interfaz pública KernelDefOrBuilder
Subclases indirectas conocidas |
Métodos públicos
resumen KernelDef.AttrConstraint | getConstraint (índice int) repeated .tensorflow.KernelDef.AttrConstraint constraint = 3; |
int abstracto | getConstraintCount () repeated .tensorflow.KernelDef.AttrConstraint constraint = 3; |
Lista abstracta < KernelDef.AttrConstraint > | getConstraintList () repeated .tensorflow.KernelDef.AttrConstraint constraint = 3; |
resumen KernelDef.AttrConstraintOrBuilder | getConstraintOrBuilder (índice int) repeated .tensorflow.KernelDef.AttrConstraint constraint = 3; |
Resumen Lista <? extiende KernelDef.AttrConstraintOrBuilder > | getConstraintOrBuilderList () repeated .tensorflow.KernelDef.AttrConstraint constraint = 3; |
cadena abstracta | getDeviceType () Type of device this kernel runs on. |
resumen com.google.protobuf.ByteString | getDeviceTypeBytes () Type of device this kernel runs on. |
cadena abstracta | getHostMemoryArg (índice int) Names of the Op's input_/output_args that reside in host memory instead of device memory. |
resumen com.google.protobuf.ByteString | getHostMemoryArgBytes (índice int) Names of the Op's input_/output_args that reside in host memory instead of device memory. |
int abstracto | getHostMemoryArgCount () Names of the Op's input_/output_args that reside in host memory instead of device memory. |
Lista abstracta <String> | getHostMemoryArgList () Names of the Op's input_/output_args that reside in host memory instead of device memory. |
cadena abstracta | getLabel () This allows experimental kernels to be registered for an op that won't be used unless the user specifies a "_kernel" attr with value matching this. |
resumen com.google.protobuf.ByteString | getLabelBytes () This allows experimental kernels to be registered for an op that won't be used unless the user specifies a "_kernel" attr with value matching this. |
cadena abstracta | getOp () Must match the name of an Op. |
resumen com.google.protobuf.ByteString | getOpBytes () Must match the name of an Op. |
int abstracto | getPriority () Prioritization of kernel amongst different devices. |
Métodos públicos
resumen público KernelDef.AttrConstraint getConstraint (índice int)
repeated .tensorflow.KernelDef.AttrConstraint constraint = 3;
public abstract int getConstraintCount ()
repeated .tensorflow.KernelDef.AttrConstraint constraint = 3;
Lista pública abstracta < KernelDef.AttrConstraint > getConstraintList ()
repeated .tensorflow.KernelDef.AttrConstraint constraint = 3;
resumen público KernelDef.AttrConstraintOrBuilder getConstraintOrBuilder (índice int)
repeated .tensorflow.KernelDef.AttrConstraint constraint = 3;
Lista de resumen público <? extiende KernelDef.AttrConstraintOrBuilder > getConstraintOrBuilderList ()
repeated .tensorflow.KernelDef.AttrConstraint constraint = 3;
getDeviceType de cadena abstracta pública ()
Type of device this kernel runs on.
string device_type = 2;
resumen público com.google.protobuf.ByteString getDeviceTypeBytes ()
Type of device this kernel runs on.
string device_type = 2;
Cadena abstracta pública getHostMemoryArg (índice int)
Names of the Op's input_/output_args that reside in host memory instead of device memory.
repeated string host_memory_arg = 4;
resumen público com.google.protobuf.ByteString getHostMemoryArgBytes (índice int)
Names of the Op's input_/output_args that reside in host memory instead of device memory.
repeated string host_memory_arg = 4;
public abstract int getHostMemoryArgCount ()
Names of the Op's input_/output_args that reside in host memory instead of device memory.
repeated string host_memory_arg = 4;
Lista pública abstracta <String> getHostMemoryArgList ()
Names of the Op's input_/output_args that reside in host memory instead of device memory.
repeated string host_memory_arg = 4;
getLabel () de cadena abstracta pública
This allows experimental kernels to be registered for an op that won't be used unless the user specifies a "_kernel" attr with value matching this.
string label = 5;
resumen público com.google.protobuf.ByteString getLabelBytes ()
This allows experimental kernels to be registered for an op that won't be used unless the user specifies a "_kernel" attr with value matching this.
string label = 5;
getOp () de cadena abstracta pública
Must match the name of an Op.
string op = 1;
resumen público com.google.protobuf.ByteString getOpBytes ()
Must match the name of an Op.
string op = 1;
public abstract int getPriority ()
Prioritization of kernel amongst different devices. By default we assume priority is 0. The higher the priority the better. By default (i.e. if this is not set), we prefer GPU kernels over CPU.
int32 priority = 6;