IP address of the server where the NCS is located.
port
Port number of the server where the NCS is located.
device
Device used for tuning in the operating environment.
buffer_optimize
Whether to enable buffer optimization.
compress_weight_conf
Directory (including the file name) of the compression optimization configuration file. This option is used to enable the compression optimization feature specified in the configuration file to improve network performance.
precision_mode
Precision mode of a network model.
disable_reuse_memory
Memory overcommitment enable.
enable_single_stream
Switch for configuring a model to use only one stream.
aicore_num
Number of AI Cores for model compilation. This option is reserved in the current version. If you need to use this option, retain the default.
fusion_switch_file
Directory (including the file name) of the fusion switch configuration file for graph fusion and UB fusion patterns. You can disable selected fusion patterns in the configuration file.
enable_small_channel
Whether to enable small channel tuning to yield performance benefits at convolutional layers with channel size ≤ 4.
You are advised to enable this function in inference scenarios.
op_select_implmode
Operator implementation mode.
In high-precision mode, Newton's Method and Taylor's Formula is used to improve operator precision (fp16). In high-performance mode, the optimal performance is implemented without affecting the network precision (fp16).
optypelist_for_implmode
Implementation mode of an operator on the optype list. Currently, only the Pooling, SoftmaxV2, LRN, and ROIAlign operators are supported.
enable_scope_fusion_passes
Scope fusion pattern (or scope fusion patterns separated by commas) to take effect during compilation.
op_debug_level
TBE operator debug during operator compilation.
virtual_type
Whether AOE can run on virtual devices generated on Ascend virtual instances.
sparsity
Whether to enable global sparsity.
modify_mixlist
Operators on the mixed precision list.
customize_dtypes
Custom computing precision of one or more operators during model building.
framework
Framework of the original network model.
job_type
Tuning mode.
compression_optimize_conf
Path (including the file name) of the model compression configuration file. This option enables the model compression features specified in the configuration file to improve network performance.
init_bypass
Transparently transmits compilation options that are not detectable by the AOE tuning framework and tuning services in the modeling initialization phase. For details about the specific options to be transmitted, see Options (globally effective options).