org.platanios.tensorflow.api.core.client
GPU allocation strategy.
Graph optimization level.
EXPERIMENTAL Graph JIT compilation level.
A "best-fit with coalescing" GPU allocation algorithm, simplified from a version of dlmalloc
.
System-chosen default GPU allocation strategy which may change over time.
Default graph JIT compilation level, which is currently set to JIT not being used.
Level-1 graph optimizations performed, which include common subexpression elimination and constant folding.
Level-1 JIT compilation performed.
Level-1 JIT compilation performed. Higher JIT levels are more aggressive. Higher levels may reduce opportunities for parallelism and may use more memory (at present, there is no distinction between L1 and L2 but this is expected to change).
Level-2 JIT compilation performed.
Level-2 JIT compilation performed. Higher JIT levels are more aggressive. Higher levels may reduce opportunities for parallelism and may use more memory (at present, there is no distinction between L1 and L2 but this is expected to change).
No graph optimization performed.
No JIT compilation performed.
Contains helper methods for dealing with SessionConfigs.