Crate megenginelite_sys
source · [−]Structs
\brief Configuration when load and compile the graph
\brief config the network input and output item
\brief the simple layout description
\brief the input and output information when load the network the NetworkIO will remain in the network until the network is destroyed
\brief the inference options which will be translated to megenine
\brief warpper of the MegEngine Tensor
Constants
! when the device information is set in model, so set LITE_DEVICE_DEFAULT ! in lite, which equal to xpu in megengine
< The lowest level and most verbose
< Print only errors
< The lowest level and most verbose
< Print only warning and errors
! for scene adreno 630 use adreno 506 cache
! default not compat for series and vendor
! for scene adreno 640 use adreno 630 cache
! Tensor is input
! Tensor maybe input or output
! Tensor is output
Statics
! define a default TensorDesc
! define a default IO
! define a default LiteLayout
! define a default Options
Type Definitions
\brief The Error Code
\brief operation algorithm seletion strategy type, some operations have multi algorithms, different algorithm has different attribute, according to the strategy, the best algorithm will be selected.
\brief A user-implemented allocator function
\brief Model decryption function
! free the memory pointed by ptr in the given device
\brief the input and output type, include SHAPE and VALUE sometimes user only need the shape of the output tensor
\brief The log level.
\brief The network is construct form a model, implement model load, init, forward, and display some model information
\brief enum for cache compat level, for example: adreno 630 cache may be apply to
adreno 640, if you do not want search cache for adreno 640, just config SERIES_COMPAT
or VENDOR_COMPAT, adreno 506 cache may be apply to adreno 630, if you do not want
search cache for adreno 630, just config VENDOR_COMPAT
WARN: this config just let program_cache_io try use a old cache for device compile the cache do not means MegEngine will insure the compile will be ok! its a device CL driver behavior, if compile failed!, MegEngine will try build from source, Whats
more, even though compile from binary success, this cross-use-cache may affect
performance, VENDOR_COMPAT will contain SERIES_COMPAT
\brief Model information parse function
\brief the start/finish callback function \param unordered_map map from the io tensor name to the pair of which is the corresponding IO of user config and the realy input or output tensor.
\brief The pointer to a Lite Tensor object
\brief the thread affinith callback type \param thread_id thread_id is the a number begin from 0 to (nr_threads - 1), thread_id of (nr_threads - 1) is the main worker thread.