Comments (7)
It's hard to say what went wrong. If you have spark webgui, you can look at executor thread dump (or go to your executors and "jstack your_spark_pid"). it may show why the program was stuck.
from caffeonspark.
Please attach the log file into this issue. We need more context for debug.
from caffeonspark.
com.yahoo.ml.jcaffe.CaffeNet.sync(Native Method)
com.yahoo.ml.caffe.CaffeProcessor.sync(CaffeProcessor.scala:156)
com.yahoo.ml.caffe.CaffeOnSpark$$anonfun$6.apply(CaffeOnSpark.scala:176)
com.yahoo.ml.caffe.CaffeOnSpark$$anonfun$6.apply(CaffeOnSpark.scala:171)
org.apache.spark.rdd.RDD$$anonfun$mapPartitions$1$$anonfun$apply$20.apply(RDD.scala:710)
org.apache.spark.rdd.RDD$$anonfun$mapPartitions$1$$anonfun$apply$20.apply(RDD.scala:710)
org.apache.spark.rdd.MapPartitionsRDD.compute(MapPartitionsRDD.scala:38)
org.apache.spark.rdd.RDD.computeOrReadCheckpoint(RDD.scala:306)
org.apache.spark.CacheManager.getOrCompute(CacheManager.scala:69)
org.apache.spark.rdd.RDD.iterator(RDD.scala:268)
org.apache.spark.scheduler.ResultTask.runTask(ResultTask.scala:66)
org.apache.spark.scheduler.Task.run(Task.scala:89)
org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:213)
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1145)
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:615)
java.lang.Thread.run(Thread.java:724)
from caffeonspark.
@chengdianxuezi Can you post the WHOLE log (not that small section)? You could get the complete log via "yarn logs -applicationID ..."
from caffeonspark.
Container: container_1458635993280_0010_01_000006 on dlgpu10.ai.bjcc.qihoo.net_54591
LogType:stderr
Log Upload Time:23-Mar-2016 16:53:58
LogLength:31209
Log Contents:
SLF4J: Class path contains multiple SLF4J bindings.
SLF4J: Found binding in [jar:file:/dev/hadoop/nm-local-dir/usercache/hadoop/filecache/11/spark-assembly-1.6.0-hadoop2.6.4-U4.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: Found binding in [jar:file:/home/hadoop/software/hadoop-2.6.4/share/hadoop/common/lib/slf4j-log4j12-1.7.5.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: See http://www.slf4j.org/codes.html#multiple_bindings for an explanation.
SLF4J: Actual binding is of type [org.slf4j.impl.Log4jLoggerFactory]
16/03/23 16:51:14 INFO executor.CoarseGrainedExecutorBackend: Registered signal handlers for [TERM, HUP, INT]
16/03/23 16:51:14 INFO yarn.YarnSparkHadoopUtil: Set ugi in YarnConf to: hadoop,hadoop
16/03/23 16:51:14 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
16/03/23 16:51:14 INFO yarn.YarnSparkHadoopUtil: running as user: hadoop
16/03/23 16:51:14 INFO spark.SecurityManager: Changing view acls to: hadoop
16/03/23 16:51:14 INFO spark.SecurityManager: Changing modify acls to: hadoop
16/03/23 16:51:14 INFO spark.SecurityManager: SecurityManager: authentication disabled; ui acls disabled; users with view permissions: Set(hadoop); users with modify permissions: Set(hadoop)
16/03/23 16:51:15 INFO spark.SecurityManager: Changing view acls to: hadoop
16/03/23 16:51:15 INFO spark.SecurityManager: Changing modify acls to: hadoop
16/03/23 16:51:15 INFO spark.SecurityManager: SecurityManager: authentication disabled; ui acls disabled; users with view permissions: Set(hadoop); users with modify permissions: Set(hadoop)
16/03/23 16:51:16 INFO slf4j.Slf4jLogger: Slf4jLogger started
16/03/23 16:51:16 INFO Remoting: Starting remoting
16/03/23 16:51:16 INFO Remoting: Remoting started; listening on addresses :[akka.tcp://[email protected]:54395]
16/03/23 16:51:16 INFO util.Utils: Successfully started service 'sparkExecutorActorSystem' on port 54395.
16/03/23 16:51:16 INFO storage.DiskBlockManager: Created local directory at /home/hadoop/yarn/nm-local-dir/usercache/hadoop/appcache/application_1458635993280_0010/blockmgr-263aa769-42b6-40d7-8246-276962488419
16/03/23 16:51:16 INFO storage.DiskBlockManager: Created local directory at /dev/hadoop/nm-local-dir/usercache/hadoop/appcache/application_1458635993280_0010/blockmgr-3db9d9f0-d30e-4dc4-99d7-4699802069b3
16/03/23 16:51:16 INFO storage.DiskBlockManager: Created local directory at /run/hadoop/nm-local-dir/usercache/hadoop/appcache/application_1458635993280_0010/blockmgr-79eb8e7e-19b9-448f-bfbe-b2ad11e47af2
16/03/23 16:51:16 INFO storage.MemoryStore: MemoryStore started with capacity 24.9 GB
16/03/23 16:51:16 INFO yarn.YarnSparkHadoopUtil: Set ugi in YarnConf to: hadoop,hadoop
16/03/23 16:51:16 INFO executor.CoarseGrainedExecutorBackend: Connecting to driver: spark://[email protected]:45855
16/03/23 16:51:16 INFO executor.CoarseGrainedExecutorBackend: Successfully registered with driver
16/03/23 16:51:16 INFO executor.Executor: Starting executor ID 4 on host dlgpu10.ai.bjcc.qihoo.net
16/03/23 16:51:16 INFO util.Utils: Successfully started service 'org.apache.spark.network.netty.NettyBlockTransferService' on port 46213.
16/03/23 16:51:16 INFO netty.NettyBlockTransferService: Server created on 46213
16/03/23 16:51:16 INFO storage.BlockManager: external shuffle service port = 7337
16/03/23 16:51:16 INFO storage.BlockManagerMaster: Trying to register BlockManager
16/03/23 16:51:16 INFO storage.BlockManagerMaster: Registered BlockManager
16/03/23 16:51:16 INFO storage.BlockManager: Registering executor with local external shuffle service.
16/03/23 16:51:19 INFO executor.CoarseGrainedExecutorBackend: Got assigned task 4
16/03/23 16:51:19 INFO executor.Executor: Running task 4.0 in stage 0.0 (TID 4)
16/03/23 16:51:19 INFO broadcast.TorrentBroadcast: Started reading broadcast variable 0
16/03/23 16:51:20 INFO storage.MemoryStore: Block broadcast_0_piece0 stored as bytes in memory (estimated size 2.1 KB, free 2.1 KB)
16/03/23 16:51:20 INFO broadcast.TorrentBroadcast: Reading broadcast variable 0 took 296 ms
16/03/23 16:51:20 INFO storage.MemoryStore: Block broadcast_0 stored as values in memory (estimated size 3.2 KB, free 5.3 KB)
16/03/23 16:51:20 INFO caffe.CaffeProcessor: my rank is 4
16/03/23 16:51:20 INFO caffe.LMDB: Batch size:64
WARNING: Logging before InitGoogleLogging() is written to STDERR
I0323 16:51:27.078397 26145 CaffeNet.cpp:78] set root solver device id to 0
I0323 16:51:27.259912 26145 solver.cpp:48] Initializing solver from parameters:
test_iter: 1
test_interval: 10001
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: 10001
snapshot_prefix: "mnist_lenet"
solver_mode: GPU
device_id: 0
net_param {
name: "LeNet"
layer {
name: "data"
type: "MemoryData"
top: "data"
top: "label"
include {
phase: TRAIN
}
memory_data_param {
batch_size: 64
channels: 1
height: 28
width: 28
share_in_parallel: false
source: "hdfs:///projects/machine_learning/image_dataset/mnist_train_lmdb/"
}
source_class: "com.yahoo.ml.caffe.LMDB"
}
layer {
name: "data"
type: "MemoryData"
top: "data"
top: "label"
include {
phase: TEST
}
memory_data_param {
batch_size: 100
channels: 1
height: 28
width: 28
share_in_parallel: false
source: "hdfs:///projects/machine_learning/image_dataset/mnist_test_lmdb/"
}
source_class: "com.yahoo.ml.caffe.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"
}
}
test_initialization: false
I0323 16:51:27.260068 26145 solver.cpp:86] Creating training net specified in net_param.
I0323 16:51:27.260135 26145 net.cpp:322] The NetState phase (0) differed from the phase (1) specified by a rule in layer data
I0323 16:51:27.260150 26145 net.cpp:322] The NetState phase (0) differed from the phase (1) specified by a rule in layer accuracy
I0323 16:51:27.260231 26145 net.cpp:49] Initializing net from parameters:
name: "LeNet"
state {
phase: TRAIN
}
layer {
name: "data"
type: "MemoryData"
top: "data"
top: "label"
include {
phase: TRAIN
}
memory_data_param {
batch_size: 64
channels: 1
height: 28
width: 28
share_in_parallel: false
source: "hdfs:///projects/machine_learning/image_dataset/mnist_train_lmdb/"
}
source_class: "com.yahoo.ml.caffe.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"
}
I0323 16:51:27.260318 26145 layer_factory.hpp:77] Creating layer data
I0323 16:51:27.260342 26145 net.cpp:106] Creating Layer data
I0323 16:51:27.260352 26145 net.cpp:411] data -> data
I0323 16:51:27.260378 26145 net.cpp:411] data -> label
I0323 16:51:27.264107 26145 net.cpp:150] Setting up data
I0323 16:51:27.264132 26145 net.cpp:157] Top shape: 64 1 28 28 (50176)
I0323 16:51:27.264139 26145 net.cpp:157] Top shape: 64 (64)
I0323 16:51:27.264143 26145 net.cpp:165] Memory required for data: 200960
I0323 16:51:27.264153 26145 layer_factory.hpp:77] Creating layer conv1
I0323 16:51:27.264171 26145 net.cpp:106] Creating Layer conv1
I0323 16:51:27.264179 26145 net.cpp:454] conv1 <- data
I0323 16:51:27.264192 26145 net.cpp:411] conv1 -> conv1
I0323 16:51:27.266789 26145 net.cpp:150] Setting up conv1
I0323 16:51:27.266804 26145 net.cpp:157] Top shape: 64 20 24 24 (737280)
I0323 16:51:27.266808 26145 net.cpp:165] Memory required for data: 3150080
I0323 16:51:27.266825 26145 layer_factory.hpp:77] Creating layer pool1
I0323 16:51:27.266836 26145 net.cpp:106] Creating Layer pool1
I0323 16:51:27.266841 26145 net.cpp:454] pool1 <- conv1
I0323 16:51:27.266847 26145 net.cpp:411] pool1 -> pool1
I0323 16:51:27.266974 26145 net.cpp:150] Setting up pool1
I0323 16:51:27.266983 26145 net.cpp:157] Top shape: 64 20 12 12 (184320)
I0323 16:51:27.266988 26145 net.cpp:165] Memory required for data: 3887360
I0323 16:51:27.266991 26145 layer_factory.hpp:77] Creating layer conv2
I0323 16:51:27.267002 26145 net.cpp:106] Creating Layer conv2
I0323 16:51:27.267006 26145 net.cpp:454] conv2 <- pool1
I0323 16:51:27.267016 26145 net.cpp:411] conv2 -> conv2
I0323 16:51:27.267971 26145 net.cpp:150] Setting up conv2
I0323 16:51:27.267982 26145 net.cpp:157] Top shape: 64 50 8 8 (204800)
I0323 16:51:27.267985 26145 net.cpp:165] Memory required for data: 4706560
I0323 16:51:27.267994 26145 layer_factory.hpp:77] Creating layer pool2
I0323 16:51:27.268008 26145 net.cpp:106] Creating Layer pool2
I0323 16:51:27.268013 26145 net.cpp:454] pool2 <- conv2
I0323 16:51:27.268018 26145 net.cpp:411] pool2 -> pool2
I0323 16:51:27.268118 26145 net.cpp:150] Setting up pool2
I0323 16:51:27.268129 26145 net.cpp:157] Top shape: 64 50 4 4 (51200)
I0323 16:51:27.268133 26145 net.cpp:165] Memory required for data: 4911360
I0323 16:51:27.268137 26145 layer_factory.hpp:77] Creating layer ip1
I0323 16:51:27.268149 26145 net.cpp:106] Creating Layer ip1
I0323 16:51:27.268154 26145 net.cpp:454] ip1 <- pool2
I0323 16:51:27.268159 26145 net.cpp:411] ip1 -> ip1
I0323 16:51:27.274052 26145 net.cpp:150] Setting up ip1
I0323 16:51:27.274067 26145 net.cpp:157] Top shape: 64 500 (32000)
I0323 16:51:27.274071 26145 net.cpp:165] Memory required for data: 5039360
I0323 16:51:27.274081 26145 layer_factory.hpp:77] Creating layer relu1
I0323 16:51:27.274091 26145 net.cpp:106] Creating Layer relu1
I0323 16:51:27.274096 26145 net.cpp:454] relu1 <- ip1
I0323 16:51:27.274103 26145 net.cpp:397] relu1 -> ip1 (in-place)
I0323 16:51:27.274114 26145 net.cpp:150] Setting up relu1
I0323 16:51:27.274121 26145 net.cpp:157] Top shape: 64 500 (32000)
I0323 16:51:27.274123 26145 net.cpp:165] Memory required for data: 5167360
I0323 16:51:27.274127 26145 layer_factory.hpp:77] Creating layer ip2
I0323 16:51:27.274133 26145 net.cpp:106] Creating Layer ip2
I0323 16:51:27.274137 26145 net.cpp:454] ip2 <- ip1
I0323 16:51:27.274145 26145 net.cpp:411] ip2 -> ip2
I0323 16:51:27.275876 26145 net.cpp:150] Setting up ip2
I0323 16:51:27.275890 26145 net.cpp:157] Top shape: 64 10 (640)
I0323 16:51:27.275894 26145 net.cpp:165] Memory required for data: 5169920
I0323 16:51:27.275902 26145 layer_factory.hpp:77] Creating layer loss
I0323 16:51:27.275916 26145 net.cpp:106] Creating Layer loss
I0323 16:51:27.275920 26145 net.cpp:454] loss <- ip2
I0323 16:51:27.275930 26145 net.cpp:454] loss <- label
I0323 16:51:27.275941 26145 net.cpp:411] loss -> loss
I0323 16:51:27.275959 26145 layer_factory.hpp:77] Creating layer loss
I0323 16:51:27.276264 26145 net.cpp:150] Setting up loss
I0323 16:51:27.276273 26145 net.cpp:157] Top shape: (1)
I0323 16:51:27.276278 26145 net.cpp:160] with loss weight 1
I0323 16:51:27.276295 26145 net.cpp:165] Memory required for data: 5169924
I0323 16:51:27.276299 26145 net.cpp:226] loss needs backward computation.
I0323 16:51:27.276304 26145 net.cpp:226] ip2 needs backward computation.
I0323 16:51:27.276307 26145 net.cpp:226] relu1 needs backward computation.
I0323 16:51:27.276311 26145 net.cpp:226] ip1 needs backward computation.
I0323 16:51:27.276314 26145 net.cpp:226] pool2 needs backward computation.
I0323 16:51:27.276319 26145 net.cpp:226] conv2 needs backward computation.
I0323 16:51:27.276324 26145 net.cpp:226] pool1 needs backward computation.
I0323 16:51:27.276329 26145 net.cpp:226] conv1 needs backward computation.
I0323 16:51:27.276332 26145 net.cpp:228] data does not need backward computation.
I0323 16:51:27.276336 26145 net.cpp:270] This network produces output loss
I0323 16:51:27.276346 26145 net.cpp:283] Network initialization done.
I0323 16:51:27.276401 26145 solver.cpp:181] Creating test net (#0) specified by net_param
I0323 16:51:27.276422 26145 net.cpp:322] The NetState phase (1) differed from the phase (0) specified by a rule in layer data
I0323 16:51:27.276518 26145 net.cpp:49] Initializing net from parameters:
name: "LeNet"
state {
phase: TEST
}
layer {
name: "data"
type: "MemoryData"
top: "data"
top: "label"
include {
phase: TEST
}
memory_data_param {
batch_size: 100
channels: 1
height: 28
width: 28
share_in_parallel: false
source: "hdfs:///projects/machine_learning/image_dataset/mnist_test_lmdb/"
}
source_class: "com.yahoo.ml.caffe.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"
}
I0323 16:51:27.276599 26145 layer_factory.hpp:77] Creating layer data
I0323 16:51:27.276609 26145 net.cpp:106] Creating Layer data
I0323 16:51:27.276615 26145 net.cpp:411] data -> data
I0323 16:51:27.276623 26145 net.cpp:411] data -> label
I0323 16:51:27.278290 26145 net.cpp:150] Setting up data
I0323 16:51:27.278304 26145 net.cpp:157] Top shape: 100 1 28 28 (78400)
I0323 16:51:27.278311 26145 net.cpp:157] Top shape: 100 (100)
I0323 16:51:27.278313 26145 net.cpp:165] Memory required for data: 314000
I0323 16:51:27.278318 26145 layer_factory.hpp:77] Creating layer label_data_1_split
I0323 16:51:27.278328 26145 net.cpp:106] Creating Layer label_data_1_split
I0323 16:51:27.278332 26145 net.cpp:454] label_data_1_split <- label
I0323 16:51:27.278342 26145 net.cpp:411] label_data_1_split -> label_data_1_split_0
I0323 16:51:27.278349 26145 net.cpp:411] label_data_1_split -> label_data_1_split_1
I0323 16:51:27.278456 26145 net.cpp:150] Setting up label_data_1_split
I0323 16:51:27.278465 26145 net.cpp:157] Top shape: 100 (100)
I0323 16:51:27.278470 26145 net.cpp:157] Top shape: 100 (100)
I0323 16:51:27.278472 26145 net.cpp:165] Memory required for data: 314800
I0323 16:51:27.278476 26145 layer_factory.hpp:77] Creating layer conv1
I0323 16:51:27.278488 26145 net.cpp:106] Creating Layer conv1
I0323 16:51:27.278492 26145 net.cpp:454] conv1 <- data
I0323 16:51:27.278501 26145 net.cpp:411] conv1 -> conv1
I0323 16:51:27.279278 26145 net.cpp:150] Setting up conv1
I0323 16:51:27.279289 26145 net.cpp:157] Top shape: 100 20 24 24 (1152000)
I0323 16:51:27.279294 26145 net.cpp:165] Memory required for data: 4922800
I0323 16:51:27.279305 26145 layer_factory.hpp:77] Creating layer pool1
I0323 16:51:27.279311 26145 net.cpp:106] Creating Layer pool1
I0323 16:51:27.279315 26145 net.cpp:454] pool1 <- conv1
I0323 16:51:27.279321 26145 net.cpp:411] pool1 -> pool1
I0323 16:51:27.279425 26145 net.cpp:150] Setting up pool1
I0323 16:51:27.279433 26145 net.cpp:157] Top shape: 100 20 12 12 (288000)
I0323 16:51:27.279436 26145 net.cpp:165] Memory required for data: 6074800
I0323 16:51:27.279440 26145 layer_factory.hpp:77] Creating layer conv2
I0323 16:51:27.279450 26145 net.cpp:106] Creating Layer conv2
I0323 16:51:27.279454 26145 net.cpp:454] conv2 <- pool1
I0323 16:51:27.279464 26145 net.cpp:411] conv2 -> conv2
I0323 16:51:27.280467 26145 net.cpp:150] Setting up conv2
I0323 16:51:27.280479 26145 net.cpp:157] Top shape: 100 50 8 8 (320000)
I0323 16:51:27.280483 26145 net.cpp:165] Memory required for data: 7354800
I0323 16:51:27.280491 26145 layer_factory.hpp:77] Creating layer pool2
I0323 16:51:27.280498 26145 net.cpp:106] Creating Layer pool2
I0323 16:51:27.280503 26145 net.cpp:454] pool2 <- conv2
I0323 16:51:27.280508 26145 net.cpp:411] pool2 -> pool2
I0323 16:51:27.280607 26145 net.cpp:150] Setting up pool2
I0323 16:51:27.280616 26145 net.cpp:157] Top shape: 100 50 4 4 (80000)
I0323 16:51:27.280622 26145 net.cpp:165] Memory required for data: 7674800
I0323 16:51:27.280624 26145 layer_factory.hpp:77] Creating layer ip1
I0323 16:51:27.280634 26145 net.cpp:106] Creating Layer ip1
I0323 16:51:27.280638 26145 net.cpp:454] ip1 <- pool2
I0323 16:51:27.280643 26145 net.cpp:411] ip1 -> ip1
I0323 16:51:27.285825 26145 net.cpp:150] Setting up ip1
I0323 16:51:27.285840 26145 net.cpp:157] Top shape: 100 500 (50000)
I0323 16:51:27.285845 26145 net.cpp:165] Memory required for data: 7874800
I0323 16:51:27.285856 26145 layer_factory.hpp:77] Creating layer relu1
I0323 16:51:27.285863 26145 net.cpp:106] Creating Layer relu1
I0323 16:51:27.285867 26145 net.cpp:454] relu1 <- ip1
I0323 16:51:27.285873 26145 net.cpp:397] relu1 -> ip1 (in-place)
I0323 16:51:27.285881 26145 net.cpp:150] Setting up relu1
I0323 16:51:27.285887 26145 net.cpp:157] Top shape: 100 500 (50000)
I0323 16:51:27.285889 26145 net.cpp:165] Memory required for data: 8074800
I0323 16:51:27.285892 26145 layer_factory.hpp:77] Creating layer ip2
I0323 16:51:27.285903 26145 net.cpp:106] Creating Layer ip2
I0323 16:51:27.285907 26145 net.cpp:454] ip2 <- ip1
I0323 16:51:27.285912 26145 net.cpp:411] ip2 -> ip2
I0323 16:51:27.286298 26145 net.cpp:150] Setting up ip2
I0323 16:51:27.286308 26145 net.cpp:157] Top shape: 100 10 (1000)
I0323 16:51:27.286311 26145 net.cpp:165] Memory required for data: 8078800
I0323 16:51:27.286319 26145 layer_factory.hpp:77] Creating layer ip2_ip2_0_split
I0323 16:51:27.286324 26145 net.cpp:106] Creating Layer ip2_ip2_0_split
I0323 16:51:27.286327 26145 net.cpp:454] ip2_ip2_0_split <- ip2
I0323 16:51:27.286335 26145 net.cpp:411] ip2_ip2_0_split -> ip2_ip2_0_split_0
I0323 16:51:27.286342 26145 net.cpp:411] ip2_ip2_0_split -> ip2_ip2_0_split_1
I0323 16:51:27.286450 26145 net.cpp:150] Setting up ip2_ip2_0_split
I0323 16:51:27.286458 26145 net.cpp:157] Top shape: 100 10 (1000)
I0323 16:51:27.286463 26145 net.cpp:157] Top shape: 100 10 (1000)
I0323 16:51:27.286466 26145 net.cpp:165] Memory required for data: 8086800
I0323 16:51:27.286469 26145 layer_factory.hpp:77] Creating layer accuracy
I0323 16:51:27.286481 26145 net.cpp:106] Creating Layer accuracy
I0323 16:51:27.286485 26145 net.cpp:454] accuracy <- ip2_ip2_0_split_0
I0323 16:51:27.286490 26145 net.cpp:454] accuracy <- label_data_1_split_0
I0323 16:51:27.286497 26145 net.cpp:411] accuracy -> accuracy
I0323 16:51:27.286509 26145 net.cpp:150] Setting up accuracy
I0323 16:51:27.286514 26145 net.cpp:157] Top shape: (1)
I0323 16:51:27.286516 26145 net.cpp:165] Memory required for data: 8086804
I0323 16:51:27.286520 26145 layer_factory.hpp:77] Creating layer loss
I0323 16:51:27.286525 26145 net.cpp:106] Creating Layer loss
I0323 16:51:27.286530 26145 net.cpp:454] loss <- ip2_ip2_0_split_1
I0323 16:51:27.286533 26145 net.cpp:454] loss <- label_data_1_split_1
I0323 16:51:27.286540 26145 net.cpp:411] loss -> loss
I0323 16:51:27.286546 26145 layer_factory.hpp:77] Creating layer loss
I0323 16:51:27.286836 26145 net.cpp:150] Setting up loss
I0323 16:51:27.286844 26145 net.cpp:157] Top shape: (1)
I0323 16:51:27.286849 26145 net.cpp:160] with loss weight 1
I0323 16:51:27.286855 26145 net.cpp:165] Memory required for data: 8086808
I0323 16:51:27.286859 26145 net.cpp:226] loss needs backward computation.
I0323 16:51:27.286862 26145 net.cpp:228] accuracy does not need backward computation.
I0323 16:51:27.286867 26145 net.cpp:226] ip2_ip2_0_split needs backward computation.
I0323 16:51:27.286870 26145 net.cpp:226] ip2 needs backward computation.
I0323 16:51:27.286875 26145 net.cpp:226] relu1 needs backward computation.
I0323 16:51:27.286877 26145 net.cpp:226] ip1 needs backward computation.
I0323 16:51:27.286881 26145 net.cpp:226] pool2 needs backward computation.
I0323 16:51:27.286885 26145 net.cpp:226] conv2 needs backward computation.
I0323 16:51:27.286888 26145 net.cpp:226] pool1 needs backward computation.
I0323 16:51:27.286892 26145 net.cpp:226] conv1 needs backward computation.
I0323 16:51:27.286898 26145 net.cpp:228] label_data_1_split does not need backward computation.
I0323 16:51:27.286906 26145 net.cpp:228] data does not need backward computation.
I0323 16:51:27.286909 26145 net.cpp:270] This network produces output accuracy
I0323 16:51:27.286912 26145 net.cpp:270] This network produces output loss
I0323 16:51:27.286924 26145 net.cpp:283] Network initialization done.
I0323 16:51:27.286973 26145 solver.cpp:60] Solver scaffolding done.
I0323 16:51:27.288064 26145 socket.cpp:219] Waiting for valid port [0]
I0323 16:51:27.288095 26165 socket.cpp:158] Assigned socket server port [46110]
I0323 16:51:27.290241 26165 socket.cpp:171] Socket Server ready [0.0.0.0]
I0323 16:51:27.298146 26145 socket.cpp:219] Waiting for valid port [46110]
I0323 16:51:27.298151 26145 socket.cpp:227] Valid port found [46110]
I0323 16:51:27.298162 26145 CaffeNet.cpp:186] Socket adapter: dlgpu10.ai.bjcc.qihoo.net:46110
I0323 16:51:27.298370 26145 CaffeNet.cpp:325] 0-th Socket addr: dlgpu10.ai.bjcc.qihoo.net:46110
I0323 16:51:27.298380 26145 CaffeNet.cpp:325] 1-th Socket addr: dlgpu10.ai.bjcc.qihoo.net:46110
I0323 16:51:27.298384 26145 CaffeNet.cpp:325] 2-th Socket addr: dlgpu10.ai.bjcc.qihoo.net:46110
I0323 16:51:27.298388 26145 CaffeNet.cpp:325] 3-th Socket addr: dlgpu10.ai.bjcc.qihoo.net:46110
I0323 16:51:27.298393 26145 CaffeNet.cpp:325] 4-th Socket addr:
I0323 16:51:27.298401 26145 JniCaffeNet.cpp:110] 0-th local addr: dlgpu10.ai.bjcc.qihoo.net:46110
I0323 16:51:27.298406 26145 JniCaffeNet.cpp:110] 1-th local addr: dlgpu10.ai.bjcc.qihoo.net:46110
I0323 16:51:27.298410 26145 JniCaffeNet.cpp:110] 2-th local addr: dlgpu10.ai.bjcc.qihoo.net:46110
I0323 16:51:27.298413 26145 JniCaffeNet.cpp:110] 3-th local addr: dlgpu10.ai.bjcc.qihoo.net:46110
I0323 16:51:27.298418 26145 JniCaffeNet.cpp:110] 4-th local addr:
16/03/23 16:51:27 INFO executor.Executor: Finished task 4.0 in stage 0.0 (TID 4). 1069 bytes result sent to driver
16/03/23 16:51:27 INFO executor.CoarseGrainedExecutorBackend: Got assigned task 7
16/03/23 16:51:27 INFO executor.Executor: Running task 2.0 in stage 1.0 (TID 7)
16/03/23 16:51:27 INFO broadcast.TorrentBroadcast: Started reading broadcast variable 2
16/03/23 16:51:27 INFO storage.MemoryStore: Block broadcast_2_piece0 stored as bytes in memory (estimated size 1622.0 B, free 6.9 KB)
16/03/23 16:51:27 INFO broadcast.TorrentBroadcast: Reading broadcast variable 2 took 15 ms
16/03/23 16:51:27 INFO storage.MemoryStore: Block broadcast_2 stored as values in memory (estimated size 2.6 KB, free 9.5 KB)
16/03/23 16:51:27 INFO broadcast.TorrentBroadcast: Started reading broadcast variable 1
16/03/23 16:51:27 INFO storage.MemoryStore: Block broadcast_1_piece0 stored as bytes in memory (estimated size 235.0 B, free 9.7 KB)
16/03/23 16:51:27 INFO broadcast.TorrentBroadcast: Reading broadcast variable 1 took 12 ms
16/03/23 16:51:27 INFO storage.MemoryStore: Block broadcast_1 stored as values in memory (estimated size 3.1 KB, free 12.8 KB)
I0323 16:51:27.552399 26145 common.cpp:61] 4-th string is NULL
I0323 16:51:27.552476 26145 socket.cpp:250] Trying to connect with ...[dlgpu19.ai.bjcc.qihoo.net:57402]
I0323 16:51:27.552788 26145 socket.cpp:309] Connected to server [dlgpu19.ai.bjcc.qihoo.net:57402] with client_fd [315]
I0323 16:51:37.552959 26145 socket.cpp:250] Trying to connect with ...[dlgpu19.ai.bjcc.qihoo.net:41630]
I0323 16:51:37.553390 26145 socket.cpp:309] Connected to server [dlgpu19.ai.bjcc.qihoo.net:41630] with client_fd [316]
I0323 16:51:47.553521 26145 socket.cpp:250] Trying to connect with ...[dlgpu10.ai.bjcc.qihoo.net:58067]
I0323 16:51:47.553771 26145 socket.cpp:309] Connected to server [dlgpu10.ai.bjcc.qihoo.net:58067] with client_fd [317]
I0323 16:51:57.553901 26145 socket.cpp:250] Trying to connect with ...[dlgpu20.ai.bjcc.qihoo.net:47155]
I0323 16:51:57.554342 26145 socket.cpp:309] Connected to server [dlgpu20.ai.bjcc.qihoo.net:47155] with client_fd [318]
I0323 16:51:57.555569 26165 socket.cpp:184] Accepted the connection from client [dlgpu10.ai.bjcc.qihoo.net]
I0323 16:51:57.558529 26165 socket.cpp:184] Accepted the connection from client [dlgpu20.ai.bjcc.qihoo.net]
I0323 16:51:57.560233 26165 socket.cpp:184] Accepted the connection from client [dlgpu19.ai.bjcc.qihoo.net]
I0323 16:51:57.562348 26165 socket.cpp:184] Accepted the connection from client [dlgpu19.ai.bjcc.qihoo.net]
I0323 16:52:07.600042 26145 parallel.cpp:392] GPUs pairs 0:1, 2:3, 0:2, 0:4
I0323 16:52:07.676470 26145 parallel.cpp:234] GPU 4 does not have p2p access to GPU 0
I0323 16:52:07.710218 26292 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
16/03/23 16:52:07 INFO executor.Executor: Finished task 2.0 in stage 1.0 (TID 7). 918 bytes result sent to driver
I0323 16:52:08.245970 26298 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
I0323 16:52:08.270828 26299 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
I0323 16:52:08.277041 26304 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
I0323 16:52:08.279109 26310 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
16/03/23 16:52:08 INFO executor.CoarseGrainedExecutorBackend: Got assigned task 10
16/03/23 16:52:08 INFO executor.Executor: Running task 0.0 in stage 2.0 (TID 10)
16/03/23 16:52:08 INFO broadcast.TorrentBroadcast: Started reading broadcast variable 3
16/03/23 16:52:08 INFO storage.MemoryStore: Block broadcast_3_piece0 stored as bytes in memory (estimated size 1450.0 B, free 4.7 KB)
16/03/23 16:52:08 INFO broadcast.TorrentBroadcast: Reading broadcast variable 3 took 12 ms
16/03/23 16:52:08 INFO storage.MemoryStore: Block broadcast_3 stored as values in memory (estimated size 2.2 KB, free 6.9 KB)
16/03/23 16:52:08 INFO spark.CacheManager: Partition rdd_5_0 not found, computing it
16/03/23 16:52:08 INFO spark.CacheManager: Partition rdd_0_0 not found, computing it
16/03/23 16:52:08 INFO caffe.LmdbRDD: Processing partition 0
16/03/23 16:52:10 INFO caffe.LmdbRDD: Completed partition 0
16/03/23 16:52:10 INFO storage.BlockManager: Found block rdd_0_0 locally
16/03/23 16:53:58 INFO executor.CoarseGrainedExecutorBackend: Driver commanded a shutdown
16/03/23 16:53:58 INFO storage.MemoryStore: MemoryStore cleared
16/03/23 16:53:58 INFO storage.BlockManager: BlockManager stopped
16/03/23 16:53:58 WARN executor.CoarseGrainedExecutorBackend: An unknown (dlgpu20.ai.bjcc.qihoo.net:45855) driver disconnected.
16/03/23 16:53:58 ERROR executor.CoarseGrainedExecutorBackend: Driver 10.142.118.172:45855 disassociated! Shutting down.
16/03/23 16:53:58 INFO util.ShutdownHookManager: Shutdown hook called
16/03/23 16:53:58 INFO remote.RemoteActorRefProvider$RemotingTerminator: Shutting down remote daemon.
LogType:stdout
Log Upload Time:23-Mar-2016 16:53:58
LogLength:0
Log Contents:
Container: container_1458635993280_0010_01_000003 on dlgpu10.ai.bjcc.qihoo.net_54591
LogType:stderr
Log Upload Time:23-Mar-2016 16:53:58
LogLength:31209
Log Contents:
SLF4J: Class path contains multiple SLF4J bindings.
SLF4J: Found binding in [jar:file:/dev/hadoop/nm-local-dir/usercache/hadoop/filecache/11/spark-assembly-1.6.0-hadoop2.6.4-U4.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: Found binding in [jar:file:/home/hadoop/software/hadoop-2.6.4/share/hadoop/common/lib/slf4j-log4j12-1.7.5.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: See http://www.slf4j.org/codes.html#multiple_bindings for an explanation.
SLF4J: Actual binding is of type [org.slf4j.impl.Log4jLoggerFactory]
16/03/23 16:51:12 INFO executor.CoarseGrainedExecutorBackend: Registered signal handlers for [TERM, HUP, INT]
16/03/23 16:51:13 INFO yarn.YarnSparkHadoopUtil: Set ugi in YarnConf to: hadoop,hadoop
16/03/23 16:51:13 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
16/03/23 16:51:13 INFO yarn.YarnSparkHadoopUtil: running as user: hadoop
16/03/23 16:51:13 INFO spark.SecurityManager: Changing view acls to: hadoop
16/03/23 16:51:13 INFO spark.SecurityManager: Changing modify acls to: hadoop
16/03/23 16:51:13 INFO spark.SecurityManager: SecurityManager: authentication disabled; ui acls disabled; users with view permissions: Set(hadoop); users with modify permissions: Set(hadoop)
16/03/23 16:51:14 INFO spark.SecurityManager: Changing view acls to: hadoop
16/03/23 16:51:14 INFO spark.SecurityManager: Changing modify acls to: hadoop
16/03/23 16:51:14 INFO spark.SecurityManager: SecurityManager: authentication disabled; ui acls disabled; users with view permissions: Set(hadoop); users with modify permissions: Set(hadoop)
16/03/23 16:51:14 INFO slf4j.Slf4jLogger: Slf4jLogger started
16/03/23 16:51:14 INFO Remoting: Starting remoting
16/03/23 16:51:14 INFO Remoting: Remoting started; listening on addresses :[akka.tcp://[email protected]:39671]
16/03/23 16:51:14 INFO util.Utils: Successfully started service 'sparkExecutorActorSystem' on port 39671.
16/03/23 16:51:14 INFO storage.DiskBlockManager: Created local directory at /home/hadoop/yarn/nm-local-dir/usercache/hadoop/appcache/application_1458635993280_0010/blockmgr-2691b3ef-44ad-456c-8e25-6ca9a207b8d8
16/03/23 16:51:14 INFO storage.DiskBlockManager: Created local directory at /dev/hadoop/nm-local-dir/usercache/hadoop/appcache/application_1458635993280_0010/blockmgr-8983bf2d-3fed-4721-8ca6-f9ff555b5b7b
16/03/23 16:51:14 INFO storage.DiskBlockManager: Created local directory at /run/hadoop/nm-local-dir/usercache/hadoop/appcache/application_1458635993280_0010/blockmgr-88fabbbb-2ab7-4079-bfe2-8f7ff6bdb746
16/03/23 16:51:14 INFO storage.MemoryStore: MemoryStore started with capacity 24.9 GB
16/03/23 16:51:14 INFO yarn.YarnSparkHadoopUtil: Set ugi in YarnConf to: hadoop,hadoop
16/03/23 16:51:15 INFO executor.CoarseGrainedExecutorBackend: Connecting to driver: spark://[email protected]:45855
16/03/23 16:51:15 INFO executor.CoarseGrainedExecutorBackend: Successfully registered with driver
16/03/23 16:51:15 INFO executor.Executor: Starting executor ID 2 on host dlgpu10.ai.bjcc.qihoo.net
16/03/23 16:51:15 INFO util.Utils: Successfully started service 'org.apache.spark.network.netty.NettyBlockTransferService' on port 33341.
16/03/23 16:51:15 INFO netty.NettyBlockTransferService: Server created on 33341
16/03/23 16:51:15 INFO storage.BlockManager: external shuffle service port = 7337
16/03/23 16:51:15 INFO storage.BlockManagerMaster: Trying to register BlockManager
16/03/23 16:51:15 INFO storage.BlockManagerMaster: Registered BlockManager
16/03/23 16:51:15 INFO storage.BlockManager: Registering executor with local external shuffle service.
16/03/23 16:51:19 INFO executor.CoarseGrainedExecutorBackend: Got assigned task 2
16/03/23 16:51:19 INFO executor.Executor: Running task 2.0 in stage 0.0 (TID 2)
16/03/23 16:51:19 INFO broadcast.TorrentBroadcast: Started reading broadcast variable 0
16/03/23 16:51:20 INFO storage.MemoryStore: Block broadcast_0_piece0 stored as bytes in memory (estimated size 2.1 KB, free 2.1 KB)
16/03/23 16:51:20 INFO broadcast.TorrentBroadcast: Reading broadcast variable 0 took 277 ms
16/03/23 16:51:20 INFO storage.MemoryStore: Block broadcast_0 stored as values in memory (estimated size 3.2 KB, free 5.3 KB)
16/03/23 16:51:20 INFO caffe.CaffeProcessor: my rank is 2
16/03/23 16:51:20 INFO caffe.LMDB: Batch size:64
WARNING: Logging before InitGoogleLogging() is written to STDERR
I0323 16:51:27.065438 26144 CaffeNet.cpp:78] set root solver device id to 0
I0323 16:51:27.236641 26144 solver.cpp:48] Initializing solver from parameters:
test_iter: 1
test_interval: 10001
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: 10001
snapshot_prefix: "mnist_lenet"
solver_mode: GPU
device_id: 0
net_param {
name: "LeNet"
layer {
name: "data"
type: "MemoryData"
top: "data"
top: "label"
include {
phase: TRAIN
}
memory_data_param {
batch_size: 64
channels: 1
height: 28
width: 28
share_in_parallel: false
source: "hdfs:///projects/machine_learning/image_dataset/mnist_train_lmdb/"
}
source_class: "com.yahoo.ml.caffe.LMDB"
}
layer {
name: "data"
type: "MemoryData"
top: "data"
top: "label"
include {
phase: TEST
}
memory_data_param {
batch_size: 100
channels: 1
height: 28
width: 28
share_in_parallel: false
source: "hdfs:///projects/machine_learning/image_dataset/mnist_test_lmdb/"
}
source_class: "com.yahoo.ml.caffe.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"
}
}
test_initialization: false
I0323 16:51:27.236922 26144 solver.cpp:86] Creating training net specified in net_param.
I0323 16:51:27.237049 26144 net.cpp:322] The NetState phase (0) differed from the phase (1) specified by a rule in layer data
I0323 16:51:27.237072 26144 net.cpp:322] The NetState phase (0) differed from the phase (1) specified by a rule in layer accuracy
I0323 16:51:27.237207 26144 net.cpp:49] Initializing net from parameters:
name: "LeNet"
state {
phase: TRAIN
}
layer {
name: "data"
type: "MemoryData"
top: "data"
top: "label"
include {
phase: TRAIN
}
memory_data_param {
batch_size: 64
channels: 1
height: 28
width: 28
share_in_parallel: false
source: "hdfs:///projects/machine_learning/image_dataset/mnist_train_lmdb/"
}
source_class: "com.yahoo.ml.caffe.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"
}
I0323 16:51:27.237365 26144 layer_factory.hpp:77] Creating layer data
I0323 16:51:27.237407 26144 net.cpp:106] Creating Layer data
I0323 16:51:27.237423 26144 net.cpp:411] data -> data
I0323 16:51:27.237469 26144 net.cpp:411] data -> label
I0323 16:51:27.243285 26144 net.cpp:150] Setting up data
I0323 16:51:27.243335 26144 net.cpp:157] Top shape: 64 1 28 28 (50176)
I0323 16:51:27.243348 26144 net.cpp:157] Top shape: 64 (64)
I0323 16:51:27.243355 26144 net.cpp:165] Memory required for data: 200960
I0323 16:51:27.243369 26144 layer_factory.hpp:77] Creating layer conv1
I0323 16:51:27.243397 26144 net.cpp:106] Creating Layer conv1
I0323 16:51:27.243408 26144 net.cpp:454] conv1 <- data
I0323 16:51:27.243434 26144 net.cpp:411] conv1 -> conv1
I0323 16:51:27.248071 26144 net.cpp:150] Setting up conv1
I0323 16:51:27.248101 26144 net.cpp:157] Top shape: 64 20 24 24 (737280)
I0323 16:51:27.248111 26144 net.cpp:165] Memory required for data: 3150080
I0323 16:51:27.248141 26144 layer_factory.hpp:77] Creating layer pool1
I0323 16:51:27.248162 26144 net.cpp:106] Creating Layer pool1
I0323 16:51:27.248170 26144 net.cpp:454] pool1 <- conv1
I0323 16:51:27.248180 26144 net.cpp:411] pool1 -> pool1
I0323 16:51:27.248386 26144 net.cpp:150] Setting up pool1
I0323 16:51:27.248400 26144 net.cpp:157] Top shape: 64 20 12 12 (184320)
I0323 16:51:27.248407 26144 net.cpp:165] Memory required for data: 3887360
I0323 16:51:27.248414 26144 layer_factory.hpp:77] Creating layer conv2
I0323 16:51:27.248431 26144 net.cpp:106] Creating Layer conv2
I0323 16:51:27.248440 26144 net.cpp:454] conv2 <- pool1
I0323 16:51:27.248452 26144 net.cpp:411] conv2 -> conv2
I0323 16:51:27.250042 26144 net.cpp:150] Setting up conv2
I0323 16:51:27.250069 26144 net.cpp:157] Top shape: 64 50 8 8 (204800)
I0323 16:51:27.250077 26144 net.cpp:165] Memory required for data: 4706560
I0323 16:51:27.250092 26144 layer_factory.hpp:77] Creating layer pool2
I0323 16:51:27.250104 26144 net.cpp:106] Creating Layer pool2
I0323 16:51:27.250113 26144 net.cpp:454] pool2 <- conv2
I0323 16:51:27.250121 26144 net.cpp:411] pool2 -> pool2
I0323 16:51:27.250284 26144 net.cpp:150] Setting up pool2
I0323 16:51:27.250305 26144 net.cpp:157] Top shape: 64 50 4 4 (51200)
I0323 16:51:27.250313 26144 net.cpp:165] Memory required for data: 4911360
I0323 16:51:27.250319 26144 layer_factory.hpp:77] Creating layer ip1
I0323 16:51:27.250339 26144 net.cpp:106] Creating Layer ip1
I0323 16:51:27.250345 26144 net.cpp:454] ip1 <- pool2
I0323 16:51:27.250355 26144 net.cpp:411] ip1 -> ip1
I0323 16:51:27.258404 26144 net.cpp:150] Setting up ip1
I0323 16:51:27.258431 26144 net.cpp:157] Top shape: 64 500 (32000)
I0323 16:51:27.258440 26144 net.cpp:165] Memory required for data: 5039360
I0323 16:51:27.258455 26144 layer_factory.hpp:77] Creating layer relu1
I0323 16:51:27.258473 26144 net.cpp:106] Creating Layer relu1
I0323 16:51:27.258481 26144 net.cpp:454] relu1 <- ip1
I0323 16:51:27.258491 26144 net.cpp:397] relu1 -> ip1 (in-place)
I0323 16:51:27.258507 26144 net.cpp:150] Setting up relu1
I0323 16:51:27.258515 26144 net.cpp:157] Top shape: 64 500 (32000)
I0323 16:51:27.258522 26144 net.cpp:165] Memory required for data: 5167360
I0323 16:51:27.258527 26144 layer_factory.hpp:77] Creating layer ip2
I0323 16:51:27.258538 26144 net.cpp:106] Creating Layer ip2
I0323 16:51:27.258544 26144 net.cpp:454] ip2 <- ip1
I0323 16:51:27.258558 26144 net.cpp:411] ip2 -> ip2
I0323 16:51:27.260877 26144 net.cpp:150] Setting up ip2
I0323 16:51:27.260903 26144 net.cpp:157] Top shape: 64 10 (640)
I0323 16:51:27.260911 26144 net.cpp:165] Memory required for data: 5169920
I0323 16:51:27.260923 26144 layer_factory.hpp:77] Creating layer loss
I0323 16:51:27.260956 26144 net.cpp:106] Creating Layer loss
I0323 16:51:27.260963 26144 net.cpp:454] loss <- ip2
I0323 16:51:27.260972 26144 net.cpp:454] loss <- label
I0323 16:51:27.260984 26144 net.cpp:411] loss -> loss
I0323 16:51:27.261020 26144 layer_factory.hpp:77] Creating layer loss
I0323 16:51:27.261478 26144 net.cpp:150] Setting up loss
I0323 16:51:27.261492 26144 net.cpp:157] Top shape: (1)
I0323 16:51:27.261499 26144 net.cpp:160] with loss weight 1
I0323 16:51:27.261528 26144 net.cpp:165] Memory required for data: 5169924
I0323 16:51:27.261534 26144 net.cpp:226] loss needs backward computation.
I0323 16:51:27.261543 26144 net.cpp:226] ip2 needs backward computation.
I0323 16:51:27.261548 26144 net.cpp:226] relu1 needs backward computation.
I0323 16:51:27.261554 26144 net.cpp:226] ip1 needs backward computation.
I0323 16:51:27.261559 26144 net.cpp:226] pool2 needs backward computation.
I0323 16:51:27.261565 26144 net.cpp:226] conv2 needs backward computation.
I0323 16:51:27.261571 26144 net.cpp:226] pool1 needs backward computation.
I0323 16:51:27.261577 26144 net.cpp:226] conv1 needs backward computation.
I0323 16:51:27.261584 26144 net.cpp:228] data does not need backward computation.
I0323 16:51:27.261590 26144 net.cpp:270] This network produces output loss
I0323 16:51:27.261605 26144 net.cpp:283] Network initialization done.
I0323 16:51:27.261695 26144 solver.cpp:181] Creating test net (#0) specified by net_param
I0323 16:51:27.261728 26144 net.cpp:322] The NetState phase (1) differed from the phase (0) specified by a rule in layer data
I0323 16:51:27.261884 26144 net.cpp:49] Initializing net from parameters:
name: "LeNet"
state {
phase: TEST
}
layer {
name: "data"
type: "MemoryData"
top: "data"
top: "label"
include {
phase: TEST
}
memory_data_param {
batch_size: 100
channels: 1
height: 28
width: 28
share_in_parallel: false
source: "hdfs:///projects/machine_learning/image_dataset/mnist_test_lmdb/"
}
source_class: "com.yahoo.ml.caffe.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"
}
I0323 16:51:27.262022 26144 layer_factory.hpp:77] Creating layer data
I0323 16:51:27.262045 26144 net.cpp:106] Creating Layer data
I0323 16:51:27.262053 26144 net.cpp:411] data -> data
I0323 16:51:27.262068 26144 net.cpp:411] data -> label
I0323 16:51:27.264725 26144 net.cpp:150] Setting up data
I0323 16:51:27.264752 26144 net.cpp:157] Top shape: 100 1 28 28 (78400)
I0323 16:51:27.264765 26144 net.cpp:157] Top shape: 100 (100)
I0323 16:51:27.264772 26144 net.cpp:165] Memory required for data: 314000
I0323 16:51:27.264780 26144 layer_factory.hpp:77] Creating layer label_data_1_split
I0323 16:51:27.264796 26144 net.cpp:106] Creating Layer label_data_1_split
I0323 16:51:27.264802 26144 net.cpp:454] label_data_1_split <- label
I0323 16:51:27.264811 26144 net.cpp:411] label_data_1_split -> label_data_1_split_0
I0323 16:51:27.264823 26144 net.cpp:411] label_data_1_split -> label_data_1_split_1
I0323 16:51:27.264999 26144 net.cpp:150] Setting up label_data_1_split
I0323 16:51:27.265017 26144 net.cpp:157] Top shape: 100 (100)
I0323 16:51:27.265025 26144 net.cpp:157] Top shape: 100 (100)
I0323 16:51:27.265032 26144 net.cpp:165] Memory required for data: 314800
I0323 16:51:27.265038 26144 layer_factory.hpp:77] Creating layer conv1
I0323 16:51:27.265058 26144 net.cpp:106] Creating Layer conv1
I0323 16:51:27.265064 26144 net.cpp:454] conv1 <- data
I0323 16:51:27.265079 26144 net.cpp:411] conv1 -> conv1
I0323 16:51:27.266289 26144 net.cpp:150] Setting up conv1
I0323 16:51:27.266309 26144 net.cpp:157] Top shape: 100 20 24 24 (1152000)
I0323 16:51:27.266315 26144 net.cpp:165] Memory required for data: 4922800
I0323 16:51:27.266330 26144 layer_factory.hpp:77] Creating layer pool1
I0323 16:51:27.266345 26144 net.cpp:106] Creating Layer pool1
I0323 16:51:27.266351 26144 net.cpp:454] pool1 <- conv1
I0323 16:51:27.266362 26144 net.cpp:411] pool1 -> pool1
I0323 16:51:27.266513 26144 net.cpp:150] Setting up pool1
I0323 16:51:27.266525 26144 net.cpp:157] Top shape: 100 20 12 12 (288000)
I0323 16:51:27.266530 26144 net.cpp:165] Memory required for data: 6074800
I0323 16:51:27.266535 26144 layer_factory.hpp:77] Creating layer conv2
I0323 16:51:27.266552 26144 net.cpp:106] Creating Layer conv2
I0323 16:51:27.266558 26144 net.cpp:454] conv2 <- pool1
I0323 16:51:27.266571 26144 net.cpp:411] conv2 -> conv2
I0323 16:51:27.268051 26144 net.cpp:150] Setting up conv2
I0323 16:51:27.268070 26144 net.cpp:157] Top shape: 100 50 8 8 (320000)
I0323 16:51:27.268076 26144 net.cpp:165] Memory required for data: 7354800
I0323 16:51:27.268093 26144 layer_factory.hpp:77] Creating layer pool2
I0323 16:51:27.268105 26144 net.cpp:106] Creating Layer pool2
I0323 16:51:27.268110 26144 net.cpp:454] pool2 <- conv2
I0323 16:51:27.268121 26144 net.cpp:411] pool2 -> pool2
I0323 16:51:27.268266 26144 net.cpp:150] Setting up pool2
I0323 16:51:27.268278 26144 net.cpp:157] Top shape: 100 50 4 4 (80000)
I0323 16:51:27.268291 26144 net.cpp:165] Memory required for data: 7674800
I0323 16:51:27.268297 26144 layer_factory.hpp:77] Creating layer ip1
I0323 16:51:27.268311 26144 net.cpp:106] Creating Layer ip1
I0323 16:51:27.268317 26144 net.cpp:454] ip1 <- pool2
I0323 16:51:27.268328 26144 net.cpp:411] ip1 -> ip1
I0323 16:51:27.276182 26144 net.cpp:150] Setting up ip1
I0323 16:51:27.276206 26144 net.cpp:157] Top shape: 100 500 (50000)
I0323 16:51:27.276213 26144 net.cpp:165] Memory required for data: 7874800
I0323 16:51:27.276228 26144 layer_factory.hpp:77] Creating layer relu1
I0323 16:51:27.276242 26144 net.cpp:106] Creating Layer relu1
I0323 16:51:27.276249 26144 net.cpp:454] relu1 <- ip1
I0323 16:51:27.276257 26144 net.cpp:397] relu1 -> ip1 (in-place)
I0323 16:51:27.276268 26144 net.cpp:150] Setting up relu1
I0323 16:51:27.276275 26144 net.cpp:157] Top shape: 100 500 (50000)
I0323 16:51:27.276280 26144 net.cpp:165] Memory required for data: 8074800
I0323 16:51:27.276285 26144 layer_factory.hpp:77] Creating layer ip2
I0323 16:51:27.276299 26144 net.cpp:106] Creating Layer ip2
I0323 16:51:27.276304 26144 net.cpp:454] ip2 <- ip1
I0323 16:51:27.276312 26144 net.cpp:411] ip2 -> ip2
I0323 16:51:27.276834 26144 net.cpp:150] Setting up ip2
I0323 16:51:27.276847 26144 net.cpp:157] Top shape: 100 10 (1000)
I0323 16:51:27.276854 26144 net.cpp:165] Memory required for data: 8078800
I0323 16:51:27.276862 26144 layer_factory.hpp:77] Creating layer ip2_ip2_0_split
I0323 16:51:27.276870 26144 net.cpp:106] Creating Layer ip2_ip2_0_split
I0323 16:51:27.276876 26144 net.cpp:454] ip2_ip2_0_split <- ip2
I0323 16:51:27.276892 26144 net.cpp:411] ip2_ip2_0_split -> ip2_ip2_0_split_0
I0323 16:51:27.276902 26144 net.cpp:411] ip2_ip2_0_split -> ip2_ip2_0_split_1
I0323 16:51:27.277052 26144 net.cpp:150] Setting up ip2_ip2_0_split
I0323 16:51:27.277067 26144 net.cpp:157] Top shape: 100 10 (1000)
I0323 16:51:27.277075 26144 net.cpp:157] Top shape: 100 10 (1000)
I0323 16:51:27.277079 26144 net.cpp:165] Memory required for data: 8086800
I0323 16:51:27.277084 26144 layer_factory.hpp:77] Creating layer accuracy
I0323 16:51:27.277096 26144 net.cpp:106] Creating Layer accuracy
I0323 16:51:27.277102 26144 net.cpp:454] accuracy <- ip2_ip2_0_split_0
I0323 16:51:27.277109 26144 net.cpp:454] accuracy <- label_data_1_split_0
I0323 16:51:27.277120 26144 net.cpp:411] accuracy -> accuracy
I0323 16:51:27.277137 26144 net.cpp:150] Setting up accuracy
I0323 16:51:27.277143 26144 net.cpp:157] Top shape: (1)
I0323 16:51:27.277149 26144 net.cpp:165] Memory required for data: 8086804
I0323 16:51:27.277153 26144 layer_factory.hpp:77] Creating layer loss
I0323 16:51:27.277161 26144 net.cpp:106] Creating Layer loss
I0323 16:51:27.277166 26144 net.cpp:454] loss <- ip2_ip2_0_split_1
I0323 16:51:27.277173 26144 net.cpp:454] loss <- label_data_1_split_1
I0323 16:51:27.277184 26144 net.cpp:411] loss -> loss
I0323 16:51:27.277195 26144 layer_factory.hpp:77] Creating layer loss
I0323 16:51:27.277585 26144 net.cpp:150] Setting up loss
I0323 16:51:27.277595 26144 net.cpp:157] Top shape: (1)
I0323 16:51:27.277601 26144 net.cpp:160] with loss weight 1
I0323 16:51:27.277611 26144 net.cpp:165] Memory required for data: 8086808
I0323 16:51:27.277616 26144 net.cpp:226] loss needs backward computation.
I0323 16:51:27.277622 26144 net.cpp:228] accuracy does not need backward computation.
I0323 16:51:27.277629 26144 net.cpp:226] ip2_ip2_0_split needs backward computation.
I0323 16:51:27.277634 26144 net.cpp:226] ip2 needs backward computation.
I0323 16:51:27.277639 26144 net.cpp:226] relu1 needs backward computation.
I0323 16:51:27.277644 26144 net.cpp:226] ip1 needs backward computation.
I0323 16:51:27.277649 26144 net.cpp:226] pool2 needs backward computation.
I0323 16:51:27.277654 26144 net.cpp:226] conv2 needs backward computation.
I0323 16:51:27.277660 26144 net.cpp:226] pool1 needs backward computation.
I0323 16:51:27.277665 26144 net.cpp:226] conv1 needs backward computation.
I0323 16:51:27.277672 26144 net.cpp:228] label_data_1_split does not need backward computation.
I0323 16:51:27.277688 26144 net.cpp:228] data does not need backward computation.
I0323 16:51:27.277693 26144 net.cpp:270] This network produces output accuracy
I0323 16:51:27.277699 26144 net.cpp:270] This network produces output loss
I0323 16:51:27.277719 26144 net.cpp:283] Network initialization done.
I0323 16:51:27.277781 26144 solver.cpp:60] Solver scaffolding done.
I0323 16:51:27.279260 26144 socket.cpp:219] Waiting for valid port [0]
I0323 16:51:27.279287 26164 socket.cpp:158] Assigned socket server port [58067]
I0323 16:51:27.281308 26164 socket.cpp:171] Socket Server ready [0.0.0.0]
I0323 16:51:27.289355 26144 socket.cpp:219] Waiting for valid port [58067]
I0323 16:51:27.289378 26144 socket.cpp:227] Valid port found [58067]
I0323 16:51:27.289403 26144 CaffeNet.cpp:186] Socket adapter: dlgpu10.ai.bjcc.qihoo.net:58067
I0323 16:51:27.289767 26144 CaffeNet.cpp:325] 0-th Socket addr: dlgpu10.ai.bjcc.qihoo.net:58067
I0323 16:51:27.289790 26144 CaffeNet.cpp:325] 1-th Socket addr: dlgpu10.ai.bjcc.qihoo.net:58067
I0323 16:51:27.289804 26144 CaffeNet.cpp:325] 2-th Socket addr:
I0323 16:51:27.289815 26144 CaffeNet.cpp:325] 3-th Socket addr: dlgpu10.ai.bjcc.qihoo.net:58067
I0323 16:51:27.289824 26144 CaffeNet.cpp:325] 4-th Socket addr: dlgpu10.ai.bjcc.qihoo.net:58067
I0323 16:51:27.289845 26144 JniCaffeNet.cpp:110] 0-th local addr: dlgpu10.ai.bjcc.qihoo.net:58067
I0323 16:51:27.289860 26144 JniCaffeNet.cpp:110] 1-th local addr: dlgpu10.ai.bjcc.qihoo.net:58067
I0323 16:51:27.289870 26144 JniCaffeNet.cpp:110] 2-th local addr:
I0323 16:51:27.289878 26144 JniCaffeNet.cpp:110] 3-th local addr: dlgpu10.ai.bjcc.qihoo.net:58067
I0323 16:51:27.289887 26144 JniCaffeNet.cpp:110] 4-th local addr: dlgpu10.ai.bjcc.qihoo.net:58067
16/03/23 16:51:27 INFO executor.Executor: Finished task 2.0 in stage 0.0 (TID 2). 1069 bytes result sent to driver
16/03/23 16:51:27 INFO executor.CoarseGrainedExecutorBackend: Got assigned task 6
16/03/23 16:51:27 INFO executor.Executor: Running task 1.0 in stage 1.0 (TID 6)
16/03/23 16:51:27 INFO broadcast.TorrentBroadcast: Started reading broadcast variable 2
16/03/23 16:51:27 INFO storage.MemoryStore: Block broadcast_2_piece0 stored as bytes in memory (estimated size 1622.0 B, free 6.9 KB)
16/03/23 16:51:27 INFO broadcast.TorrentBroadcast: Reading broadcast variable 2 took 15 ms
16/03/23 16:51:27 INFO storage.MemoryStore: Block broadcast_2 stored as values in memory (estimated size 2.6 KB, free 9.5 KB)
16/03/23 16:51:27 INFO broadcast.TorrentBroadcast: Started reading broadcast variable 1
16/03/23 16:51:27 INFO storage.MemoryStore: Block broadcast_1_piece0 stored as bytes in memory (estimated size 235.0 B, free 9.7 KB)
16/03/23 16:51:27 INFO broadcast.TorrentBroadcast: Reading broadcast variable 1 took 13 ms
16/03/23 16:51:27 INFO storage.MemoryStore: Block broadcast_1 stored as values in memory (estimated size 3.1 KB, free 12.8 KB)
I0323 16:51:27.553778 26144 common.cpp:61] 2-th string is NULL
I0323 16:51:27.553853 26144 socket.cpp:250] Trying to connect with ...[dlgpu19.ai.bjcc.qihoo.net:57402]
I0323 16:51:27.554154 26144 socket.cpp:309] Connected to server [dlgpu19.ai.bjcc.qihoo.net:57402] with client_fd [315]
I0323 16:51:37.554280 26144 socket.cpp:250] Trying to connect with ...[dlgpu19.ai.bjcc.qihoo.net:41630]
I0323 16:51:37.554639 26144 socket.cpp:309] Connected to server [dlgpu19.ai.bjcc.qihoo.net:41630] with client_fd [316]
I0323 16:51:37.559141 26164 socket.cpp:184] Accepted the connection from client [dlgpu19.ai.bjcc.qihoo.net]
I0323 16:51:37.561334 26164 socket.cpp:184] Accepted the connection from client [dlgpu19.ai.bjcc.qihoo.net]
I0323 16:51:47.553865 26164 socket.cpp:184] Accepted the connection from client [dlgpu10.ai.bjcc.qihoo.net]
I0323 16:51:47.554746 26144 socket.cpp:250] Trying to connect with ...[dlgpu20.ai.bjcc.qihoo.net:47155]
I0323 16:51:47.555070 26144 socket.cpp:309] Connected to server [dlgpu20.ai.bjcc.qihoo.net:47155] with client_fd [320]
I0323 16:51:47.558022 26164 socket.cpp:184] Accepted the connection from client [dlgpu20.ai.bjcc.qihoo.net]
I0323 16:51:57.555244 26144 socket.cpp:250] Trying to connect with ...[dlgpu10.ai.bjcc.qihoo.net:46110]
I0323 16:51:57.555477 26144 socket.cpp:309] Connected to server [dlgpu10.ai.bjcc.qihoo.net:46110] with client_fd [322]
I0323 16:52:07.601250 26144 parallel.cpp:392] GPUs pairs 0:1, 2:3, 0:2, 0:4
I0323 16:52:07.675889 26144 parallel.cpp:234] GPU 4 does not have p2p access to GPU 0
I0323 16:52:07.709733 26291 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
16/03/23 16:52:07 INFO executor.Executor: Finished task 1.0 in stage 1.0 (TID 6). 918 bytes result sent to driver
I0323 16:52:08.261286 26294 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
I0323 16:52:08.296454 26303 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
I0323 16:52:08.299464 26308 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
I0323 16:52:08.301334 26301 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
16/03/23 16:52:08 INFO executor.CoarseGrainedExecutorBackend: Got assigned task 14
16/03/23 16:52:08 INFO executor.Executor: Running task 4.0 in stage 2.0 (TID 14)
16/03/23 16:52:08 INFO broadcast.TorrentBroadcast: Started reading broadcast variable 3
16/03/23 16:52:08 INFO storage.MemoryStore: Block broadcast_3_piece0 stored as bytes in memory (estimated size 1450.0 B, free 4.7 KB)
16/03/23 16:52:08 INFO broadcast.TorrentBroadcast: Reading broadcast variable 3 took 14 ms
16/03/23 16:52:08 INFO storage.MemoryStore: Block broadcast_3 stored as values in memory (estimated size 2.2 KB, free 6.9 KB)
16/03/23 16:52:08 INFO spark.CacheManager: Partition rdd_5_4 not found, computing it
16/03/23 16:52:08 INFO spark.CacheManager: Partition rdd_0_4 not found, computing it
16/03/23 16:52:08 INFO caffe.LmdbRDD: Processing partition 4
16/03/23 16:52:10 INFO caffe.LmdbRDD: Completed partition 4
16/03/23 16:52:10 INFO storage.BlockManager: Found block rdd_0_4 locally
16/03/23 16:53:58 INFO executor.CoarseGrainedExecutorBackend: Driver commanded a shutdown
16/03/23 16:53:58 INFO storage.MemoryStore: MemoryStore cleared
16/03/23 16:53:58 INFO storage.BlockManager: BlockManager stopped
16/03/23 16:53:58 WARN executor.CoarseGrainedExecutorBackend: An unknown (dlgpu20.ai.bjcc.qihoo.net:45855) driver disconnected.
16/03/23 16:53:58 ERROR executor.CoarseGrainedExecutorBackend: Driver 10.142.118.172:45855 disassociated! Shutting down.
16/03/23 16:53:58 INFO util.ShutdownHookManager: Shutdown hook called
16/03/23 16:53:58 INFO remote.RemoteActorRefProvider$RemotingTerminator: Shutting down remote daemon.
LogType:stdout
Log Upload Time:23-Mar-2016 16:53:58
LogLength:0
Log Contents:
Container: container_1458635993280_0010_01_000007 on dlgpu19.ai.bjcc.qihoo.net_33935
LogType:stderr
Log Upload Time:23-Mar-2016 16:53:59
LogLength:32045
Log Contents:
SLF4J: Class path contains multiple SLF4J bindings.
SLF4J: Found binding in [jar:file:/dev/hadoop/nm-local-dir/usercache/hadoop/filecache/11/spark-assembly-1.6.0-hadoop2.6.4-U4.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: Found binding in [jar:file:/home/hadoop/software/hadoop-2.6.4/share/hadoop/common/lib/slf4j-log4j12-1.7.5.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: See http://www.slf4j.org/codes.html#multiple_bindings for an explanation.
SLF4J: Actual binding is of type [org.slf4j.impl.Log4jLoggerFactory]
16/03/23 16:51:14 INFO executor.CoarseGrainedExecutorBackend: Registered signal handlers for [TERM, HUP, INT]
16/03/23 16:51:14 INFO yarn.YarnSparkHadoopUtil: Set ugi in YarnConf to: hadoop,hadoop
16/03/23 16:51:14 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
16/03/23 16:51:14 INFO yarn.YarnSparkHadoopUtil: running as user: hadoop
16/03/23 16:51:14 INFO spark.SecurityManager: Changing view acls to: hadoop
16/03/23 16:51:14 INFO spark.SecurityManager: Changing modify acls to: hadoop
16/03/23 16:51:14 INFO spark.SecurityManager: SecurityManager: authentication disabled; ui acls disabled; users with view permissions: Set(hadoop); users with modify permissions: Set(hadoop)
16/03/23 16:51:15 INFO spark.SecurityManager: Changing view acls to: hadoop
16/03/23 16:51:15 INFO spark.SecurityManager: Changing modify acls to: hadoop
16/03/23 16:51:15 INFO spark.SecurityManager: SecurityManager: authentication disabled; ui acls disabled; users with view permissions: Set(hadoop); users with modify permissions: Set(hadoop)
16/03/23 16:51:15 INFO slf4j.Slf4jLogger: Slf4jLogger started
16/03/23 16:51:15 INFO Remoting: Starting remoting
16/03/23 16:51:16 INFO Remoting: Remoting started; listening on addresses :[akka.tcp://[email protected]:42763]
16/03/23 16:51:16 INFO util.Utils: Successfully started service 'sparkExecutorActorSystem' on port 42763.
16/03/23 16:51:16 INFO storage.DiskBlockManager: Created local directory at /home/hadoop/yarn/nm-local-dir/usercache/hadoop/appcache/application_1458635993280_0010/blockmgr-9024c2f5-2a3f-4a1a-99ab-76d15f01278b
16/03/23 16:51:16 INFO storage.DiskBlockManager: Created local directory at /dev/hadoop/nm-local-dir/usercache/hadoop/appcache/application_1458635993280_0010/blockmgr-ac7b2147-3d49-41d9-9c23-7d95ee0cdd1c
16/03/23 16:51:16 INFO storage.DiskBlockManager: Created local directory at /run/hadoop/nm-local-dir/usercache/hadoop/appcache/application_1458635993280_0010/blockmgr-2cbded0d-1350-44a2-9a31-f91a452b59f7
16/03/23 16:51:16 INFO storage.MemoryStore: MemoryStore started with capacity 24.9 GB
16/03/23 16:51:16 INFO yarn.YarnSparkHadoopUtil: Set ugi in YarnConf to: hadoop,hadoop
16/03/23 16:51:16 INFO executor.CoarseGrainedExecutorBackend: Connecting to driver: spark://[email protected]:45855
16/03/23 16:51:16 INFO executor.CoarseGrainedExecutorBackend: Successfully registered with driver
16/03/23 16:51:16 INFO executor.Executor: Starting executor ID 5 on host dlgpu19.ai.bjcc.qihoo.net
16/03/23 16:51:16 INFO util.Utils: Successfully started service 'org.apache.spark.network.netty.NettyBlockTransferService' on port 40461.
16/03/23 16:51:16 INFO netty.NettyBlockTransferService: Server created on 40461
16/03/23 16:51:16 INFO storage.BlockManager: external shuffle service port = 7337
16/03/23 16:51:16 INFO storage.BlockManagerMaster: Trying to register BlockManager
16/03/23 16:51:16 INFO storage.BlockManagerMaster: Registered BlockManager
16/03/23 16:51:16 INFO storage.BlockManager: Registering executor with local external shuffle service.
16/03/23 16:51:19 INFO executor.CoarseGrainedExecutorBackend: Got assigned task 0
16/03/23 16:51:19 INFO executor.Executor: Running task 0.0 in stage 0.0 (TID 0)
16/03/23 16:51:19 INFO broadcast.TorrentBroadcast: Started reading broadcast variable 0
16/03/23 16:51:20 INFO storage.MemoryStore: Block broadcast_0_piece0 stored as bytes in memory (estimated size 2.1 KB, free 2.1 KB)
16/03/23 16:51:20 INFO broadcast.TorrentBroadcast: Reading broadcast variable 0 took 292 ms
16/03/23 16:51:20 INFO storage.MemoryStore: Block broadcast_0 stored as values in memory (estimated size 3.2 KB, free 5.3 KB)
16/03/23 16:51:20 INFO caffe.CaffeProcessor: my rank is 0
16/03/23 16:51:20 INFO caffe.LMDB: Batch size:64
WARNING: Logging before InitGoogleLogging() is written to STDERR
I0323 16:51:22.875790 1378 CaffeNet.cpp:78] set root solver device id to 0
I0323 16:51:23.038626 1378 solver.cpp:48] Initializing solver from parameters:
test_iter: 1
test_interval: 10001
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: 10001
snapshot_prefix: "mnist_lenet"
solver_mode: GPU
device_id: 0
net_param {
name: "LeNet"
layer {
name: "data"
type: "MemoryData"
top: "data"
top: "label"
include {
phase: TRAIN
}
memory_data_param {
batch_size: 64
channels: 1
height: 28
width: 28
share_in_parallel: false
source: "hdfs:///projects/machine_learning/image_dataset/mnist_train_lmdb/"
}
source_class: "com.yahoo.ml.caffe.LMDB"
}
layer {
name: "data"
type: "MemoryData"
top: "data"
top: "label"
include {
phase: TEST
}
memory_data_param {
batch_size: 100
channels: 1
height: 28
width: 28
share_in_parallel: false
source: "hdfs:///projects/machine_learning/image_dataset/mnist_test_lmdb/"
}
source_class: "com.yahoo.ml.caffe.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"
}
}
test_initialization: false
I0323 16:51:23.038815 1378 solver.cpp:86] Creating training net specified in net_param.
I0323 16:51:23.038894 1378 net.cpp:322] The NetState phase (0) differed from the phase (1) specified by a rule in layer data
I0323 16:51:23.038911 1378 net.cpp:322] The NetState phase (0) differed from the phase (1) specified by a rule in layer accuracy
I0323 16:51:23.039005 1378 net.cpp:49] Initializing net from parameters:
name: "LeNet"
state {
phase: TRAIN
}
layer {
name: "data"
type: "MemoryData"
top: "data"
top: "label"
include {
phase: TRAIN
}
memory_data_param {
batch_size: 64
channels: 1
height: 28
width: 28
share_in_parallel: false
source: "hdfs:///projects/machine_learning/image_dataset/mnist_train_lmdb/"
}
source_class: "com.yahoo.ml.caffe.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"
}
I0323 16:51:23.039119 1378 layer_factory.hpp:77] Creating layer data
I0323 16:51:23.039150 1378 net.cpp:106] Creating Layer data
I0323 16:51:23.039162 1378 net.cpp:411] data -> data
I0323 16:51:23.039197 1378 net.cpp:411] data -> label
I0323 16:51:23.042510 1378 net.cpp:150] Setting up data
I0323 16:51:23.042541 1378 net.cpp:157] Top shape: 64 1 28 28 (50176)
I0323 16:51:23.042549 1378 net.cpp:157] Top shape: 64 (64)
I0323 16:51:23.042553 1378 net.cpp:165] Memory required for data: 200960
I0323 16:51:23.042563 1378 layer_factory.hpp:77] Creating layer conv1
I0323 16:51:23.042587 1378 net.cpp:106] Creating Layer conv1
I0323 16:51:23.042595 1378 net.cpp:454] conv1 <- data
I0323 16:51:23.042613 1378 net.cpp:411] conv1 -> conv1
I0323 16:51:23.045305 1378 net.cpp:150] Setting up conv1
I0323 16:51:23.045325 1378 net.cpp:157] Top shape: 64 20 24 24 (737280)
I0323 16:51:23.045331 1378 net.cpp:165] Memory required for data: 3150080
I0323 16:51:23.045349 1378 layer_factory.hpp:77] Creating layer pool1
I0323 16:51:23.045363 1378 net.cpp:106] Creating Layer pool1
I0323 16:51:23.045368 1378 net.cpp:454] pool1 <- conv1
I0323 16:51:23.045374 1378 net.cpp:411] pool1 -> pool1
I0323 16:51:23.045497 1378 net.cpp:150] Setting up pool1
I0323 16:51:23.045508 1378 net.cpp:157] Top shape: 64 20 12 12 (184320)
I0323 16:51:23.045512 1378 net.cpp:165] Memory required for data: 3887360
I0323 16:51:23.045516 1378 layer_factory.hpp:77] Creating layer conv2
I0323 16:51:23.045531 1378 net.cpp:106] Creating Layer conv2
I0323 16:51:23.045536 1378 net.cpp:454] conv2 <- pool1
I0323 16:51:23.045543 1378 net.cpp:411] conv2 -> conv2
I0323 16:51:23.046510 1378 net.cpp:150] Setting up conv2
I0323 16:51:23.046527 1378 net.cpp:157] Top shape: 64 50 8 8 (204800)
I0323 16:51:23.046532 1378 net.cpp:165] Memory required for data: 4706560
I0323 16:51:23.046542 1378 layer_factory.hpp:77] Creating layer pool2
I0323 16:51:23.046553 1378 net.cpp:106] Creating Layer pool2
I0323 16:51:23.046558 1378 net.cpp:454] pool2 <- conv2
I0323 16:51:23.046564 1378 net.cpp:411] pool2 -> pool2
I0323 16:51:23.046670 1378 net.cpp:150] Setting up pool2
I0323 16:51:23.046687 1378 net.cpp:157] Top shape: 64 50 4 4 (51200)
I0323 16:51:23.046692 1378 net.cpp:165] Memory required for data: 4911360
I0323 16:51:23.046696 1378 layer_factory.hpp:77] Creating layer ip1
I0323 16:51:23.046708 1378 net.cpp:106] Creating Layer ip1
I0323 16:51:23.046713 1378 net.cpp:454] ip1 <- pool2
I0323 16:51:23.046722 1378 net.cpp:411] ip1 -> ip1
I0323 16:51:23.051942 1378 net.cpp:150] Setting up ip1
I0323 16:51:23.051961 1378 net.cpp:157] Top shape: 64 500 (32000)
I0323 16:51:23.051966 1378 net.cpp:165] Memory required for data: 5039360
I0323 16:51:23.051977 1378 layer_factory.hpp:77] Creating layer relu1
I0323 16:51:23.051988 1378 net.cpp:106] Creating Layer relu1
I0323 16:51:23.051995 1378 net.cpp:454] relu1 <- ip1
I0323 16:51:23.052000 1378 net.cpp:397] relu1 -> ip1 (in-place)
I0323 16:51:23.052012 1378 net.cpp:150] Setting up relu1
I0323 16:51:23.052018 1378 net.cpp:157] Top shape: 64 500 (32000)
I0323 16:51:23.052022 1378 net.cpp:165] Memory required for data: 5167360
I0323 16:51:23.052026 1378 layer_factory.hpp:77] Creating layer ip2
I0323 16:51:23.052037 1378 net.cpp:106] Creating Layer ip2
I0323 16:51:23.052042 1378 net.cpp:454] ip2 <- ip1
I0323 16:51:23.052050 1378 net.cpp:411] ip2 -> ip2
I0323 16:51:23.053828 1378 net.cpp:150] Setting up ip2
I0323 16:51:23.053848 1378 net.cpp:157] Top shape: 64 10 (640)
I0323 16:51:23.053853 1378 net.cpp:165] Memory required for data: 5169920
I0323 16:51:23.053860 1378 layer_factory.hpp:77] Creating layer loss
I0323 16:51:23.053875 1378 net.cpp:106] Creating Layer loss
I0323 16:51:23.053881 1378 net.cpp:454] loss <- ip2
I0323 16:51:23.053886 1378 net.cpp:454] loss <- label
I0323 16:51:23.053896 1378 net.cpp:411] loss -> loss
I0323 16:51:23.053915 1378 layer_factory.hpp:77] Creating layer loss
I0323 16:51:23.054229 1378 net.cpp:150] Setting up loss
I0323 16:51:23.054244 1378 net.cpp:157] Top shape: (1)
I0323 16:51:23.054250 1378 net.cpp:160] with loss weight 1
I0323 16:51:23.054270 1378 net.cpp:165] Memory required for data: 5169924
I0323 16:51:23.054275 1378 net.cpp:226] loss needs backward computation.
I0323 16:51:23.054280 1378 net.cpp:226] ip2 needs backward computation.
I0323 16:51:23.054283 1378 net.cpp:226] relu1 needs backward computation.
I0323 16:51:23.054287 1378 net.cpp:226] ip1 needs backward computation.
I0323 16:51:23.054291 1378 net.cpp:226] pool2 needs backward computation.
I0323 16:51:23.054296 1378 net.cpp:226] conv2 needs backward computation.
I0323 16:51:23.054299 1378 net.cpp:226] pool1 needs backward computation.
I0323 16:51:23.054303 1378 net.cpp:226] conv1 needs backward computation.
I0323 16:51:23.054307 1378 net.cpp:228] data does not need backward computation.
I0323 16:51:23.054311 1378 net.cpp:270] This network produces output loss
I0323 16:51:23.054325 1378 net.cpp:283] Network initialization done.
I0323 16:51:23.054383 1378 solver.cpp:181] Creating test net (#0) specified by net_param
I0323 16:51:23.054406 1378 net.cpp:322] The NetState phase (1) differed from the phase (0) specified by a rule in layer data
I0323 16:51:23.054507 1378 net.cpp:49] Initializing net from parameters:
name: "LeNet"
state {
phase: TEST
}
layer {
name: "data"
type: "MemoryData"
top: "data"
top: "label"
include {
phase: TEST
}
memory_data_param {
batch_size: 100
channels: 1
height: 28
width: 28
share_in_parallel: false
source: "hdfs:///projects/machine_learning/image_dataset/mnist_test_lmdb/"
}
source_class: "com.yahoo.ml.caffe.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"
}
I0323 16:51:23.054594 1378 layer_factory.hpp:77] Creating layer data
I0323 16:51:23.054605 1378 net.cpp:106] Creating Layer data
I0323 16:51:23.054611 1378 net.cpp:411] data -> data
I0323 16:51:23.054621 1378 net.cpp:411] data -> label
I0323 16:51:23.056326 1378 net.cpp:150] Setting up data
I0323 16:51:23.056345 1378 net.cpp:157] Top shape: 100 1 28 28 (78400)
I0323 16:51:23.056352 1378 net.cpp:157] Top shape: 100 (100)
I0323 16:51:23.056356 1378 net.cpp:165] Memory required for data: 314000
I0323 16:51:23.056361 1378 layer_factory.hpp:77] Creating layer label_data_1_split
I0323 16:51:23.056372 1378 net.cpp:106] Creating Layer label_data_1_split
I0323 16:51:23.056377 1378 net.cpp:454] label_data_1_split <- label
I0323 16:51:23.056386 1378 net.cpp:411] label_data_1_split -> label_data_1_split_0
I0323 16:51:23.056396 1378 net.cpp:411] label_data_1_split -> label_data_1_split_1
I0323 16:51:23.056507 1378 net.cpp:150] Setting up label_data_1_split
I0323 16:51:23.056517 1378 net.cpp:157] Top shape: 100 (100)
I0323 16:51:23.056522 1378 net.cpp:157] Top shape: 100 (100)
I0323 16:51:23.056526 1378 net.cpp:165] Memory required for data: 314800
I0323 16:51:23.056530 1378 layer_factory.hpp:77] Creating layer conv1
I0323 16:51:23.056543 1378 net.cpp:106] Creating Layer conv1
I0323 16:51:23.056548 1378 net.cpp:454] conv1 <- data
I0323 16:51:23.056556 1378 net.cpp:411] conv1 -> conv1
I0323 16:51:23.057343 1378 net.cpp:150] Setting up conv1
I0323 16:51:23.057359 1378 net.cpp:157] Top shape: 100 20 24 24 (1152000)
I0323 16:51:23.057364 1378 net.cpp:165] Memory required for data: 4922800
I0323 16:51:23.057375 1378 layer_factory.hpp:77] Creating layer pool1
I0323 16:51:23.057385 1378 net.cpp:106] Creating Layer pool1
I0323 16:51:23.057389 1378 net.cpp:454] pool1 <- conv1
I0323 16:51:23.057396 1378 net.cpp:411] pool1 -> pool1
I0323 16:51:23.057507 1378 net.cpp:150] Setting up pool1
I0323 16:51:23.057517 1378 net.cpp:157] Top shape: 100 20 12 12 (288000)
I0323 16:51:23.057521 1378 net.cpp:165] Memory required for data: 6074800
I0323 16:51:23.057525 1378 layer_factory.hpp:77] Creating layer conv2
I0323 16:51:23.057538 1378 net.cpp:106] Creating Layer conv2
I0323 16:51:23.057543 1378 net.cpp:454] conv2 <- pool1
I0323 16:51:23.057549 1378 net.cpp:411] conv2 -> conv2
I0323 16:51:23.058568 1378 net.cpp:150] Setting up conv2
I0323 16:51:23.058583 1378 net.cpp:157] Top shape: 100 50 8 8 (320000)
I0323 16:51:23.058588 1378 net.cpp:165] Memory required for data: 7354800
I0323 16:51:23.058601 1378 layer_factory.hpp:77] Creating layer pool2
I0323 16:51:23.058609 1378 net.cpp:106] Creating Layer pool2
I0323 16:51:23.058614 1378 net.cpp:454] pool2 <- conv2
I0323 16:51:23.058619 1378 net.cpp:411] pool2 -> pool2
I0323 16:51:23.058724 1378 net.cpp:150] Setting up pool2
I0323 16:51:23.058733 1378 net.cpp:157] Top shape: 100 50 4 4 (80000)
I0323 16:51:23.058743 1378 net.cpp:165] Memory required for data: 7674800
I0323 16:51:23.058748 1378 layer_factory.hpp:77] Creating layer ip1
I0323 16:51:23.058759 1378 net.cpp:106] Creating Layer ip1
I0323 16:51:23.058764 1378 net.cpp:454] ip1 <- pool2
I0323 16:51:23.058770 1378 net.cpp:411] ip1 -> ip1
I0323 16:51:23.064004 1378 net.cpp:150] Setting up ip1
I0323 16:51:23.064024 1378 net.cpp:157] Top shape: 100 500 (50000)
I0323 16:51:23.064029 1378 net.cpp:165] Memory required for data: 7874800
I0323 16:51:23.064040 1378 layer_factory.hpp:77] Creating layer relu1
I0323 16:51:23.064051 1378 net.cpp:106] Creating Layer relu1
I0323 16:51:23.064056 1378 net.cpp:454] relu1 <- ip1
I0323 16:51:23.064062 1378 net.cpp:397] relu1 -> ip1 (in-place)
I0323 16:51:23.064070 1378 net.cpp:150] Setting up relu1
I0323 16:51:23.064076 1378 net.cpp:157] Top shape: 100 500 (50000)
I0323 16:51:23.064079 1378 net.cpp:165] Memory required for data: 8074800
I0323 16:51:23.064084 1378 layer_factory.hpp:77] Creating layer ip2
I0323 16:51:23.064092 1378 net.cpp:106] Creating Layer ip2
I0323 16:51:23.064096 1378 net.cpp:454] ip2 <- ip1
I0323 16:51:23.064106 1378 net.cpp:411] ip2 -> ip2
I0323 16:51:23.064492 1378 net.cpp:150] Setting up ip2
I0323 16:51:23.064503 1378 net.cpp:157] Top shape: 100 10 (1000)
I0323 16:51:23.064508 1378 net.cpp:165] Memory required for data: 8078800
I0323 16:51:23.064515 1378 layer_factory.hpp:77] Creating layer ip2_ip2_0_split
I0323 16:51:23.064525 1378 net.cpp:106] Creating Layer ip2_ip2_0_split
I0323 16:51:23.064530 1378 net.cpp:454] ip2_ip2_0_split <- ip2
I0323 16:51:23.064537 1378 net.cpp:411] ip2_ip2_0_split -> ip2_ip2_0_split_0
I0323 16:51:23.064546 1378 net.cpp:411] ip2_ip2_0_split -> ip2_ip2_0_split_1
I0323 16:51:23.064656 1378 net.cpp:150] Setting up ip2_ip2_0_split
I0323 16:51:23.064664 1378 net.cpp:157] Top shape: 100 10 (1000)
I0323 16:51:23.064671 1378 net.cpp:157] Top shape: 100 10 (1000)
I0323 16:51:23.064673 1378 net.cpp:165] Memory required for data: 8086800
I0323 16:51:23.064678 1378 layer_factory.hpp:77] Creating layer accuracy
I0323 16:51:23.064690 1378 net.cpp:106] Creating Layer accuracy
I0323 16:51:23.064697 1378 net.cpp:454] accuracy <- ip2_ip2_0_split_0
I0323 16:51:23.064702 1378 net.cpp:454] accuracy <- label_data_1_split_0
I0323 16:51:23.064707 1378 net.cpp:411] accuracy -> accuracy
I0323 16:51:23.064721 1378 net.cpp:150] Setting up accuracy
I0323 16:51:23.064728 1378 net.cpp:157] Top shape: (1)
I0323 16:51:23.064733 1378 net.cpp:165] Memory required for data: 8086804
I0323 16:51:23.064736 1378 layer_factory.hpp:77] Creating layer loss
I0323 16:51:23.064744 1378 net.cpp:106] Creating Layer loss
I0323 16:51:23.064749 1378 net.cpp:454] loss <- ip2_ip2_0_split_1
I0323 16:51:23.064754 1378 net.cpp:454] loss <- label_data_1_split_1
I0323 16:51:23.064759 1378 net.cpp:411] loss -> loss
I0323 16:51:23.064769 1378 layer_factory.hpp:77] Creating layer loss
I0323 16:51:23.065067 1378 net.cpp:150] Setting up loss
I0323 16:51:23.065081 1378 net.cpp:157] Top shape: (1)
I0323 16:51:23.065085 1378 net.cpp:160] with loss weight 1
I0323 16:51:23.065093 1378 net.cpp:165] Memory required for data: 8086808
I0323 16:51:23.065098 1378 net.cpp:226] loss needs backward computation.
I0323 16:51:23.065102 1378 net.cpp:228] accuracy does not need backward computation.
I0323 16:51:23.065107 1378 net.cpp:226] ip2_ip2_0_split needs backward computation.
I0323 16:51:23.065111 1378 net.cpp:226] ip2 needs backward computation.
I0323 16:51:23.065115 1378 net.cpp:226] relu1 needs backward computation.
I0323 16:51:23.065119 1378 net.cpp:226] ip1 needs backward computation.
I0323 16:51:23.065122 1378 net.cpp:226] pool2 needs backward computation.
I0323 16:51:23.065127 1378 net.cpp:226] conv2 needs backward computation.
I0323 16:51:23.065131 1378 net.cpp:226] pool1 needs backward computation.
I0323 16:51:23.065135 1378 net.cpp:226] conv1 needs backward computation.
I0323 16:51:23.065140 1378 net.cpp:228] label_data_1_split does not need backward computation.
I0323 16:51:23.065150 1378 net.cpp:228] data does not need backward computation.
I0323 16:51:23.065155 1378 net.cpp:270] This network produces output accuracy
I0323 16:51:23.065160 1378 net.cpp:270] This network produces output loss
I0323 16:51:23.065173 1378 net.cpp:283] Network initialization done.
I0323 16:51:23.065218 1378 solver.cpp:60] Solver scaffolding done.
I0323 16:51:23.066324 1378 socket.cpp:224] Waiting for valid port [0]
I0323 16:51:23.066359 1414 socket.cpp:163] Assigned socket server port [57402]
I0323 16:51:23.068367 1414 socket.cpp:176] Socket Server ready [0.0.0.0]
I0323 16:51:23.076409 1378 socket.cpp:224] Waiting for valid port [57402]
I0323 16:51:23.076423 1378 socket.cpp:232] Valid port found [57402]
I0323 16:51:23.076437 1378 CaffeNet.cpp:186] Socket adapter: dlgpu19.ai.bjcc.qihoo.net:57402
I0323 16:51:23.076686 1378 CaffeNet.cpp:325] 0-th Socket addr:
I0323 16:51:23.076701 1378 CaffeNet.cpp:325] 1-th Socket addr: dlgpu19.ai.bjcc.qihoo.net:57402
I0323 16:51:23.076707 1378 CaffeNet.cpp:325] 2-th Socket addr: dlgpu19.ai.bjcc.qihoo.net:57402
I0323 16:51:23.076711 1378 CaffeNet.cpp:325] 3-th Socket addr: dlgpu19.ai.bjcc.qihoo.net:57402
I0323 16:51:23.076715 1378 CaffeNet.cpp:325] 4-th Socket addr: dlgpu19.ai.bjcc.qihoo.net:57402
I0323 16:51:23.076730 1378 JniCaffeNet.cpp:110] 0-th local addr:
I0323 16:51:23.076736 1378 JniCaffeNet.cpp:110] 1-th local addr: dlgpu19.ai.bjcc.qihoo.net:57402
I0323 16:51:23.076740 1378 JniCaffeNet.cpp:110] 2-th local addr: dlgpu19.ai.bjcc.qihoo.net:57402
I0323 16:51:23.076745 1378 JniCaffeNet.cpp:110] 3-th local addr: dlgpu19.ai.bjcc.qihoo.net:57402
I0323 16:51:23.076748 1378 JniCaffeNet.cpp:110] 4-th local addr: dlgpu19.ai.bjcc.qihoo.net:57402
16/03/23 16:51:23 INFO executor.Executor: Finished task 0.0 in stage 0.0 (TID 0). 1069 bytes result sent to driver
16/03/23 16:51:27 INFO executor.CoarseGrainedExecutorBackend: Got assigned task 5
16/03/23 16:51:27 INFO executor.Executor: Running task 0.0 in stage 1.0 (TID 5)
16/03/23 16:51:27 INFO broadcast.TorrentBroadcast: Started reading broadcast variable 2
16/03/23 16:51:27 INFO storage.MemoryStore: Block broadcast_2_piece0 stored as bytes in memory (estimated size 1622.0 B, free 6.9 KB)
16/03/23 16:51:27 INFO broadcast.TorrentBroadcast: Reading broadcast variable 2 took 17 ms
16/03/23 16:51:27 INFO storage.MemoryStore: Block broadcast_2 stored as values in memory (estimated size 2.6 KB, free 9.5 KB)
16/03/23 16:51:27 INFO broadcast.TorrentBroadcast: Started reading broadcast variable 1
16/03/23 16:51:27 INFO storage.MemoryStore: Block broadcast_1_piece0 stored as bytes in memory (estimated size 235.0 B, free 9.7 KB)
16/03/23 16:51:27 INFO broadcast.TorrentBroadcast: Reading broadcast variable 1 took 13 ms
I0323 16:51:27.553076 1414 socket.cpp:189] Accepted the connection from client [dlgpu10.ai.bjcc.qihoo.net]
I0323 16:51:27.554393 1414 socket.cpp:189] Accepted the connection from client [dlgpu10.ai.bjcc.qihoo.net]
16/03/23 16:51:27 INFO storage.MemoryStore: Block broadcast_1 stored as values in memory (estimated size 3.1 KB, free 12.8 KB)
I0323 16:51:27.556990 1414 socket.cpp:189] Accepted the connection from client [dlgpu20.ai.bjcc.qihoo.net]
I0323 16:51:27.558328 1378 common.cpp:61] 0-th string is NULL
I0323 16:51:27.558416 1378 socket.cpp:255] Trying to connect with ...[dlgpu19.ai.bjcc.qihoo.net:41630]
I0323 16:51:27.558573 1378 socket.cpp:314] Connected to server [dlgpu19.ai.bjcc.qihoo.net:41630] with client_fd [318]
I0323 16:51:27.560863 1414 socket.cpp:189] Accepted the connection from client [dlgpu19.ai.bjcc.qihoo.net]
I0323 16:51:37.558701 1378 socket.cpp:255] Trying to connect with ...[dlgpu10.ai.bjcc.qihoo.net:58067]
I0323 16:51:37.559094 1378 socket.cpp:314] Connected to server [dlgpu10.ai.bjcc.qihoo.net:58067] with client_fd [320]
I0323 16:51:47.559253 1378 socket.cpp:255] Trying to connect with ...[dlgpu20.ai.bjcc.qihoo.net:47155]
I0323 16:51:47.559638 1378 socket.cpp:314] Connected to server [dlgpu20.ai.bjcc.qihoo.net:47155] with client_fd [321]
I0323 16:51:57.559800 1378 socket.cpp:255] Trying to connect with ...[dlgpu10.ai.bjcc.qihoo.net:46110]
I0323 16:51:57.560225 1378 socket.cpp:314] Connected to server [dlgpu10.ai.bjcc.qihoo.net:46110] with client_fd [322]
I0323 16:52:07.604482 1378 parallel.cpp:392] GPUs pairs 0:1, 2:3, 0:2, 0:4
I0323 16:52:07.675818 1378 parallel.cpp:234] GPU 4 does not have p2p access to GPU 0
I0323 16:52:07.740978 1756 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
16/03/23 16:52:08 INFO executor.Executor: Finished task 0.0 in stage 1.0 (TID 5). 918 bytes result sent to driver
I0323 16:52:08.425240 1758 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
I0323 16:52:08.427775 1763 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
I0323 16:52:08.447262 1772 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
I0323 16:52:08.454452 1774 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
16/03/23 16:52:08 INFO executor.CoarseGrainedExecutorBackend: Got assigned task 12
16/03/23 16:52:08 INFO executor.Executor: Running task 2.0 in stage 2.0 (TID 12)
16/03/23 16:52:08 INFO broadcast.TorrentBroadcast: Started reading broadcast variable 3
16/03/23 16:52:08 INFO storage.MemoryStore: Block broadcast_3_piece0 stored as bytes in memory (estimated size 1450.0 B, free 4.7 KB)
16/03/23 16:52:08 INFO broadcast.TorrentBroadcast: Reading broadcast variable 3 took 18 ms
16/03/23 16:52:08 INFO storage.MemoryStore: Block broadcast_3 stored as values in memory (estimated size 2.2 KB, free 6.9 KB)
16/03/23 16:52:08 INFO spark.CacheManager: Partition rdd_5_2 not found, computing it
16/03/23 16:52:08 INFO spark.CacheManager: Partition rdd_0_2 not found, computing it
16/03/23 16:52:08 INFO caffe.LmdbRDD: Processing partition 2
16/03/23 16:52:09 INFO caffe.LmdbRDD: Completed partition 2
16/03/23 16:52:09 INFO storage.BlockManager: Found block rdd_0_2 locally
I0323 16:52:10.079346 1424 socket.cpp:58] receive message: 3 0
I0323 16:52:10.079407 1424 socket.cpp:59] socket: 12 message header size: 12
I0323 16:52:10.159688 1378 socket.cpp:44] send essage: 0 0 344864
I0323 16:52:10.161568 1378 socket.cpp:44] send essage: 0 0 344864
I0323 16:52:10.162847 1427 socket.cpp:58] receive message: 1 0
I0323 16:52:10.162863 1427 socket.cpp:59] socket: 12 message header size: 12
I0323 16:52:10.169373 1378 socket.cpp:44] send essage: 0 0 344864
I0323 16:52:10.214555 1378 socket.cpp:44] send essage: 0 0 344864
I0323 16:52:10.938413 1423 socket.cpp:58] receive message: 2 0
I0323 16:52:10.938458 1423 socket.cpp:59] socket: 12 message header size: 12
I0323 16:52:11.040642 1422 socket.cpp:58] receive message: 4 0
I0323 16:52:11.040699 1422 socket.cpp:59] socket: 12 message header size: 12
16/03/23 16:53:58 INFO executor.CoarseGrainedExecutorBackend: Driver commanded a shutdown
16/03/23 16:53:58 INFO storage.MemoryStore: MemoryStore cleared
16/03/23 16:53:58 INFO storage.BlockManager: BlockManager stopped
16/03/23 16:53:58 WARN executor.CoarseGrainedExecutorBackend: An unknown (dlgpu20.ai.bjcc.qihoo.net:45855) driver disconnected.
16/03/23 16:53:58 ERROR executor.CoarseGrainedExecutorBackend: Driver 10.142.118.172:45855 disassociated! Shutting down.
16/03/23 16:53:58 INFO util.ShutdownHookManager: Shutdown hook called
16/03/23 16:53:58 INFO remote.RemoteActorRefProvider$RemotingTerminator: Shutting down remote daemon.
LogType:stdout
Log Upload Time:23-Mar-2016 16:53:59
LogLength:0
Log Contents:
Container: container_1458635993280_0010_01_000004 on dlgpu19.ai.bjcc.qihoo.net_33935
LogType:stderr
Log Upload Time:23-Mar-2016 16:53:59
LogLength:32045
Log Contents:
SLF4J: Class path contains multiple SLF4J bindings.
SLF4J: Found binding in [jar:file:/dev/hadoop/nm-local-dir/usercache/hadoop/filecache/11/spark-assembly-1.6.0-hadoop2.6.4-U4.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: Found binding in [jar:file:/home/hadoop/software/hadoop-2.6.4/share/hadoop/common/lib/slf4j-log4j12-1.7.5.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: See http://www.slf4j.org/codes.html#multiple_bindings for an explanation.
SLF4J: Actual binding is of type [org.slf4j.impl.Log4jLoggerFactory]
16/03/23 16:51:13 INFO executor.CoarseGrainedExecutorBackend: Registered signal handlers for [TERM, HUP, INT]
16/03/23 16:51:13 INFO yarn.YarnSparkHadoopUtil: Set ugi in YarnConf to: hadoop,hadoop
16/03/23 16:51:14 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
16/03/23 16:51:14 INFO yarn.YarnSparkHadoopUtil: running as user: hadoop
16/03/23 16:51:14 INFO spark.SecurityManager: Changing view acls to: hadoop
16/03/23 16:51:14 INFO spark.SecurityManager: Changing modify acls to: hadoop
16/03/23 16:51:14 INFO spark.SecurityManager: SecurityManager: authentication disabled; ui acls disabled; users with view permissions: Set(hadoop); users with modify permissions: Set(hadoop)
16/03/23 16:51:14 INFO spark.SecurityManager: Changing view acls to: hadoop
16/03/23 16:51:14 INFO spark.SecurityManager: Changing modify acls to: hadoop
16/03/23 16:51:14 INFO spark.SecurityManager: SecurityManager: authentication disabled; ui acls disabled; users with view permissions: Set(hadoop); users with modify permissions: Set(hadoop)
16/03/23 16:51:15 INFO slf4j.Slf4jLogger: Slf4jLogger started
16/03/23 16:51:15 INFO Remoting: Starting remoting
16/03/23 16:51:15 INFO Remoting: Remoting started; listening on addresses :[akka.tcp://[email protected]:35824]
16/03/23 16:51:15 INFO util.Utils: Successfully started service 'sparkExecutorActorSystem' on port 35824.
16/03/23 16:51:15 INFO storage.DiskBlockManager: Created local directory at /home/hadoop/yarn/nm-local-dir/usercache/hadoop/appcache/application_1458635993280_0010/blockmgr-b5a31fe0-8b93-47bd-b631-81a11b0a4124
16/03/23 16:51:15 INFO storage.DiskBlockManager: Created local directory at /dev/hadoop/nm-local-dir/usercache/hadoop/appcache/application_1458635993280_0010/blockmgr-159de318-affd-49f0-839f-46ea110a371b
16/03/23 16:51:15 INFO storage.DiskBlockManager: Created local directory at /run/hadoop/nm-local-dir/usercache/hadoop/appcache/application_1458635993280_0010/blockmgr-1801ddca-5d6e-4b58-913e-cf72dd8cb4d8
16/03/23 16:51:15 INFO storage.MemoryStore: MemoryStore started with capacity 24.9 GB
16/03/23 16:51:15 INFO yarn.YarnSparkHadoopUtil: Set ugi in YarnConf to: hadoop,hadoop
16/03/23 16:51:15 INFO executor.CoarseGrainedExecutorBackend: Connecting to driver: spark://[email protected]:45855
16/03/23 16:51:15 INFO executor.CoarseGrainedExecutorBackend: Successfully registered with driver
16/03/23 16:51:15 INFO executor.Executor: Starting executor ID 3 on host dlgpu19.ai.bjcc.qihoo.net
16/03/23 16:51:15 INFO util.Utils: Successfully started service 'org.apache.spark.network.netty.NettyBlockTransferService' on port 41146.
16/03/23 16:51:15 INFO netty.NettyBlockTransferService: Server created on 41146
16/03/23 16:51:15 INFO storage.BlockManager: external shuffle service port = 7337
16/03/23 16:51:15 INFO storage.BlockManagerMaster: Trying to register BlockManager
16/03/23 16:51:15 INFO storage.BlockManagerMaster: Registered BlockManager
16/03/23 16:51:15 INFO storage.BlockManager: Registering executor with local external shuffle service.
16/03/23 16:51:19 INFO executor.CoarseGrainedExecutorBackend: Got assigned task 1
16/03/23 16:51:19 INFO executor.Executor: Running task 1.0 in stage 0.0 (TID 1)
16/03/23 16:51:19 INFO broadcast.TorrentBroadcast: Started reading broadcast variable 0
16/03/23 16:51:20 INFO storage.MemoryStore: Block broadcast_0_piece0 stored as bytes in memory (estimated size 2.1 KB, free 2.1 KB)
16/03/23 16:51:20 INFO broadcast.TorrentBroadcast: Reading broadcast variable 0 took 283 ms
16/03/23 16:51:20 INFO storage.MemoryStore: Block broadcast_0 stored as values in memory (estimated size 3.2 KB, free 5.3 KB)
16/03/23 16:51:20 INFO caffe.CaffeProcessor: my rank is 1
16/03/23 16:51:20 INFO caffe.LMDB: Batch size:64
WARNING: Logging before InitGoogleLogging() is written to STDERR
I0323 16:51:22.752178 1377 CaffeNet.cpp:78] set root solver device id to 0
I0323 16:51:22.980456 1377 solver.cpp:48] Initializing solver from parameters:
test_iter: 1
test_interval: 10001
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: 10001
snapshot_prefix: "mnist_lenet"
solver_mode: GPU
device_id: 0
net_param {
name: "LeNet"
layer {
name: "data"
type: "MemoryData"
top: "data"
top: "label"
include {
phase: TRAIN
}
memory_data_param {
batch_size: 64
channels: 1
height: 28
width: 28
share_in_parallel: false
source: "hdfs:///projects/machine_learning/image_dataset/mnist_train_lmdb/"
}
source_class: "com.yahoo.ml.caffe.LMDB"
}
layer {
name: "data"
type: "MemoryData"
top: "data"
top: "label"
include {
phase: TEST
}
memory_data_param {
batch_size: 100
channels: 1
height: 28
width: 28
share_in_parallel: false
source: "hdfs:///projects/machine_learning/image_dataset/mnist_test_lmdb/"
}
source_class: "com.yahoo.ml.caffe.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"
}
}
test_initialization: false
I0323 16:51:22.980608 1377 solver.cpp:86] Creating training net specified in net_param.
I0323 16:51:22.980680 1377 net.cpp:322] The NetState phase (0) differed from the phase (1) specified by a rule in layer data
I0323 16:51:22.980692 1377 net.cpp:322] The NetState phase (0) differed from the phase (1) specified by a rule in layer accuracy
I0323 16:51:22.980774 1377 net.cpp:49] Initializing net from parameters:
name: "LeNet"
state {
phase: TRAIN
}
layer {
name: "data"
type: "MemoryData"
top: "data"
top: "label"
include {
phase: TRAIN
}
memory_data_param {
batch_size: 64
channels: 1
height: 28
width: 28
share_in_parallel: false
source: "hdfs:///projects/machine_learning/image_dataset/mnist_train_lmdb/"
}
source_class: "com.yahoo.ml.caffe.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"
}
I0323 16:51:22.980877 1377 layer_factory.hpp:77] Creating layer data
I0323 16:51:22.980906 1377 net.cpp:106] Creating Layer data
I0323 16:51:22.980917 1377 net.cpp:411] data -> data
I0323 16:51:22.980954 1377 net.cpp:411] data -> label
I0323 16:51:22.984134 1377 net.cpp:150] Setting up data
I0323 16:51:22.984165 1377 net.cpp:157] Top shape: 64 1 28 28 (50176)
I0323 16:51:22.984174 1377 net.cpp:157] Top shape: 64 (64)
I0323 16:51:22.984177 1377 net.cpp:165] Memory required for data: 200960
I0323 16:51:22.984187 1377 layer_factory.hpp:77] Creating layer conv1
I0323 16:51:22.984207 1377 net.cpp:106] Creating Layer conv1
I0323 16:51:22.984215 1377 net.cpp:454] conv1 <- data
I0323 16:51:22.984228 1377 net.cpp:411] conv1 -> conv1
I0323 16:51:22.986796 1377 net.cpp:150] Setting up conv1
I0323 16:51:22.986815 1377 net.cpp:157] Top shape: 64 20 24 24 (737280)
I0323 16:51:22.986821 1377 net.cpp:165] Memory required for data: 3150080
I0323 16:51:22.986840 1377 layer_factory.hpp:77] Creating layer pool1
I0323 16:51:22.986851 1377 net.cpp:106] Creating Layer pool1
I0323 16:51:22.986856 1377 net.cpp:454] pool1 <- conv1
I0323 16:51:22.986863 1377 net.cpp:411] pool1 -> pool1
I0323 16:51:22.986991 1377 net.cpp:150] Setting up pool1
I0323 16:51:22.987004 1377 net.cpp:157] Top shape: 64 20 12 12 (184320)
I0323 16:51:22.987009 1377 net.cpp:165] Memory required for data: 3887360
I0323 16:51:22.987013 1377 layer_factory.hpp:77] Creating layer conv2
I0323 16:51:22.987023 1377 net.cpp:106] Creating Layer conv2
I0323 16:51:22.987027 1377 net.cpp:454] conv2 <- pool1
I0323 16:51:22.987035 1377 net.cpp:411] conv2 -> conv2
I0323 16:51:22.987975 1377 net.cpp:150] Setting up conv2
I0323 16:51:22.987990 1377 net.cpp:157] Top shape: 64 50 8 8 (204800)
I0323 16:51:22.987995 1377 net.cpp:165] Memory required for data: 4706560
I0323 16:51:22.988005 1377 layer_factory.hpp:77] Creating layer pool2
I0323 16:51:22.988013 1377 net.cpp:106] Creating Layer pool2
I0323 16:51:22.988018 1377 net.cpp:454] pool2 <- conv2
I0323 16:51:22.988024 1377 net.cpp:411] pool2 -> pool2
I0323 16:51:22.988122 1377 net.cpp:150] Setting up pool2
I0323 16:51:22.988136 1377 net.cpp:157] Top shape: 64 50 4 4 (51200)
I0323 16:51:22.988140 1377 net.cpp:165] Memory required for data: 4911360
I0323 16:51:22.988144 1377 layer_factory.hpp:77] Creating layer ip1
I0323 16:51:22.988157 1377 net.cpp:106] Creating Layer ip1
I0323 16:51:22.988160 1377 net.cpp:454] ip1 <- pool2
I0323 16:51:22.988168 1377 net.cpp:411] ip1 -> ip1
I0323 16:51:22.993342 1377 net.cpp:150] Setting up ip1
I0323 16:51:22.993361 1377 net.cpp:157] Top shape: 64 500 (32000)
I0323 16:51:22.993366 1377 net.cpp:165] Memory required for data: 5039360
I0323 16:51:22.993377 1377 layer_factory.hpp:77] Creating layer relu1
I0323 16:51:22.993387 1377 net.cpp:106] Creating Layer relu1
I0323 16:51:22.993393 1377 net.cpp:454] relu1 <- ip1
I0323 16:51:22.993399 1377 net.cpp:397] relu1 -> ip1 (in-place)
I0323 16:51:22.993410 1377 net.cpp:150] Setting up relu1
I0323 16:51:22.993415 1377 net.cpp:157] Top shape: 64 500 (32000)
I0323 16:51:22.993419 1377 net.cpp:165] Memory required for data: 5167360
I0323 16:51:22.993423 1377 layer_factory.hpp:77] Creating layer ip2
I0323 16:51:22.993430 1377 net.cpp:106] Creating Layer ip2
I0323 16:51:22.993434 1377 net.cpp:454] ip2 <- ip1
I0323 16:51:22.993441 1377 net.cpp:411] ip2 -> ip2
I0323 16:51:22.995192 1377 net.cpp:150] Setting up ip2
I0323 16:51:22.995210 1377 net.cpp:157] Top shape: 64 10 (640)
I0323 16:51:22.995216 1377 net.cpp:165] Memory required for data: 5169920
I0323 16:51:22.995225 1377 layer_factory.hpp:77] Creating layer loss
I0323 16:51:22.995237 1377 net.cpp:106] Creating Layer loss
I0323 16:51:22.995242 1377 net.cpp:454] loss <- ip2
I0323 16:51:22.995247 1377 net.cpp:454] loss <- label
I0323 16:51:22.995256 1377 net.cpp:411] loss -> loss
I0323 16:51:22.995275 1377 layer_factory.hpp:77] Creating layer loss
I0323 16:51:22.995570 1377 net.cpp:150] Setting up loss
I0323 16:51:22.995580 1377 net.cpp:157] Top shape: (1)
I0323 16:51:22.995584 1377 net.cpp:160] with loss weight 1
I0323 16:51:22.995605 1377 net.cpp:165] Memory required for data: 5169924
I0323 16:51:22.995610 1377 net.cpp:226] loss needs backward computation.
I0323 16:51:22.995615 1377 net.cpp:226] ip2 needs backward computation.
I0323 16:51:22.995620 1377 net.cpp:226] relu1 needs backward computation.
I0323 16:51:22.995623 1377 net.cpp:226] ip1 needs backward computation.
I0323 16:51:22.995626 1377 net.cpp:226] pool2 needs backward computation.
I0323 16:51:22.995630 1377 net.cpp:226] conv2 needs backward computation.
I0323 16:51:22.995635 1377 net.cpp:226] pool1 needs backward computation.
I0323 16:51:22.995638 1377 net.cpp:226] conv1 needs backward computation.
I0323 16:51:22.995643 1377 net.cpp:228] data does not need backward computation.
I0323 16:51:22.995646 1377 net.cpp:270] This network produces output loss
I0323 16:51:22.995657 1377 net.cpp:283] Network initialization done.
I0323 16:51:22.995712 1377 solver.cpp:181] Creating test net (#0) specified by net_param
I0323 16:51:22.995733 1377 net.cpp:322] The NetState phase (1) differed from the phase (0) specified by a rule in layer data
I0323 16:51:22.995823 1377 net.cpp:49] Initializing net from parameters:
name: "LeNet"
state {
phase: TEST
}
layer {
name: "data"
type: "MemoryData"
top: "data"
top: "label"
include {
phase: TEST
}
memory_data_param {
batch_size: 100
channels: 1
height: 28
width: 28
share_in_parallel: false
source: "hdfs:///projects/machine_learning/image_dataset/mnist_test_lmdb/"
}
source_class: "com.yahoo.ml.caffe.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"
}
I0323 16:51:22.995904 1377 layer_factory.hpp:77] Creating layer data
I0323 16:51:22.995915 1377 net.cpp:106] Creating Layer data
I0323 16:51:22.995921 1377 net.cpp:411] data -> data
I0323 16:51:22.995930 1377 net.cpp:411] data -> label
I0323 16:51:22.997612 1377 net.cpp:150] Setting up data
I0323 16:51:22.997632 1377 net.cpp:157] Top shape: 100 1 28 28 (78400)
I0323 16:51:22.997638 1377 net.cpp:157] Top shape: 100 (100)
I0323 16:51:22.997643 1377 net.cpp:165] Memory required for data: 314000
I0323 16:51:22.997648 1377 layer_factory.hpp:77] Creating layer label_data_1_split
I0323 16:51:22.997659 1377 net.cpp:106] Creating Layer label_data_1_split
I0323 16:51:22.997664 1377 net.cpp:454] label_data_1_split <- label
I0323 16:51:22.997670 1377 net.cpp:411] label_data_1_split -> label_data_1_split_0
I0323 16:51:22.997679 1377 net.cpp:411] label_data_1_split -> label_data_1_split_1
I0323 16:51:22.997789 1377 net.cpp:150] Setting up label_data_1_split
I0323 16:51:22.997797 1377 net.cpp:157] Top shape: 100 (100)
I0323 16:51:22.997802 1377 net.cpp:157] Top shape: 100 (100)
I0323 16:51:22.997807 1377 net.cpp:165] Memory required for data: 314800
I0323 16:51:22.997810 1377 layer_factory.hpp:77] Creating layer conv1
I0323 16:51:22.997822 1377 net.cpp:106] Creating Layer conv1
I0323 16:51:22.997827 1377 net.cpp:454] conv1 <- data
I0323 16:51:22.997833 1377 net.cpp:411] conv1 -> conv1
I0323 16:51:22.998605 1377 net.cpp:150] Setting up conv1
I0323 16:51:22.998620 1377 net.cpp:157] Top shape: 100 20 24 24 (1152000)
I0323 16:51:22.998625 1377 net.cpp:165] Memory required for data: 4922800
I0323 16:51:22.998636 1377 layer_factory.hpp:77] Creating layer pool1
I0323 16:51:22.998642 1377 net.cpp:106] Creating Layer pool1
I0323 16:51:22.998647 1377 net.cpp:454] pool1 <- conv1
I0323 16:51:22.998653 1377 net.cpp:411] pool1 -> pool1
I0323 16:51:22.998760 1377 net.cpp:150] Setting up pool1
I0323 16:51:22.998769 1377 net.cpp:157] Top shape: 100 20 12 12 (288000)
I0323 16:51:22.998772 1377 net.cpp:165] Memory required for data: 6074800
I0323 16:51:22.998777 1377 layer_factory.hpp:77] Creating layer conv2
I0323 16:51:22.998787 1377 net.cpp:106] Creating Layer conv2
I0323 16:51:22.998792 1377 net.cpp:454] conv2 <- pool1
I0323 16:51:22.998800 1377 net.cpp:411] conv2 -> conv2
I0323 16:51:22.999790 1377 net.cpp:150] Setting up conv2
I0323 16:51:22.999804 1377 net.cpp:157] Top shape: 100 50 8 8 (320000)
I0323 16:51:22.999809 1377 net.cpp:165] Memory required for data: 7354800
I0323 16:51:22.999819 1377 layer_factory.hpp:77] Creating layer pool2
I0323 16:51:22.999825 1377 net.cpp:106] Creating Layer pool2
I0323 16:51:22.999830 1377 net.cpp:454] pool2 <- conv2
I0323 16:51:22.999836 1377 net.cpp:411] pool2 -> pool2
I0323 16:51:22.999944 1377 net.cpp:150] Setting up pool2
I0323 16:51:22.999956 1377 net.cpp:157] Top shape: 100 50 4 4 (80000)
I0323 16:51:22.999968 1377 net.cpp:165] Memory required for data: 7674800
I0323 16:51:22.999972 1377 layer_factory.hpp:77] Creating layer ip1
I0323 16:51:22.999980 1377 net.cpp:106] Creating Layer ip1
I0323 16:51:22.999985 1377 net.cpp:454] ip1 <- pool2
I0323 16:51:22.999992 1377 net.cpp:411] ip1 -> ip1
I0323 16:51:23.005182 1377 net.cpp:150] Setting up ip1
I0323 16:51:23.005201 1377 net.cpp:157] Top shape: 100 500 (50000)
I0323 16:51:23.005208 1377 net.cpp:165] Memory required for data: 7874800
I0323 16:51:23.005218 1377 layer_factory.hpp:77] Creating layer relu1
I0323 16:51:23.005225 1377 net.cpp:106] Creating Layer relu1
I0323 16:51:23.005229 1377 net.cpp:454] relu1 <- ip1
I0323 16:51:23.005235 1377 net.cpp:397] relu1 -> ip1 (in-place)
I0323 16:51:23.005244 1377 net.cpp:150] Setting up relu1
I0323 16:51:23.005249 1377 net.cpp:157] Top shape: 100 500 (50000)
I0323 16:51:23.005252 1377 net.cpp:165] Memory required for data: 8074800
I0323 16:51:23.005256 1377 layer_factory.hpp:77] Creating layer ip2
I0323 16:51:23.005264 1377 net.cpp:106] Creating Layer ip2
I0323 16:51:23.005269 1377 net.cpp:454] ip2 <- ip1
I0323 16:51:23.005275 1377 net.cpp:411] ip2 -> ip2
I0323 16:51:23.005650 1377 net.cpp:150] Setting up ip2
I0323 16:51:23.005661 1377 net.cpp:157] Top shape: 100 10 (1000)
I0323 16:51:23.005664 1377 net.cpp:165] Memory required for data: 8078800
I0323 16:51:23.005671 1377 layer_factory.hpp:77] Creating layer ip2_ip2_0_split
I0323 16:51:23.005678 1377 net.cpp:106] Creating Layer ip2_ip2_0_split
I0323 16:51:23.005683 1377 net.cpp:454] ip2_ip2_0_split <- ip2
I0323 16:51:23.005689 1377 net.cpp:411] ip2_ip2_0_split -> ip2_ip2_0_split_0
I0323 16:51:23.005697 1377 net.cpp:411] ip2_ip2_0_split -> ip2_ip2_0_split_1
I0323 16:51:23.005798 1377 net.cpp:150] Setting up ip2_ip2_0_split
I0323 16:51:23.005806 1377 net.cpp:157] Top shape: 100 10 (1000)
I0323 16:51:23.005811 1377 net.cpp:157] Top shape: 100 10 (1000)
I0323 16:51:23.005815 1377 net.cpp:165] Memory required for data: 8086800
I0323 16:51:23.005820 1377 layer_factory.hpp:77] Creating layer accuracy
I0323 16:51:23.005830 1377 net.cpp:106] Creating Layer accuracy
I0323 16:51:23.005833 1377 net.cpp:454] accuracy <- ip2_ip2_0_split_0
I0323 16:51:23.005838 1377 net.cpp:454] accuracy <- label_data_1_split_0
I0323 16:51:23.005846 1377 net.cpp:411] accuracy -> accuracy
I0323 16:51:23.005856 1377 net.cpp:150] Setting up accuracy
I0323 16:51:23.005862 1377 net.cpp:157] Top shape: (1)
I0323 16:51:23.005867 1377 net.cpp:165] Memory required for data: 8086804
I0323 16:51:23.005870 1377 layer_factory.hpp:77] Creating layer loss
I0323 16:51:23.005877 1377 net.cpp:106] Creating Layer loss
I0323 16:51:23.005880 1377 net.cpp:454] loss <- ip2_ip2_0_split_1
I0323 16:51:23.005885 1377 net.cpp:454] loss <- label_data_1_split_1
I0323 16:51:23.005892 1377 net.cpp:411] loss -> loss
I0323 16:51:23.005899 1377 layer_factory.hpp:77] Creating layer loss
I0323 16:51:23.006204 1377 net.cpp:150] Setting up loss
I0323 16:51:23.006217 1377 net.cpp:157] Top shape: (1)
I0323 16:51:23.006222 1377 net.cpp:160] with loss weight 1
I0323 16:51:23.006229 1377 net.cpp:165] Memory required for data: 8086808
I0323 16:51:23.006233 1377 net.cpp:226] loss needs backward computation.
I0323 16:51:23.006238 1377 net.cpp:228] accuracy does not need backward computation.
I0323 16:51:23.006243 1377 net.cpp:226] ip2_ip2_0_split needs backward computation.
I0323 16:51:23.006247 1377 net.cpp:226] ip2 needs backward computation.
I0323 16:51:23.006252 1377 net.cpp:226] relu1 needs backward computation.
I0323 16:51:23.006255 1377 net.cpp:226] ip1 needs backward computation.
I0323 16:51:23.006259 1377 net.cpp:226] pool2 needs backward computation.
I0323 16:51:23.006263 1377 net.cpp:226] conv2 needs backward computation.
I0323 16:51:23.006268 1377 net.cpp:226] pool1 needs backward computation.
I0323 16:51:23.006271 1377 net.cpp:226] conv1 needs backward computation.
I0323 16:51:23.006276 1377 net.cpp:228] label_data_1_split does not need backward computation.
I0323 16:51:23.006288 1377 net.cpp:228] data does not need backward computation.
I0323 16:51:23.006291 1377 net.cpp:270] This network produces output accuracy
I0323 16:51:23.006295 1377 net.cpp:270] This network produces output loss
I0323 16:51:23.006310 1377 net.cpp:283] Network initialization done.
I0323 16:51:23.006355 1377 solver.cpp:60] Solver scaffolding done.
I0323 16:51:23.007391 1377 socket.cpp:224] Waiting for valid port [0]
I0323 16:51:23.007429 1413 socket.cpp:163] Assigned socket server port [41630]
I0323 16:51:23.009763 1413 socket.cpp:176] Socket Server ready [0.0.0.0]
I0323 16:51:23.017469 1377 socket.cpp:224] Waiting for valid port [41630]
I0323 16:51:23.017478 1377 socket.cpp:232] Valid port found [41630]
I0323 16:51:23.017490 1377 CaffeNet.cpp:186] Socket adapter: dlgpu19.ai.bjcc.qihoo.net:41630
I0323 16:51:23.017716 1377 CaffeNet.cpp:325] 0-th Socket addr: dlgpu19.ai.bjcc.qihoo.net:41630
I0323 16:51:23.017729 1377 CaffeNet.cpp:325] 1-th Socket addr:
I0323 16:51:23.017734 1377 CaffeNet.cpp:325] 2-th Socket addr: dlgpu19.ai.bjcc.qihoo.net:41630
I0323 16:51:23.017737 1377 CaffeNet.cpp:325] 3-th Socket addr: dlgpu19.ai.bjcc.qihoo.net:41630
I0323 16:51:23.017742 1377 CaffeNet.cpp:325] 4-th Socket addr: dlgpu19.ai.bjcc.qihoo.net:41630
I0323 16:51:23.017755 1377 JniCaffeNet.cpp:110] 0-th local addr: dlgpu19.ai.bjcc.qihoo.net:41630
I0323 16:51:23.017761 1377 JniCaffeNet.cpp:110] 1-th local addr:
I0323 16:51:23.017765 1377 JniCaffeNet.cpp:110] 2-th local addr: dlgpu19.ai.bjcc.qihoo.net:41630
I0323 16:51:23.017768 1377 JniCaffeNet.cpp:110] 3-th local addr: dlgpu19.ai.bjcc.qihoo.net:41630
I0323 16:51:23.017772 1377 JniCaffeNet.cpp:110] 4-th local addr: dlgpu19.ai.bjcc.qihoo.net:41630
16/03/23 16:51:23 INFO executor.Executor: Finished task 1.0 in stage 0.0 (TID 1). 1069 bytes result sent to driver
16/03/23 16:51:27 INFO executor.CoarseGrainedExecutorBackend: Got assigned task 9
16/03/23 16:51:27 INFO executor.Executor: Running task 4.0 in stage 1.0 (TID 9)
16/03/23 16:51:27 INFO broadcast.TorrentBroadcast: Started reading broadcast variable 2
16/03/23 16:51:27 INFO storage.MemoryStore: Block broadcast_2_piece0 stored as bytes in memory (estimated size 1622.0 B, free 6.9 KB)
16/03/23 16:51:27 INFO broadcast.TorrentBroadcast: Reading broadcast variable 2 took 16 ms
16/03/23 16:51:27 INFO storage.MemoryStore: Block broadcast_2 stored as values in memory (estimated size 2.6 KB, free 9.5 KB)
16/03/23 16:51:27 INFO broadcast.TorrentBroadcast: Started reading broadcast variable 1
16/03/23 16:51:27 INFO storage.MemoryStore: Block broadcast_1_piece0 stored as bytes in memory (estimated size 235.0 B, free 9.7 KB)
16/03/23 16:51:27 INFO broadcast.TorrentBroadcast: Reading broadcast variable 1 took 13 ms
16/03/23 16:51:27 INFO storage.MemoryStore: Block broadcast_1 stored as values in memory (estimated size 3.1 KB, free 12.8 KB)
I0323 16:51:27.558655 1413 socket.cpp:189] Accepted the connection from client [dlgpu19.ai.bjcc.qihoo.net]
I0323 16:51:27.560506 1377 common.cpp:61] 1-th string is NULL
I0323 16:51:27.560600 1377 socket.cpp:255] Trying to connect with ...[dlgpu19.ai.bjcc.qihoo.net:57402]
I0323 16:51:27.560781 1377 socket.cpp:314] Connected to server [dlgpu19.ai.bjcc.qihoo.net:57402] with client_fd [316]
I0323 16:51:37.553697 1413 socket.cpp:189] Accepted the connection from client [dlgpu10.ai.bjcc.qihoo.net]
I0323 16:51:37.554890 1413 socket.cpp:189] Accepted the connection from client [dlgpu10.ai.bjcc.qihoo.net]
I0323 16:51:37.557639 1413 socket.cpp:189] Accepted the connection from client [dlgpu20.ai.bjcc.qihoo.net]
I0323 16:51:37.560904 1377 socket.cpp:255] Trying to connect with ...[dlgpu10.ai.bjcc.qihoo.net:58067]
I0323 16:51:37.561306 1377 socket.cpp:314] Connected to server [dlgpu10.ai.bjcc.qihoo.net:58067] with client_fd [320]
I0323 16:51:47.561446 1377 socket.cpp:255] Trying to connect with ...[dlgpu20.ai.bjcc.qihoo.net:47155]
I0323 16:51:47.561790 1377 socket.cpp:314] Connected to server [dlgpu20.ai.bjcc.qihoo.net:47155] with client_fd [321]
I0323 16:51:57.561954 1377 socket.cpp:255] Trying to connect with ...[dlgpu10.ai.bjcc.qihoo.net:46110]
I0323 16:51:57.562314 1377 socket.cpp:314] Connected to server [dlgpu10.ai.bjcc.qihoo.net:46110] with client_fd [322]
I0323 16:52:07.606834 1377 parallel.cpp:392] GPUs pairs 0:1, 2:3, 0:2, 0:4
I0323 16:52:07.678092 1377 parallel.cpp:234] GPU 4 does not have p2p access to GPU 0
I0323 16:52:07.715546 1759 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
16/03/23 16:52:07 INFO executor.Executor: Finished task 4.0 in stage 1.0 (TID 9). 918 bytes result sent to driver
I0323 16:52:08.365959 1761 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
I0323 16:52:08.395136 1765 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
I0323 16:52:08.398850 1767 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
I0323 16:52:08.408880 1769 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
16/03/23 16:52:08 INFO executor.CoarseGrainedExecutorBackend: Got assigned task 13
16/03/23 16:52:08 INFO executor.Executor: Running task 3.0 in stage 2.0 (TID 13)
16/03/23 16:52:08 INFO broadcast.TorrentBroadcast: Started reading broadcast variable 3
16/03/23 16:52:08 INFO storage.MemoryStore: Block broadcast_3_piece0 stored as bytes in memory (estimated size 1450.0 B, free 4.7 KB)
16/03/23 16:52:08 INFO broadcast.TorrentBroadcast: Reading broadcast variable 3 took 17 ms
16/03/23 16:52:08 INFO storage.MemoryStore: Block broadcast_3 stored as values in memory (estimated size 2.2 KB, free 6.9 KB)
16/03/23 16:52:08 INFO spark.CacheManager: Partition rdd_5_3 not found, computing it
16/03/23 16:52:08 INFO spark.CacheManager: Partition rdd_0_3 not found, computing it
16/03/23 16:52:08 INFO caffe.LmdbRDD: Processing partition 3
16/03/23 16:52:09 INFO caffe.LmdbRDD: Completed partition 3
16/03/23 16:52:09 INFO storage.BlockManager: Found block rdd_0_3 locally
I0323 16:52:10.081959 1452 socket.cpp:58] receive message: 3 0
I0323 16:52:10.081997 1452 socket.cpp:59] socket: 12 message header size: 12
I0323 16:52:10.128891 1377 socket.cpp:44] send essage: 1 0 344864
I0323 16:52:10.140643 1377 socket.cpp:44] send essage: 1 0 344864
I0323 16:52:10.149338 1377 socket.cpp:44] send essage: 1 0 344864
I0323 16:52:10.159688 1425 socket.cpp:58] receive message: 0 0
I0323 16:52:10.159714 1425 socket.cpp:59] socket: 12 message header size: 12
I0323 16:52:10.162822 1377 socket.cpp:44] send essage: 1 0 344864
I0323 16:52:10.940675 1451 socket.cpp:58] receive message: 2 0
I0323 16:52:10.940747 1451 socket.cpp:59] socket: 12 message header size: 12
I0323 16:52:11.121742 1450 socket.cpp:58] receive message: 4 0
I0323 16:52:11.121800 1450 socket.cpp:59] socket: 12 message header size: 12
16/03/23 16:53:58 INFO executor.CoarseGrainedExecutorBackend: Driver commanded a shutdown
16/03/23 16:53:58 INFO storage.MemoryStore: MemoryStore cleared
16/03/23 16:53:58 INFO storage.BlockManager: BlockManager stopped
16/03/23 16:53:58 WARN executor.CoarseGrainedExecutorBackend: An unknown (dlgpu20.ai.bjcc.qihoo.net:45855) driver disconnected.
16/03/23 16:53:58 ERROR executor.CoarseGrainedExecutorBackend: Driver 10.142.118.172:45855 disassociated! Shutting down.
16/03/23 16:53:58 INFO util.ShutdownHookManager: Shutdown hook called
16/03/23 16:53:58 INFO remote.RemoteActorRefProvider$RemotingTerminator: Shutting down remote daemon.
LogType:stdout
Log Upload Time:23-Mar-2016 16:53:59
LogLength:0
Log Contents:
Container: container_1458635993280_0010_01_000002 on dlgpu20.ai.bjcc.qihoo.net_35190
LogType:stderr
Log Upload Time:23-Mar-2016 16:53:58
LogLength:30799
Log Contents:
SLF4J: Class path contains multiple SLF4J bindings.
SLF4J: Found binding in [jar:file:/run/hadoop/nm-local-dir/usercache/hadoop/filecache/12/spark-assembly-1.6.0-hadoop2.6.4-U4.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: Found binding in [jar:file:/home/hadoop/software/hadoop-2.6.4/share/hadoop/common/lib/slf4j-log4j12-1.7.5.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: See http://www.slf4j.org/codes.html#multiple_bindings for an explanation.
SLF4J: Actual binding is of type [org.slf4j.impl.Log4jLoggerFactory]
16/03/23 16:51:12 INFO executor.CoarseGrainedExecutorBackend: Registered signal handlers for [TERM, HUP, INT]
16/03/23 16:51:12 INFO yarn.YarnSparkHadoopUtil: Set ugi in YarnConf to: hadoop,hadoop
16/03/23 16:51:13 WARN util.NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
16/03/23 16:51:13 INFO yarn.YarnSparkHadoopUtil: running as user: hadoop
16/03/23 16:51:13 INFO spark.SecurityManager: Changing view acls to: hadoop
16/03/23 16:51:13 INFO spark.SecurityManager: Changing modify acls to: hadoop
16/03/23 16:51:13 INFO spark.SecurityManager: SecurityManager: authentication disabled; ui acls disabled; users with view permissions: Set(hadoop); users with modify permissions: Set(hadoop)
16/03/23 16:51:14 INFO spark.SecurityManager: Changing view acls to: hadoop
16/03/23 16:51:14 INFO spark.SecurityManager: Changing modify acls to: hadoop
16/03/23 16:51:14 INFO spark.SecurityManager: SecurityManager: authentication disabled; ui acls disabled; users with view permissions: Set(hadoop); users with modify permissions: Set(hadoop)
16/03/23 16:51:14 INFO slf4j.Slf4jLogger: Slf4jLogger started
16/03/23 16:51:14 INFO Remoting: Starting remoting
16/03/23 16:51:14 INFO Remoting: Remoting started; listening on addresses :[akka.tcp://[email protected]:45365]
16/03/23 16:51:14 INFO util.Utils: Successfully started service 'sparkExecutorActorSystem' on port 45365.
16/03/23 16:51:14 INFO storage.DiskBlockManager: Created local directory at /home/hadoop/yarn/nm-local-dir/usercache/hadoop/appcache/application_1458635993280_0010/blockmgr-3159017c-e654-4541-8292-47d1e6039345
16/03/23 16:51:14 INFO storage.DiskBlockManager: Created local directory at /dev/hadoop/nm-local-dir/usercache/hadoop/appcache/application_1458635993280_0010/blockmgr-4cf722e0-fd5e-42e3-af3a-05ff17d88dc1
16/03/23 16:51:14 INFO storage.DiskBlockManager: Created local directory at /run/hadoop/nm-local-dir/usercache/hadoop/appcache/application_1458635993280_0010/blockmgr-1ed10eb6-8a21-4b13-b983-fea3b40c7120
16/03/23 16:51:14 INFO storage.MemoryStore: MemoryStore started with capacity 24.9 GB
16/03/23 16:51:14 INFO yarn.YarnSparkHadoopUtil: Set ugi in YarnConf to: hadoop,hadoop
16/03/23 16:51:15 INFO executor.CoarseGrainedExecutorBackend: Connecting to driver: spark://[email protected]:45855
16/03/23 16:51:15 INFO executor.CoarseGrainedExecutorBackend: Successfully registered with driver
16/03/23 16:51:15 INFO executor.Executor: Starting executor ID 1 on host dlgpu20.ai.bjcc.qihoo.net
16/03/23 16:51:15 INFO util.Utils: Successfully started service 'org.apache.spark.network.netty.NettyBlockTransferService' on port 60205.
16/03/23 16:51:15 INFO netty.NettyBlockTransferService: Server created on 60205
16/03/23 16:51:15 INFO storage.BlockManager: external shuffle service port = 7337
16/03/23 16:51:15 INFO storage.BlockManagerMaster: Trying to register BlockManager
16/03/23 16:51:15 INFO storage.BlockManagerMaster: Registered BlockManager
16/03/23 16:51:15 INFO storage.BlockManager: Registering executor with local external shuffle service.
16/03/23 16:51:19 INFO executor.CoarseGrainedExecutorBackend: Got assigned task 3
16/03/23 16:51:19 INFO executor.Executor: Running task 3.0 in stage 0.0 (TID 3)
16/03/23 16:51:19 INFO broadcast.TorrentBroadcast: Started reading broadcast variable 0
16/03/23 16:51:20 INFO storage.MemoryStore: Block broadcast_0_piece0 stored as bytes in memory (estimated size 2.1 KB, free 2.1 KB)
16/03/23 16:51:20 INFO broadcast.TorrentBroadcast: Reading broadcast variable 0 took 310 ms
16/03/23 16:51:20 INFO storage.MemoryStore: Block broadcast_0 stored as values in memory (estimated size 3.2 KB, free 5.3 KB)
16/03/23 16:51:20 INFO caffe.CaffeProcessor: my rank is 3
16/03/23 16:51:20 INFO caffe.LMDB: Batch size:64
WARNING: Logging before InitGoogleLogging() is written to STDERR
I0323 16:51:26.091801 29967 CaffeNet.cpp:78] set root solver device id to 0
I0323 16:51:26.250738 29967 solver.cpp:48] Initializing solver from parameters:
test_iter: 1
test_interval: 10001
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: 10001
snapshot_prefix: "mnist_lenet"
solver_mode: GPU
device_id: 0
net_param {
name: "LeNet"
layer {
name: "data"
type: "MemoryData"
top: "data"
top: "label"
include {
phase: TRAIN
}
memory_data_param {
batch_size: 64
channels: 1
height: 28
width: 28
share_in_parallel: false
source: "hdfs:///projects/machine_learning/image_dataset/mnist_train_lmdb/"
}
source_class: "com.yahoo.ml.caffe.LMDB"
}
layer {
name: "data"
type: "MemoryData"
top: "data"
top: "label"
include {
phase: TEST
}
memory_data_param {
batch_size: 100
channels: 1
height: 28
width: 28
share_in_parallel: false
source: "hdfs:///projects/machine_learning/image_dataset/mnist_test_lmdb/"
}
source_class: "com.yahoo.ml.caffe.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"
}
}
test_initialization: false
I0323 16:51:26.250965 29967 solver.cpp:86] Creating training net specified in net_param.
I0323 16:51:26.251049 29967 net.cpp:322] The NetState phase (0) differed from the phase (1) specified by a rule in layer data
I0323 16:51:26.251065 29967 net.cpp:322] The NetState phase (0) differed from the phase (1) specified by a rule in layer accuracy
I0323 16:51:26.251157 29967 net.cpp:49] Initializing net from parameters:
name: "LeNet"
state {
phase: TRAIN
}
layer {
name: "data"
type: "MemoryData"
top: "data"
top: "label"
include {
phase: TRAIN
}
memory_data_param {
batch_size: 64
channels: 1
height: 28
width: 28
share_in_parallel: false
source: "hdfs:///projects/machine_learning/image_dataset/mnist_train_lmdb/"
}
source_class: "com.yahoo.ml.caffe.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"
}
I0323 16:51:26.251276 29967 layer_factory.hpp:77] Creating layer data
I0323 16:51:26.251309 29967 net.cpp:106] Creating Layer data
I0323 16:51:26.251322 29967 net.cpp:411] data -> data
I0323 16:51:26.251359 29967 net.cpp:411] data -> label
I0323 16:51:26.254645 29967 net.cpp:150] Setting up data
I0323 16:51:26.254678 29967 net.cpp:157] Top shape: 64 1 28 28 (50176)
I0323 16:51:26.254685 29967 net.cpp:157] Top shape: 64 (64)
I0323 16:51:26.254689 29967 net.cpp:165] Memory required for data: 200960
I0323 16:51:26.254700 29967 layer_factory.hpp:77] Creating layer conv1
I0323 16:51:26.254724 29967 net.cpp:106] Creating Layer conv1
I0323 16:51:26.254732 29967 net.cpp:454] conv1 <- data
I0323 16:51:26.254747 29967 net.cpp:411] conv1 -> conv1
I0323 16:51:26.257411 29967 net.cpp:150] Setting up conv1
I0323 16:51:26.257429 29967 net.cpp:157] Top shape: 64 20 24 24 (737280)
I0323 16:51:26.257434 29967 net.cpp:165] Memory required for data: 3150080
I0323 16:51:26.257454 29967 layer_factory.hpp:77] Creating layer pool1
I0323 16:51:26.257468 29967 net.cpp:106] Creating Layer pool1
I0323 16:51:26.257473 29967 net.cpp:454] pool1 <- conv1
I0323 16:51:26.257479 29967 net.cpp:411] pool1 -> pool1
I0323 16:51:26.257601 29967 net.cpp:150] Setting up pool1
I0323 16:51:26.257608 29967 net.cpp:157] Top shape: 64 20 12 12 (184320)
I0323 16:51:26.257612 29967 net.cpp:165] Memory required for data: 3887360
I0323 16:51:26.257616 29967 layer_factory.hpp:77] Creating layer conv2
I0323 16:51:26.257628 29967 net.cpp:106] Creating Layer conv2
I0323 16:51:26.257632 29967 net.cpp:454] conv2 <- pool1
I0323 16:51:26.257642 29967 net.cpp:411] conv2 -> conv2
I0323 16:51:26.258602 29967 net.cpp:150] Setting up conv2
I0323 16:51:26.258618 29967 net.cpp:157] Top shape: 64 50 8 8 (204800)
I0323 16:51:26.258622 29967 net.cpp:165] Memory required for data: 4706560
I0323 16:51:26.258632 29967 layer_factory.hpp:77] Creating layer pool2
I0323 16:51:26.258641 29967 net.cpp:106] Creating Layer pool2
I0323 16:51:26.258644 29967 net.cpp:454] pool2 <- conv2
I0323 16:51:26.258651 29967 net.cpp:411] pool2 -> pool2
I0323 16:51:26.258749 29967 net.cpp:150] Setting up pool2
I0323 16:51:26.258764 29967 net.cpp:157] Top shape: 64 50 4 4 (51200)
I0323 16:51:26.258769 29967 net.cpp:165] Memory required for data: 4911360
I0323 16:51:26.258772 29967 layer_factory.hpp:77] Creating layer ip1
I0323 16:51:26.258790 29967 net.cpp:106] Creating Layer ip1
I0323 16:51:26.258795 29967 net.cpp:454] ip1 <- pool2
I0323 16:51:26.258800 29967 net.cpp:411] ip1 -> ip1
I0323 16:51:26.264073 29967 net.cpp:150] Setting up ip1
I0323 16:51:26.264094 29967 net.cpp:157] Top shape: 64 500 (32000)
I0323 16:51:26.264099 29967 net.cpp:165] Memory required for data: 5039360
I0323 16:51:26.264112 29967 layer_factory.hpp:77] Creating layer relu1
I0323 16:51:26.264122 29967 net.cpp:106] Creating Layer relu1
I0323 16:51:26.264127 29967 net.cpp:454] relu1 <- ip1
I0323 16:51:26.264133 29967 net.cpp:397] relu1 -> ip1 (in-place)
I0323 16:51:26.264145 29967 net.cpp:150] Setting up relu1
I0323 16:51:26.264152 29967 net.cpp:157] Top shape: 64 500 (32000)
I0323 16:51:26.264154 29967 net.cpp:165] Memory required for data: 5167360
I0323 16:51:26.264158 29967 layer_factory.hpp:77] Creating layer ip2
I0323 16:51:26.264165 29967 net.cpp:106] Creating Layer ip2
I0323 16:51:26.264169 29967 net.cpp:454] ip2 <- ip1
I0323 16:51:26.264178 29967 net.cpp:411] ip2 -> ip2
I0323 16:51:26.265939 29967 net.cpp:150] Setting up ip2
I0323 16:51:26.265956 29967 net.cpp:157] Top shape: 64 10 (640)
I0323 16:51:26.265961 29967 net.cpp:165] Memory required for data: 5169920
I0323 16:51:26.265969 29967 layer_factory.hpp:77] Creating layer loss
I0323 16:51:26.265980 29967 net.cpp:106] Creating Layer loss
I0323 16:51:26.265985 29967 net.cpp:454] loss <- ip2
I0323 16:51:26.265990 29967 net.cpp:454] loss <- label
I0323 16:51:26.266002 29967 net.cpp:411] loss -> loss
I0323 16:51:26.266022 29967 layer_factory.hpp:77] Creating layer loss
I0323 16:51:26.266337 29967 net.cpp:150] Setting up loss
I0323 16:51:26.266350 29967 net.cpp:157] Top shape: (1)
I0323 16:51:26.266355 29967 net.cpp:160] with loss weight 1
I0323 16:51:26.266376 29967 net.cpp:165] Memory required for data: 5169924
I0323 16:51:26.266381 29967 net.cpp:226] loss needs backward computation.
I0323 16:51:26.266386 29967 net.cpp:226] ip2 needs backward computation.
I0323 16:51:26.266389 29967 net.cpp:226] relu1 needs backward computation.
I0323 16:51:26.266392 29967 net.cpp:226] ip1 needs backward computation.
I0323 16:51:26.266396 29967 net.cpp:226] pool2 needs backward computation.
I0323 16:51:26.266401 29967 net.cpp:226] conv2 needs backward computation.
I0323 16:51:26.266404 29967 net.cpp:226] pool1 needs backward computation.
I0323 16:51:26.266407 29967 net.cpp:226] conv1 needs backward computation.
I0323 16:51:26.266412 29967 net.cpp:228] data does not need backward computation.
I0323 16:51:26.266415 29967 net.cpp:270] This network produces output loss
I0323 16:51:26.266427 29967 net.cpp:283] Network initialization done.
I0323 16:51:26.266485 29967 solver.cpp:181] Creating test net (#0) specified by net_param
I0323 16:51:26.266506 29967 net.cpp:322] The NetState phase (1) differed from the phase (0) specified by a rule in layer data
I0323 16:51:26.266605 29967 net.cpp:49] Initializing net from parameters:
name: "LeNet"
state {
phase: TEST
}
layer {
name: "data"
type: "MemoryData"
top: "data"
top: "label"
include {
phase: TEST
}
memory_data_param {
batch_size: 100
channels: 1
height: 28
width: 28
share_in_parallel: false
source: "hdfs:///projects/machine_learning/image_dataset/mnist_test_lmdb/"
}
source_class: "com.yahoo.ml.caffe.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"
}
I0323 16:51:26.266691 29967 layer_factory.hpp:77] Creating layer data
I0323 16:51:26.266703 29967 net.cpp:106] Creating Layer data
I0323 16:51:26.266710 29967 net.cpp:411] data -> data
I0323 16:51:26.266718 29967 net.cpp:411] data -> label
I0323 16:51:26.268417 29967 net.cpp:150] Setting up data
I0323 16:51:26.268438 29967 net.cpp:157] Top shape: 100 1 28 28 (78400)
I0323 16:51:26.268445 29967 net.cpp:157] Top shape: 100 (100)
I0323 16:51:26.268448 29967 net.cpp:165] Memory required for data: 314000
I0323 16:51:26.268453 29967 layer_factory.hpp:77] Creating layer label_data_1_split
I0323 16:51:26.268463 29967 net.cpp:106] Creating Layer label_data_1_split
I0323 16:51:26.268468 29967 net.cpp:454] label_data_1_split <- label
I0323 16:51:26.268474 29967 net.cpp:411] label_data_1_split -> label_data_1_split_0
I0323 16:51:26.268482 29967 net.cpp:411] label_data_1_split -> label_data_1_split_1
I0323 16:51:26.268594 29967 net.cpp:150] Setting up label_data_1_split
I0323 16:51:26.268602 29967 net.cpp:157] Top shape: 100 (100)
I0323 16:51:26.268607 29967 net.cpp:157] Top shape: 100 (100)
I0323 16:51:26.268610 29967 net.cpp:165] Memory required for data: 314800
I0323 16:51:26.268615 29967 layer_factory.hpp:77] Creating layer conv1
I0323 16:51:26.268626 29967 net.cpp:106] Creating Layer conv1
I0323 16:51:26.268630 29967 net.cpp:454] conv1 <- data
I0323 16:51:26.268641 29967 net.cpp:411] conv1 -> conv1
I0323 16:51:26.269433 29967 net.cpp:150] Setting up conv1
I0323 16:51:26.269448 29967 net.cpp:157] Top shape: 100 20 24 24 (1152000)
I0323 16:51:26.269451 29967 net.cpp:165] Memory required for data: 4922800
I0323 16:51:26.269461 29967 layer_factory.hpp:77] Creating layer pool1
I0323 16:51:26.269476 29967 net.cpp:106] Creating Layer pool1
I0323 16:51:26.269481 29967 net.cpp:454] pool1 <- conv1
I0323 16:51:26.269489 29967 net.cpp:411] pool1 -> pool1
I0323 16:51:26.269598 29967 net.cpp:150] Setting up pool1
I0323 16:51:26.269605 29967 net.cpp:157] Top shape: 100 20 12 12 (288000)
I0323 16:51:26.269609 29967 net.cpp:165] Memory required for data: 6074800
I0323 16:51:26.269613 29967 layer_factory.hpp:77] Creating layer conv2
I0323 16:51:26.269626 29967 net.cpp:106] Creating Layer conv2
I0323 16:51:26.269631 29967 net.cpp:454] conv2 <- pool1
I0323 16:51:26.269639 29967 net.cpp:411] conv2 -> conv2
I0323 16:51:26.270656 29967 net.cpp:150] Setting up conv2
I0323 16:51:26.270670 29967 net.cpp:157] Top shape: 100 50 8 8 (320000)
I0323 16:51:26.270674 29967 net.cpp:165] Memory required for data: 7354800
I0323 16:51:26.270684 29967 layer_factory.hpp:77] Creating layer pool2
I0323 16:51:26.270690 29967 net.cpp:106] Creating Layer pool2
I0323 16:51:26.270695 29967 net.cpp:454] pool2 <- conv2
I0323 16:51:26.270702 29967 net.cpp:411] pool2 -> pool2
I0323 16:51:26.270805 29967 net.cpp:150] Setting up pool2
I0323 16:51:26.270812 29967 net.cpp:157] Top shape: 100 50 4 4 (80000)
I0323 16:51:26.270823 29967 net.cpp:165] Memory required for data: 7674800
I0323 16:51:26.270828 29967 layer_factory.hpp:77] Creating layer ip1
I0323 16:51:26.270838 29967 net.cpp:106] Creating Layer ip1
I0323 16:51:26.270841 29967 net.cpp:454] ip1 <- pool2
I0323 16:51:26.270849 29967 net.cpp:411] ip1 -> ip1
I0323 16:51:26.276105 29967 net.cpp:150] Setting up ip1
I0323 16:51:26.276124 29967 net.cpp:157] Top shape: 100 500 (50000)
I0323 16:51:26.276127 29967 net.cpp:165] Memory required for data: 7874800
I0323 16:51:26.276139 29967 layer_factory.hpp:77] Creating layer relu1
I0323 16:51:26.276149 29967 net.cpp:106] Creating Layer relu1
I0323 16:51:26.276154 29967 net.cpp:454] relu1 <- ip1
I0323 16:51:26.276159 29967 net.cpp:397] relu1 -> ip1 (in-place)
I0323 16:51:26.276166 29967 net.cpp:150] Setting up relu1
I0323 16:51:26.276171 29967 net.cpp:157] Top shape: 100 500 (50000)
I0323 16:51:26.276175 29967 net.cpp:165] Memory required for data: 8074800
I0323 16:51:26.276178 29967 layer_factory.hpp:77] Creating layer ip2
I0323 16:51:26.276188 29967 net.cpp:106] Creating Layer ip2
I0323 16:51:26.276192 29967 net.cpp:454] ip2 <- ip1
I0323 16:51:26.276198 29967 net.cpp:411] ip2 -> ip2
I0323 16:51:26.276582 29967 net.cpp:150] Setting up ip2
I0323 16:51:26.276592 29967 net.cpp:157] Top shape: 100 10 (1000)
I0323 16:51:26.276595 29967 net.cpp:165] Memory required for data: 8078800
I0323 16:51:26.276602 29967 layer_factory.hpp:77] Creating layer ip2_ip2_0_split
I0323 16:51:26.276608 29967 net.cpp:106] Creating Layer ip2_ip2_0_split
I0323 16:51:26.276612 29967 net.cpp:454] ip2_ip2_0_split <- ip2
I0323 16:51:26.276621 29967 net.cpp:411] ip2_ip2_0_split -> ip2_ip2_0_split_0
I0323 16:51:26.276628 29967 net.cpp:411] ip2_ip2_0_split -> ip2_ip2_0_split_1
I0323 16:51:26.276732 29967 net.cpp:150] Setting up ip2_ip2_0_split
I0323 16:51:26.276739 29967 net.cpp:157] Top shape: 100 10 (1000)
I0323 16:51:26.276744 29967 net.cpp:157] Top shape: 100 10 (1000)
I0323 16:51:26.276748 29967 net.cpp:165] Memory required for data: 8086800
I0323 16:51:26.276751 29967 layer_factory.hpp:77] Creating layer accuracy
I0323 16:51:26.276762 29967 net.cpp:106] Creating Layer accuracy
I0323 16:51:26.276765 29967 net.cpp:454] accuracy <- ip2_ip2_0_split_0
I0323 16:51:26.276770 29967 net.cpp:454] accuracy <- label_data_1_split_0
I0323 16:51:26.276778 29967 net.cpp:411] accuracy -> accuracy
I0323 16:51:26.276792 29967 net.cpp:150] Setting up accuracy
I0323 16:51:26.276796 29967 net.cpp:157] Top shape: (1)
I0323 16:51:26.276800 29967 net.cpp:165] Memory required for data: 8086804
I0323 16:51:26.276803 29967 layer_factory.hpp:77] Creating layer loss
I0323 16:51:26.276809 29967 net.cpp:106] Creating Layer loss
I0323 16:51:26.276813 29967 net.cpp:454] loss <- ip2_ip2_0_split_1
I0323 16:51:26.276818 29967 net.cpp:454] loss <- label_data_1_split_1
I0323 16:51:26.276830 29967 net.cpp:411] loss -> loss
I0323 16:51:26.276839 29967 layer_factory.hpp:77] Creating layer loss
I0323 16:51:26.277124 29967 net.cpp:150] Setting up loss
I0323 16:51:26.277137 29967 net.cpp:157] Top shape: (1)
I0323 16:51:26.277140 29967 net.cpp:160] with loss weight 1
I0323 16:51:26.277148 29967 net.cpp:165] Memory required for data: 8086808
I0323 16:51:26.277153 29967 net.cpp:226] loss needs backward computation.
I0323 16:51:26.277156 29967 net.cpp:228] accuracy does not need backward computation.
I0323 16:51:26.277161 29967 net.cpp:226] ip2_ip2_0_split needs backward computation.
I0323 16:51:26.277165 29967 net.cpp:226] ip2 needs backward computation.
I0323 16:51:26.277169 29967 net.cpp:226] relu1 needs backward computation.
I0323 16:51:26.277173 29967 net.cpp:226] ip1 needs backward computation.
I0323 16:51:26.277176 29967 net.cpp:226] pool2 needs backward computation.
I0323 16:51:26.277181 29967 net.cpp:226] conv2 needs backward computation.
I0323 16:51:26.277187 29967 net.cpp:226] pool1 needs backward computation.
I0323 16:51:26.277191 29967 net.cpp:226] conv1 needs backward computation.
I0323 16:51:26.277195 29967 net.cpp:228] label_data_1_split does not need backward computation.
I0323 16:51:26.277205 29967 net.cpp:228] data does not need backward computation.
I0323 16:51:26.277209 29967 net.cpp:270] This network produces output accuracy
I0323 16:51:26.277212 29967 net.cpp:270] This network produces output loss
I0323 16:51:26.277226 29967 net.cpp:283] Network initialization done.
I0323 16:51:26.277271 29967 solver.cpp:60] Solver scaffolding done.
I0323 16:51:26.278825 29967 socket.cpp:219] Waiting for valid port [0]
I0323 16:51:26.278879 29991 socket.cpp:158] Assigned socket server port [47155]
I0323 16:51:26.281419 29991 socket.cpp:171] Socket Server ready [0.0.0.0]
I0323 16:51:26.288908 29967 socket.cpp:219] Waiting for valid port [47155]
I0323 16:51:26.288918 29967 socket.cpp:227] Valid port found [47155]
I0323 16:51:26.288931 29967 CaffeNet.cpp:186] Socket adapter: dlgpu20.ai.bjcc.qihoo.net:47155
I0323 16:51:26.289212 29967 CaffeNet.cpp:325] 0-th Socket addr: dlgpu20.ai.bjcc.qihoo.net:47155
I0323 16:51:26.289224 29967 CaffeNet.cpp:325] 1-th Socket addr: dlgpu20.ai.bjcc.qihoo.net:47155
I0323 16:51:26.289228 29967 CaffeNet.cpp:325] 2-th Socket addr: dlgpu20.ai.bjcc.qihoo.net:47155
I0323 16:51:26.289233 29967 CaffeNet.cpp:325] 3-th Socket addr:
I0323 16:51:26.289237 29967 CaffeNet.cpp:325] 4-th Socket addr: dlgpu20.ai.bjcc.qihoo.net:47155
I0323 16:51:26.289249 29967 JniCaffeNet.cpp:110] 0-th local addr: dlgpu20.ai.bjcc.qihoo.net:47155
I0323 16:51:26.289255 29967 JniCaffeNet.cpp:110] 1-th local addr: dlgpu20.ai.bjcc.qihoo.net:47155
I0323 16:51:26.289259 29967 JniCaffeNet.cpp:110] 2-th local addr: dlgpu20.ai.bjcc.qihoo.net:47155
I0323 16:51:26.289263 29967 JniCaffeNet.cpp:110] 3-th local addr:
I0323 16:51:26.289266 29967 JniCaffeNet.cpp:110] 4-th local addr: dlgpu20.ai.bjcc.qihoo.net:47155
16/03/23 16:51:26 INFO executor.Executor: Finished task 3.0 in stage 0.0 (TID 3). 1069 bytes result sent to driver
16/03/23 16:51:27 INFO executor.CoarseGrainedExecutorBackend: Got assigned task 8
16/03/23 16:51:27 INFO executor.Executor: Running task 3.0 in stage 1.0 (TID 8)
16/03/23 16:51:27 INFO broadcast.TorrentBroadcast: Started reading broadcast variable 2
16/03/23 16:51:27 INFO storage.MemoryStore: Block broadcast_2_piece0 stored as bytes in memory (estimated size 1622.0 B, free 6.9 KB)
16/03/23 16:51:27 INFO broadcast.TorrentBroadcast: Reading broadcast variable 2 took 16 ms
16/03/23 16:51:27 INFO storage.MemoryStore: Block broadcast_2 stored as values in memory (estimated size 2.6 KB, free 9.5 KB)
16/03/23 16:51:27 INFO broadcast.TorrentBroadcast: Started reading broadcast variable 1
16/03/23 16:51:27 INFO storage.MemoryStore: Block broadcast_1_piece0 stored as bytes in memory (estimated size 235.0 B, free 9.7 KB)
16/03/23 16:51:27 INFO broadcast.TorrentBroadcast: Reading broadcast variable 1 took 12 ms
16/03/23 16:51:27 INFO storage.MemoryStore: Block broadcast_1 stored as values in memory (estimated size 3.1 KB, free 12.8 KB)
I0323 16:51:27.556499 29967 common.cpp:61] 3-th string is NULL
I0323 16:51:27.556599 29967 socket.cpp:250] Trying to connect with ...[dlgpu19.ai.bjcc.qihoo.net:57402]
I0323 16:51:27.556901 29967 socket.cpp:309] Connected to server [dlgpu19.ai.bjcc.qihoo.net:57402] with client_fd [315]
I0323 16:51:37.557061 29967 socket.cpp:250] Trying to connect with ...[dlgpu19.ai.bjcc.qihoo.net:41630]
I0323 16:51:37.557479 29967 socket.cpp:309] Connected to server [dlgpu19.ai.bjcc.qihoo.net:41630] with client_fd [316]
I0323 16:51:47.555366 29991 socket.cpp:184] Accepted the connection from client [dlgpu10.ai.bjcc.qihoo.net]
I0323 16:51:47.557629 29967 socket.cpp:250] Trying to connect with ...[dlgpu10.ai.bjcc.qihoo.net:58067]
I0323 16:51:47.557984 29967 socket.cpp:309] Connected to server [dlgpu10.ai.bjcc.qihoo.net:58067] with client_fd [318]
I0323 16:51:47.559761 29991 socket.cpp:184] Accepted the connection from client [dlgpu19.ai.bjcc.qihoo.net]
I0323 16:51:47.561936 29991 socket.cpp:184] Accepted the connection from client [dlgpu19.ai.bjcc.qihoo.net]
I0323 16:51:57.554638 29991 socket.cpp:184] Accepted the connection from client [dlgpu10.ai.bjcc.qihoo.net]
I0323 16:51:57.558133 29967 socket.cpp:250] Trying to connect with ...[dlgpu10.ai.bjcc.qihoo.net:46110]
I0323 16:51:57.558480 29967 socket.cpp:309] Connected to server [dlgpu10.ai.bjcc.qihoo.net:46110] with client_fd [322]
I0323 16:52:07.586596 29967 parallel.cpp:392] GPUs pairs 0:1, 2:3, 0:2, 0:4
I0323 16:52:07.660895 29967 parallel.cpp:234] GPU 4 does not have p2p access to GPU 0
I0323 16:52:07.694207 30111 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
16/03/23 16:52:07 INFO executor.Executor: Finished task 3.0 in stage 1.0 (TID 8). 918 bytes result sent to driver
I0323 16:52:08.215100 30113 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
I0323 16:52:08.250818 30115 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
I0323 16:52:08.287663 30117 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
I0323 16:52:08.299569 30119 MemoryInputAdapter.cpp:15] MemoryInputAdapter is used
16/03/23 16:52:08 INFO executor.CoarseGrainedExecutorBackend: Got assigned task 11
16/03/23 16:52:08 INFO executor.Executor: Running task 1.0 in stage 2.0 (TID 11)
16/03/23 16:52:08 INFO broadcast.TorrentBroadcast: Started reading broadcast variable 3
16/03/23 16:52:08 INFO storage.MemoryStore: Block broadcast_3_piece0 stored as bytes in memory (estimated size 1450.0 B, free 4.7 KB)
16/03/23 16:52:08 INFO broadcast.TorrentBroadcast: Reading broadcast variable 3 took 11 ms
16/03/23 16:52:08 INFO storage.MemoryStore: Block broadcast_3 stored as values in memory (estimated size 2.2 KB, free 6.9 KB)
16/03/23 16:52:08 INFO spark.CacheManager: Partition rdd_5_1 not found, computing it
16/03/23 16:52:08 INFO spark.CacheManager: Partition rdd_0_1 not found, computing it
16/03/23 16:52:08 INFO caffe.LmdbRDD: Processing partition 1
16/03/23 16:52:09 INFO caffe.LmdbRDD: Completed partition 1
16/03/23 16:52:09 INFO storage.BlockManager: Found block rdd_0_1 locally
16/03/23 16:53:58 ERROR executor.CoarseGrainedExecutorBackend: RECEIVED SIGNAL 15: SIGTERM
16/03/23 16:53:58 INFO storage.DiskBlockManager: Shutdown hook called
16/03/23 16:53:58 INFO util.ShutdownHookManager: Shutdown hook called
LogType:stdout
Log Upload Time:23-Mar-2016 16:53:58
LogLength:0
Log Contents:
Container: container_1458635993280_0010_01_000001 on dlgpu20.ai.bjcc.qihoo.net_35190
LogType:stderr
Log Upload Time:23-Mar-2016 16:53:58
LogLength:1577
Log Contents:
SLF4J: Class path contains multiple SLF4J bindings.
SLF4J: Found binding in [jar:file:/run/hadoop/nm-local-dir/usercache/hadoop/filecache/12/spark-assembly-1.6.0-hadoop2.6.4-U4.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: Found binding in [jar:file:/home/hadoop/software/hadoop-2.6.4/share/hadoop/common/lib/slf4j-log4j12-1.7.5.jar!/org/slf4j/impl/StaticLoggerBinder.class]
SLF4J: See http://www.slf4j.org/codes.html#multiple_bindings for an explanation.
SLF4J: Actual binding is of type [org.slf4j.impl.Log4jLoggerFactory]
[Stage 0:> (0 + 0) / 5]
[Stage 0:> (0 + 5) / 5]
[Stage 0:===========> (1 + 4) / 5]
[Stage 0:=======================> (2 + 3) / 5]
[Stage 0:===================================> (3 + 2) / 5]
[Stage 0:===========================================================(5 + 0) / 5]
[Stage 1:> (0 + 5) / 5]
[Stage 1:===================================> (3 + 2) / 5]
[Stage 2:> (0 + 5) / 5]
[Stage 2:> (0 + 5) / 5]16/03/23 16:53:58 ERROR ApplicationMaster: RECEIVED SIGNAL 15: SIGTERM
LogType:stdout
Log Upload Time:23-Mar-2016 16:53:58
LogLength:0
Log Contents:
from caffeonspark.
Your CLI asks for 5 executors, each need 5 GPUs. Each your server has 8 GPUs. 2 executors were allocated to 1 server, and these 2 executors will need 10 GPUs. As a result, we see the error messages such as:
I0323 16:52:07.676470 26145 parallel.cpp:234] GPU 4 does not have p2p access to GPU 0
That will disable the peer-to-peer communication among GPUs.
Please adjust your settings.
from caffeonspark.
@anfeng Thanks, this problem has been solved. I change the settings as your desciption.
from caffeonspark.
Related Issues (20)
- MNIST example at one out of four worker nodes only HOT 2
- CaffeOnSpark CPU model failed but GPU model success(same scripts and same data) HOT 1
- Trouble in performing test with existing model, dataframe empty. HOT 2
- Could anyone help about build CaffeOnSpak while caffe-distri failed ? HOT 5
- Core dump failures HOT 3
- CaffeOnSpark use infiniband but Cannot find the address of another infiniband host. HOT 4
- Use infiniband HOT 2
- Infiniband not work, Help me
- Feature extraction mode running slow HOT 7
- DataLayer use data_param instead of memory_data_param HOT 2
- Parameter synchronization mode HOT 2
- Error: Exception in thread "AWT-EventQueue-0" java.lang.UnsatisfiedLinkError: /Applications/Alice 2.4.app/Contents/Required/lib/osx/libjogl_awt.jnilib: Library not loaded: /System/Library/Frameworks/JavaVM.framework/Libraries/libjawt.dylib
- err “java.lang.UnsupportedOperationException: empty.reduceLeft”
- Error running javah command: Error executing command line HOT 2
- hive java.io.filenotfoundexception system cannot find specified path
- Is CaffeOnSpark still being maintained and developed? HOT 2
- Attribute protoFile not valid
- Exception in thread "main" org.apache.spark.SparkException: Cannot load main class from JAR file:/data/lenet_memory_solver.prototxt,/data/lenet_memory_train_test.prototxt
- have anyone faced that cannot fine class google V3 when doing test HOT 2
- java.lang.RuntimeException: Error while encoding: java.lang.ArrayIndexOutOfBoundsException: 1
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from caffeonspark.