在tensorflow中实现屏蔽输出的log信息
tensorflow中可以通过配置环境变量'TF_CPP_MIN_LOG_LEVEL'的值,控制tensorflow是否屏蔽通知信息、警告、报错等输出信息。
使用方法:
importos importtensorflowastf os.environ['TF_CPP_MIN_LOG_LEVEL']='3'#orany{'0','1','2'}
TF_CPP_MIN_LOG_LEVEL取值0:0也是默认值,输出所有信息
TF_CPP_MIN_LOG_LEVEL取值1:屏蔽通知信息
TF_CPP_MIN_LOG_LEVEL取值2:屏蔽通知信息和警告信息
TF_CPP_MIN_LOG_LEVEL取值3:屏蔽通知信息、警告信息和报错信息
测试代码:
importtensorflowastf importos os.environ['TF_CPP_MIN_LOG_LEVEL']='0' #os.environ['TF_CPP_MIN_LOG_LEVEL']='1' #os.environ['TF_CPP_MIN_LOG_LEVEL']='2' #os.environ['TF_CPP_MIN_LOG_LEVEL']='3' v1=tf.constant([1.0,2.0,3.0],shape=[3],name='v1') v2=tf.constant([1.0,2.0,3.0],shape=[3],name='v2') sumV12=v1+v2 withtf.Session(config=tf.ConfigProto(log_device_placement=True))assess: printsess.run(sumV12)
TF_CPP_MIN_LOG_LEVEL为0的输出:
2018-04-2114:59:09.910415:Wtensorflow/core/platform/cpu_feature_guard.cc:45]TheTensorFlowlibrarywasn'tcompiledtouseSSE4.1instructions,buttheseareavailableonyourmachineandcouldspeedupCPUcomputations. 2018-04-2114:59:09.910442:Wtensorflow/core/platform/cpu_feature_guard.cc:45]TheTensorFlowlibrarywasn'tcompiledtouseSSE4.2instructions,buttheseareavailableonyourmachineandcouldspeedupCPUcomputations. 2018-04-2114:59:09.910448:Wtensorflow/core/platform/cpu_feature_guard.cc:45]TheTensorFlowlibrarywasn'tcompiledtouseAVXinstructions,buttheseareavailableonyourmachineandcouldspeedupCPUcomputations. 2018-04-2114:59:09.910453:Wtensorflow/core/platform/cpu_feature_guard.cc:45]TheTensorFlowlibrarywasn'tcompiledtouseAVX2instructions,buttheseareavailableonyourmachineandcouldspeedupCPUcomputations. 2018-04-2114:59:09.910457:Wtensorflow/core/platform/cpu_feature_guard.cc:45]TheTensorFlowlibrarywasn'tcompiledtouseFMAinstructions,buttheseareavailableonyourmachineandcouldspeedupCPUcomputations. 2018-04-2114:59:09.911260:Itensorflow/core/common_runtime/direct_session.cc:300]Devicemapping: 2018-04-2114:59:09.911816:Itensorflow/core/common_runtime/simple_placer.cc:872]add:(Add)/job:localhost/replica:0/task:0/cpu:0 2018-04-2114:59:09.911835:Itensorflow/core/common_runtime/simple_placer.cc:872]v2:(Const)/job:localhost/replica:0/task:0/cpu:0 2018-04-2114:59:09.911841:Itensorflow/core/common_runtime/simple_placer.cc:872]v1:(Const)/job:localhost/replica:0/task:0/cpu:0 Devicemapping:noknowndevices. add:(Add):/job:localhost/replica:0/task:0/cpu:0 v2:(Const):/job:localhost/replica:0/task:0/cpu:0 v1:(Const):/job:localhost/replica:0/task:0/cpu:0 [2.4.6.]
值为0也是默认的输出,分为三部分,一个是警告信息说没有优化加速,二是通知信息告知操作所用的设备,三是程序中代码指定要输出的结果信息
TF_CPP_MIN_LOG_LEVEL为1的输出,没有通知信息了: 2018-04-2114:59:09.910415:Wtensorflow/core/platform/cpu_feature_guard.cc:45]TheTensorFlowlibrarywasn'tcompiledtouseSSE4.1instructions,buttheseareavailableonyourmachineandcouldspeedupCPUcomputations. 2018-04-2114:59:09.910442:Wtensorflow/core/platform/cpu_feature_guard.cc:45]TheTensorFlowlibrarywasn'tcompiledtouseSSE4.2instructions,buttheseareavailableonyourmachineandcouldspeedupCPUcomputations. 2018-04-2114:59:09.910448:Wtensorflow/core/platform/cpu_feature_guard.cc:45]TheTensorFlowlibrarywasn'tcompiledtouseAVXinstructions,buttheseareavailableonyourmachineandcouldspeedupCPUcomputations. 2018-04-2114:59:09.910453:Wtensorflow/core/platform/cpu_feature_guard.cc:45]TheTensorFlowlibrarywasn'tcompiledtouseAVX2instructions,buttheseareavailableonyourmachineandcouldspeedupCPUcomputations. 2018-04-2114:59:09.910457:Wtensorflow/core/platform/cpu_feature_guard.cc:45]TheTensorFlowlibrarywasn'tcompiledtouseFMAinstructions,buttheseareavailableonyourmachineandcouldspeedupCPUcomputations. Devicemapping:noknowndevices. add:(Add):/job:localhost/replica:0/task:0/cpu:0 v2:(Const):/job:localhost/replica:0/task:0/cpu:0 v1:(Const):/job:localhost/replica:0/task:0/cpu:0 [2.4.6.]
TF_CPP_MIN_LOG_LEVEL为2和3的输出,设置为2就没有警告信息了,设置为3警告和报错信息(如果有)就都没有了:
Devicemapping:noknowndevices. add:(Add):/job:localhost/replica:0/task:0/cpu:0 v2:(Const):/job:localhost/replica:0/task:0/cpu:0 v1:(Const):/job:localhost/replica:0/task:0/cpu:0 [2.4.6.]
以上这篇在tensorflow中实现屏蔽输出的log信息就是小编分享给大家的全部内容了,希望能给大家一个参考,也希望大家多多支持毛票票。
声明:本文内容来源于网络,版权归原作者所有,内容由互联网用户自发贡献自行上传,本网站不拥有所有权,未作人工编辑处理,也不承担相关法律责任。如果您发现有涉嫌版权的内容,欢迎发送邮件至:czq8825#qq.com(发邮件时,请将#更换为@)进行举报,并提供相关证据,一经查实,本站将立刻删除涉嫌侵权内容。