공개 정적 최종 열거형 RewriterConfig.MemOptType
Protobuf 열거형 tensorflow.RewriterConfig.MemOptType
상수
정수 | DEFAULT_MEM_OPT_VALUE | The default setting (SCHEDULING and SWAPPING HEURISTICS only) DEFAULT_MEM_OPT = 0; |
정수 | HEURISTICS_VALUE | Use any combination of swapping and recomputation heuristics. |
정수 | MANUAL_VALUE | Driven by manual op-level annotations. |
정수 | NO_MEM_OPT_VALUE | Disabled in the meta-optimizer. |
정수 | RECOMPUTATION_HEURISTICS_VALUE | Recomputation heuristics will recompute ops (such as Relu activation) during backprop instead of storing them, reducing peak memory usage. |
정수 | 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. |
정수 | 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. |
상속된 메서드
열거형 값
공개 정적 최종 RewriterConfig.MemOptType DEFAULT_MEM_OPT
The default setting (SCHEDULING and SWAPPING HEURISTICS only)
DEFAULT_MEM_OPT = 0;
공개 정적 최종 RewriterConfig.MemOptType 휴리스틱스
Use any combination of swapping and recomputation heuristics.
HEURISTICS = 3;
공개 정적 최종 RewriterConfig.MemOptType 매뉴얼
Driven by manual op-level annotations.
MANUAL = 2;
공개 정적 최종 RewriterConfig.MemOptType NO_MEM_OPT
Disabled in the meta-optimizer.
NO_MEM_OPT = 1;
공개 정적 최종 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;
공개 정적 최종 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;
공개 정적 최종 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;
공개 정적 최종 RewriterConfig.MemOptType 인식되지 않음
상수
공개 정적 최종 int DEFAULT_MEM_OPT_VALUE
The default setting (SCHEDULING and SWAPPING HEURISTICS only)
DEFAULT_MEM_OPT = 0;
상수값: 0
공개 정적 최종 int HEURISTICS_VALUE
Use any combination of swapping and recomputation heuristics.
HEURISTICS = 3;
상수값: 3
공개 정적 최종 int MANUAL_VALUE
Driven by manual op-level annotations.
MANUAL = 2;
상수값: 2
공개 정적 최종 int NO_MEM_OPT_VALUE
Disabled in the meta-optimizer.
NO_MEM_OPT = 1;
상수값: 1
공개 정적 최종 int 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;
상수값: 5
공개 정적 최종 int 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;
상수값: 6
공개 정적 최종 int 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;
상수값: 4