RewriterConfig.MemOptType

enumeración final estática pública RewriterConfig.MemOptType

Protobuf enumeración tensorflow.RewriterConfig.MemOptType

Constantes

En t DEFAULT_MEM_OPT_VALUE
 The default setting (SCHEDULING and SWAPPING HEURISTICS only)
DEFAULT_MEM_OPT = 0;
En t HEURÍSTICA_VALUE
 Use any combination of swapping and recomputation heuristics.
En t VALOR_MANUAL
 Driven by manual op-level annotations.
En t NO_MEM_OPT_VALUE
 Disabled in the meta-optimizer.
En t RECOMPUTATION_HEURISTICS_VALUE
 Recomputation heuristics will recompute ops (such as Relu activation)
 during backprop instead of storing them, reducing peak memory usage.
En t SCHEDULING_HEURISTICS_VALUE
 Scheduling will split big ops such as AddN and try to enforce a schedule
 of the new computations that decreases peak memory usage.
En t INTERCAMBIO_HEURISTICS_VALUE
 Swapping heuristic will move a tensor from the GPU to the CPU and move
 it back when needed to reduce peak memory usage.

Métodos heredados

Valores de enumeración

público estático final RewriterConfig.MemOptType DEFAULT_MEM_OPT

 The default setting (SCHEDULING and SWAPPING HEURISTICS only)
DEFAULT_MEM_OPT = 0;

HEURÍSTICA pública estática final RewriterConfig.MemOptType

 Use any combination of swapping and recomputation heuristics.
HEURISTICS = 3;

público estático final RewriterConfig.MemOptType MANUAL

 Driven by manual op-level annotations.
MANUAL = 2;

público estático final RewriterConfig.MemOptType NO_MEM_OPT

 Disabled in the meta-optimizer.
NO_MEM_OPT = 1;

público estático final RewriterConfig.MemOptType RECOMPUTATION_HEURISTICS

 Recomputation heuristics will recompute ops (such as Relu activation)
 during backprop instead of storing them, reducing peak memory usage.
RECOMPUTATION_HEURISTICS = 5;

público estático final RewriterConfig.MemOptType SCHEDULING_HEURISTICS

 Scheduling will split big ops such as AddN and try to enforce a schedule
 of the new computations that decreases peak memory usage.
SCHEDULING_HEURISTICS = 6;

público estático final RewriterConfig.MemOptType SWAPPING_HEURISTICS

 Swapping heuristic will move a tensor from the GPU to the CPU and move
 it back when needed to reduce peak memory usage.
SWAPPING_HEURISTICS = 4;

público estático final RewriterConfig.MemOptType NO RECONOCIDO

Constantes

int final estático público DEFAULT_MEM_OPT_VALUE

 The default setting (SCHEDULING and SWAPPING HEURISTICS only)
DEFAULT_MEM_OPT = 0;

Valor constante: 0

público estático final int HEURISTICS_VALUE

 Use any combination of swapping and recomputation heuristics.
HEURISTICS = 3;

Valor constante: 3

público estático final int MANUAL_VALUE

 Driven by manual op-level annotations.
MANUAL = 2;

Valor constante: 2

int final estático público NO_MEM_OPT_VALUE

 Disabled in the meta-optimizer.
NO_MEM_OPT = 1;

Valor constante: 1

int final estático público RECOMPUTATION_HEURISTICS_VALUE

 Recomputation heuristics will recompute ops (such as Relu activation)
 during backprop instead of storing them, reducing peak memory usage.
RECOMPUTATION_HEURISTICS = 5;

Valor constante: 5

int final estático público SCHEDULING_HEURISTICS_VALUE

 Scheduling will split big ops such as AddN and try to enforce a schedule
 of the new computations that decreases peak memory usage.
SCHEDULING_HEURISTICS = 6;

Valor constante: 6

int final estático público SWAPPING_HEURISTICS_VALUE

 Swapping heuristic will move a tensor from the GPU to the CPU and move
 it back when needed to reduce peak memory usage.
SWAPPING_HEURISTICS = 4;

Valor constante: 4