batch 4, accuracy = 0.99
I1229 08:30:12.255257  4159 caffe.cpp:308] Batch 4, loss = 0.042092
I1229 08:30:12.312587  4159 caffe.cpp:308] Batch 5, accuracy = 0.99
I1229 08:30:12.312680  4159 caffe.cpp:308] Batch 5, loss = 0.0517442
I1229 08:30:12.369791  4159 caffe.cpp:308] Batch 6, accuracy = 0.98
I1229 08:30:12.369884  4159 caffe.cpp:308] Batch 6, loss = 0.0612323
I1229 08:30:12.427008  4159 caffe.cpp:308] Batch 7, accuracy = 0.99
I1229 08:30:12.427106  4159 caffe.cpp:308] Batch 7, loss = 0.041436
I1229 08:30:12.484457  4159 caffe.cpp:308] Batch 8, accuracy = 0.99
I1229 08:30:12.484560  4159 caffe.cpp:308] Batch 8, loss = 0.0160605
I1229 08:30:12.541947  4159 caffe.cpp:308] Batch 9, accuracy = 0.98
I1229 08:30:12.542069  4159 caffe.cpp:308] Batch 9, loss = 0.0331129
I1229 08:30:12.599222  4159 caffe.cpp:308] Batch 10, accuracy = 0.98
I1229 08:30:12.599315  4159 caffe.cpp:308] Batch 10, loss = 0.0573284
I1229 08:30:12.656574  4159 caffe.cpp:308] Batch 11, accuracy = 0.98
I1229 08:30:12.656666  4159 caffe.cpp:308] Batch 11, loss = 0.0382567
I1229 08:30:12.714007  4159 caffe.cpp:308] Batch 12, accuracy = 0.96
I1229 08:30:12.714097  4159 caffe.cpp:308] Batch 12, loss = 0.138023
I1229 08:30:12.771430  4159 caffe.cpp:308] Batch 13, accuracy = 0.98
I1229 08:30:12.771523  4159 caffe.cpp:308] Batch 13, loss = 0.0342435
I1229 08:30:12.829116  4159 caffe.cpp:308] Batch 14, accuracy = 1
I1229 08:30:12.829265  4159 caffe.cpp:308] Batch 14, loss = 0.0119491
I1229 08:30:12.887130  4159 caffe.cpp:308] Batch 15, accuracy = 0.97
I1229 08:30:12.887259  4159 caffe.cpp:308] Batch 15, loss = 0.0427386
I1229 08:30:12.944875  4159 caffe.cpp:308] Batch 16, accuracy = 0.99
I1229 08:30:12.944993  4159 caffe.cpp:308] Batch 16, loss = 0.0290488
I1229 08:30:13.002748  4159 caffe.cpp:308] Batch 17, accuracy = 0.99
I1229 08:30:13.002882  4159 caffe.cpp:308] Batch 17, loss = 0.0328682
I1229 08:30:13.060534  4159 caffe.cpp:308] Batch 18, accuracy = 1
I1229 08:30:13.060644  4159 caffe.cpp:308] Batch 18, loss = 0.0121858
I1229 08:30:13.118321  4159 caffe.cpp:308] Batch 19, accuracy = 0.99
I1229 08:30:13.118448  4159 caffe.cpp:308] Batch 19, loss = 0.0610938
I1229 08:30:13.175621  4159 caffe.cpp:308] Batch 20, accuracy = 0.98
I1229 08:30:13.175751  4159 caffe.cpp:308] Batch 20, loss = 0.0924508
I1229 08:30:13.233448  4159 caffe.cpp:308] Batch 21, accuracy = 0.97
I1229 08:30:13.233566  4159 caffe.cpp:308] Batch 21, loss = 0.0909196
I1229 08:30:13.291142  4159 caffe.cpp:308] Batch 22, accuracy = 0.99
I1229 08:30:13.291272  4159 caffe.cpp:308] Batch 22, loss = 0.0139248
I1229 08:30:13.348948  4159 caffe.cpp:308] Batch 23, accuracy = 1
I1229 08:30:13.349058  4159 caffe.cpp:308] Batch 23, loss = 0.0198691
I1229 08:30:13.406636  4159 caffe.cpp:308] Batch 24, accuracy = 0.98
I1229 08:30:13.406746  4159 caffe.cpp:308] Batch 24, loss = 0.0412692
I1229 08:30:13.464308  4159 caffe.cpp:308] Batch 25, accuracy = 0.99
I1229 08:30:13.464418  4159 caffe.cpp:308] Batch 25, loss = 0.0626006
I1229 08:30:13.522017  4159 caffe.cpp:308] Batch 26, accuracy = 0.99
I1229 08:30:13.522138  4159 caffe.cpp:308] Batch 26, loss = 0.129049
I1229 08:30:13.579826  4159 caffe.cpp:308] Batch 27, accuracy = 0.99
I1229 08:30:13.579942  4159 caffe.cpp:308] Batch 27, loss = 0.0203413
I1229 08:30:13.637822  4159 caffe.cpp:308] Batch 28, accuracy = 0.99
I1229 08:30:13.637951  4159 caffe.cpp:308] Batch 28, loss = 0.0473171
I1229 08:30:13.695449  4159 caffe.cpp:308] Batch 29, accuracy = 0.95
I1229 08:30:13.695564  4159 caffe.cpp:308] Batch 29, loss = 0.130175
I1229 08:30:13.753072  4159 caffe.cpp:308] Batch 30, accuracy = 0.99
I1229 08:30:13.753187  4159 caffe.cpp:308] Batch 30, loss = 0.0209849
I1229 08:30:13.810647  4159 caffe.cpp:308] Batch 31, accuracy = 1
I1229 08:30:13.810770  4159 caffe.cpp:308] Batch 31, loss = 0.00194038
I1229 08:30:13.868512  4159 caffe.cpp:308] Batch 32, accuracy = 0.99
I1229 08:30:13.868621  4159 caffe.cpp:308] Batch 32, loss = 0.0168306
I1229 08:30:13.926297  4159 caffe.cpp:308] Batch 33, accuracy = 1
I1229 08:30:13.926412  4159 caffe.cpp:308] Batch 33, loss = 0.00384053
I1229 08:30:13.984012  4159 caffe.cpp:308] Batch 34, accuracy = 0.98
I1229 08:30:13.984127  4159 caffe.cpp:308] Batch 34, loss = 0.0660918
I1229 08:30:14.041839  4159 caffe.cpp:308] Batch 35, accuracy = 0.95
I1229 08:30:14.041949  4159 caffe.cpp:308] Batch 35, loss = 0.118765
I1229 08:30:14.099517  4159 caffe.cpp:308] Batch 36, accuracy = 1
I1229 08:30:14.099625  4159 caffe.cpp:308] Batch 36, loss = 0.00772053
I1229 08:30:14.157325  4159 caffe.cpp:308] Batch 37, accuracy = 0.98
I1229 08:30:14.157438  4159 caffe.cpp:308] Batch 37, loss = 0.0586364
I1229 08:30:14.215265  4159 caffe.cpp:308] Batch 38, accuracy = 1
I1229 08:30:14.215390  4159 caffe.cpp:308] Batch 38, loss = 0.0119813
I1229 08:30:14.273030  4159 caffe.cpp:308] Batch 39, accuracy = 0.99
I1229 08:30:14.273146  4159 caffe.cpp:308] Batch 39, loss = 0.0247472
I1229 08:30:14.330781  4159 caffe.cpp:308] Batch 40, accuracy = 1
I1229 08:30:14.330905  4159 caffe.cpp:308] Batch 40, loss = 0.0181162
I1229 08:30:14.388540  4159 caffe.cpp:308] Batch 41, accuracy = 0.98
I1229 08:30:14.388654  4159 caffe.cpp:308] Batch 41, loss = 0.0517173
I1229 08:30:14.446426  4159 caffe.cpp:308] Batch 42, accuracy = 0.98
I1229 08:30:14.446543  4159 caffe.cpp:308] Batch 42, loss = 0.0362782
I1229 08:30:14.504629  4159 caffe.cpp:308] Batch 43, accuracy = 1
I1229 08:30:14.504768  4159 caffe.cpp:308] Batch 43, loss = 0.00602194
I1229 08:30:14.562206  4159 caffe.cpp:308] Batch 44, accuracy = 1
I1229 08:30:14.562331  4159 caffe.cpp:308] Batch 44, loss = 0.0127355
I1229 08:30:14.619529  4159 caffe.cpp:308] Batch 45, accuracy = 0.99
I1229 08:30:14.619657  4159 caffe.cpp:308] Batch 45, loss = 0.0432564
I1229 08:30:14.676769  4159 caffe.cpp:308] Batch 46, accuracy = 1
I1229 08:30:14.676884  4159 caffe.cpp:308] Batch 46, loss = 0.00795186
I1229 08:30:14.734272  4159 caffe.cpp:308] Batch 47, accuracy = 1
I1229 08:30:14.734398  4159 caffe.cpp:308] Batch 47, loss = 0.00829059
I1229 08:30:14.791604  4159 caffe.cpp:308] Batch 48, accuracy = 0.96
I1229 08:30:14.791720  4159 caffe.cpp:308] Batch 48, loss = 0.062666
I1229 08:30:14.849009  4159 caffe.cpp:308] Batch 49, accuracy = 1
I1229 08:30:14.849145  4159 caffe.cpp:308] Batch 49, loss = 0.012306
I1229 08:30:14.906517  4159 caffe.cpp:308] Batch 50, accuracy = 1
I1229 08:30:14.906632  4159 caffe.cpp:308] Batch 50, loss = 0.00024388
I1229 08:30:14.963975  4159 caffe.cpp:308] Batch 51, accuracy = 1
I1229 08:30:14.964099  4159 caffe.cpp:308] Batch 51, loss = 0.00388429
I1229 08:30:15.021663  4159 caffe.cpp:308] Batch 52, accuracy = 1
I1229 08:30:15.021785  4159 caffe.cpp:308] Batch 52, loss = 0.00335451
I1229 08:30:15.079262  4159 caffe.cpp:308] Batch 53, accuracy = 1
I1229 08:30:15.079386  4159 caffe.cpp:308] Batch 53, loss = 0.00241596
I1229 08:30:15.136984  4159 caffe.cpp:308] Batch 54, accuracy = 0.99
I1229 08:30:15.137111  4159 caffe.cpp:308] Batch 54, loss = 0.0142214
I1229 08:30:15.194694  4159 caffe.cpp:308] Batch 55, accuracy = 1
I1229 08:30:15.194811  4159 caffe.cpp:308] Batch 55, loss = 0.000404293
I1229 08:30:15.252305  4159 caffe.cpp:308] Batch 56, accuracy = 1
I1229 08:30:15.252424  4159 caffe.cpp:308] Batch 56, loss = 0.00623298
I1229 08:30:15.310086  4159 caffe.cpp:308] Batch 57, accuracy = 0.99
I1229 08:30:15.310197  4159 caffe.cpp:308] Batch 57, loss = 0.0176835
I1229 08:30:15.367950  4159 caffe.cpp:308] Batch 58, accuracy = 1
I1229 08:30:15.368067  4159 caffe.cpp:308] Batch 58, loss = 0.00307279
I1229 08:30:15.425412  4159 caffe.cpp:308] Batch 59, accuracy = 0.96
I1229 08:30:15.425529  4159 caffe.cpp:308] Batch 59, loss = 0.0957446
I1229 08:30:15.482833  4159 caffe.cpp:308] Batch 60, accuracy = 1
I1229 08:30:15.482949  4159 caffe.cpp:308] Batch 60, loss = 0.0111263
I1229 08:30:15.540472  4159 caffe.cpp:308] Batch 61, accuracy = 1
I1229 08:30:15.540583  4159 caffe.cpp:308] Batch 61, loss = 0.00620457
I1229 08:30:15.598043  4159 caffe.cpp:308] Batch 62, accuracy = 1
I1229 08:30:15.598151  4159 caffe.cpp:308] Batch 62, loss = 3.23028e-05
I1229 08:30:15.655606  4159 caffe.cpp:308] Batch 63, accuracy = 1
I1229 08:30:15.655725  4159 caffe.cpp:308] Batch 63, loss = 0.000154275
I1229 08:30:15.713317  4159 caffe.cpp:308] Batch 64, accuracy = 1
I1229 08:30:15.713426  4159 caffe.cpp:308] Batch 64, loss = 0.000742791
I1229 08:30:15.770719  4159 caffe.cpp:308] Batch 65, accuracy = 0.95
I1229 08:30:15.770843  4159 caffe.cpp:308] Batch 65, loss = 0.11012
I1229 08:30:15.828063  4159 caffe.cpp:308] Batch 66, accuracy = 0.98
I1229 08:30:15.828179  4159 caffe.cpp:308] Batch 66, loss = 0.0724991
I1229 08:30:15.886113  4159 caffe.cpp:308] Batch 67, accuracy = 0.99
I1229 08:30:15.886236  4159 caffe.cpp:308] Batch 67, loss = 0.0278719
I1229 08:30:15.943961  4159 caffe.cpp:308] Batch 68, accuracy = 1
I1229 08:30:15.944084  4159 caffe.cpp:308] Batch 68, loss = 0.00463332
I1229 08:30:16.001819  4159 caffe.cpp:308] Batch 69, accuracy = 1
I1229 08:30:16.001940  4159 caffe.cpp:308] Batch 69, loss = 0.000807585
I1229 08:30:16.059298  4159 caffe.cpp:308] Batch 70, accuracy = 1
I1229 08:30:16.059428  4159 caffe.cpp:308] Batch 70, loss = 0.000793165
I1229 08:30:16.116801  4159 caffe.cpp:308] Batch 71, accuracy = 1
I1229 08:30:16.116930  4159 caffe.cpp:308] Batch 71, loss = 0.00123211
I1229 08:30:16.174362  4159 caffe.cpp:308] Batch 72, accuracy = 1
I1229 08:30:16.174492  4159 caffe.cpp:308] Batch 72, loss = 0.012733
I1229 08:30:16.231920  4159 caffe.cpp:308] Batch 73, accuracy = 1
I1229 08:30:16.232060  4159 caffe.cpp:308] Batch 73, loss = 7.07769e-05
I1229 08:30:16.289247  4159 caffe.cpp:308] Batch 74, accuracy = 1
I1229 08:30:16.289373  4159 caffe.cpp:308] Batch 74, loss = 0.00166314
I1229 08:30:16.346583  4159 caffe.cpp:308] Batch 75, accuracy = 1
I1229 08:30:16.346707  4159 caffe.cpp:308] Batch 75, loss = 0.00133687
I1229 08:30:16.404086  4159 caffe.cpp:308] Batch 76, accuracy = 1
I1229 08:30:16.404251  4159 caffe.cpp:308] Batch 76, loss = 0.00015642
I1229 08:30:16.461724  4159 caffe.cpp:308] Batch 77, accuracy = 1
I1229 08:30:16.461879  4159 caffe.cpp:308] Batch 77, loss = 0.000267936
I1229 08:30:16.519292  4159 caffe.cpp:308] Batch 78, accuracy = 1
I1229 08:30:16.519414  4159 caffe.cpp:308] Batch 78, loss = 0.00102958
I1229 08:30:16.576902  4159 caffe.cpp:308] Batch 79, accuracy = 1
I1229 08:30:16.577018  4159 caffe.cpp:308] Batch 79, loss = 0.0019392
I1229 08:30:16.634402  4159 caffe.cpp:308] Batch 80, accuracy = 1
I1229 08:30:16.634522  4159 caffe.cpp:308] Batch 80, loss = 0.00750751
I1229 08:30:16.691875  4159 caffe.cpp:308] Batch 81, accuracy = 1
I1229 08:30:16.691999  4159 caffe.cpp:308] Batch 81, loss = 0.00151326
I1229 08:30:16.751351  4159 caffe.cpp:308] Batch 82, accuracy = 1
I1229 08:30:16.751477  4159 caffe.cpp:308] Batch 82, loss = 0.00455984
I1229 08:30:16.808917  4159 caffe.cpp:308] Batch 83, accuracy = 1
I1229 08:30:16.809056  4159 caffe.cpp:308] Batch 83, loss = 0.0110726
I1229 08:30:16.867089  4159 caffe.cpp:308] Batch 84, accuracy = 0.99
I1229 08:30:16.867207  4159 caffe.cpp:308] Batch 84, loss = 0.0229276
I1229 08:30:16.924979  4159 caffe.cpp:308] Batch 85, accuracy = 0.99
I1229 08:30:16.925104  4159 caffe.cpp:308] Batch 85, loss = 0.0181782
I1229 08:30:16.982977  4159 caffe.cpp:308] Batch 86, accuracy = 1
I1229 08:30:16.983095  4159 caffe.cpp:308] Batch 86, loss = 0.000105467
I1229 08:30:17.040793  4159 caffe.cpp:308] Batch 87, accuracy = 1
I1229 08:30:17.040918  4159 caffe.cpp:308] Batch 87, loss = 8.85461e-05
I1229 08:30:17.098577  4159 caffe.cpp:308] Batch 88, accuracy = 1
I1229 08:30:17.098690  4159 caffe.cpp:308] Batch 88, loss = 1.90067e-05
I1229 08:30:17.156373  4159 caffe.cpp:308] Batch 89, accuracy = 1
I1229 08:30:17.156487  4159 caffe.cpp:308] Batch 89, loss = 2.68037e-05
I1229 08:30:17.214165  4159 caffe.cpp:308] Batch 90, accuracy = 0.97
I1229 08:30:17.214289  4159 caffe.cpp:308] Batch 90, loss = 0.0997072
I1229 08:30:17.272058  4159 caffe.cpp:308] Batch 91, accuracy = 1
I1229 08:30:17.272197  4159 caffe.cpp:308] Batch 91, loss = 1.74496e-05
I1229 08:30:17.329752  4159 caffe.cpp:308] Batch 92, accuracy = 1
I1229 08:30:17.329874  4159 caffe.cpp:308] Batch 92, loss = 0.000416481
I1229 08:30:17.387363  4159 caffe.cpp:308] Batch 93, accuracy = 1
I1229 08:30:17.387485  4159 caffe.cpp:308] Batch 93, loss = 0.00121327
I1229 08:30:17.444933  4159 caffe.cpp:308] Batch 94, accuracy = 1
I1229 08:30:17.445055  4159 caffe.cpp:308] Batch 94, loss = 0.000482124
I1229 08:30:17.502532  4159 caffe.cpp:308] Batch 95, accuracy = 1
I1229 08:30:17.502652  4159 caffe.cpp:308] Batch 95, loss = 0.00230492
I1229 08:30:17.560281  4159 caffe.cpp:308] Batch 96, accuracy = 0.99
I1229 08:30:17.560406  4159 caffe.cpp:308] Batch 96, loss = 0.0368026
I1229 08:30:17.617797  4159 caffe.cpp:308] Batch 97, accuracy = 0.98
I1229 08:30:17.617933  4159 caffe.cpp:308] Batch 97, loss = 0.0985361
I1229 08:30:17.674965  4159 caffe.cpp:308] Batch 98, accuracy = 1
I1229 08:30:17.675092  4159 caffe.cpp:308] Batch 98, loss = 0.00306735
I1229 08:30:17.732564  4159 caffe.cpp:308] Batch 99, accuracy = 1
I1229 08:30:17.732704  4159 caffe.cpp:308] Batch 99, loss = 0.0103088
I1229 08:30:17.732751  4159 caffe.cpp:313] Loss: 0.0273851
I1229 08:30:17.732795  4159 caffe.cpp:325] accuracy = 0.9913
I1229 08:30:17.732837  4159 caffe.cpp:325] loss = 0.0273851 (* 1 = 0.0273851 loss)
root@ip-172-30-0-251:/caffe#