Crate caffe2_util

Source

Macros§

NEXT_UNSAFE
const_assertx
Usage: const_assert!(Var1: Ty, Var2: Ty, ... => expression)
impl_align_n
impl_signed
test_fatal_signal
test_fatal_signal_no_print

Structs§

AVCodecContext
AVFrame
AVIOContext
AVMediaType
AVPacket
AVPixelFormat
Align16
| The data blocks processed by the | SIMD instructions (e.g. SSE2), | which must always be before any | other variables in the class, in | order to be properly aligned to 16 | bytes.
Align32
| The data blocks processed by the | SIMD instructions (e.g. SSE2), | which must always be before any | other variables in the class, in | order to be properly aligned to 16 | bytes.
Align64
| The data blocks processed by the | SIMD instructions (e.g. SSE2), | which must always be before any | other variables in the class, in | order to be properly aligned to 16 | bytes.
Align128
| The data blocks processed by the | SIMD instructions (e.g. SSE2), | which must always be before any | other variables in the class, in | order to be properly aligned to 16 | bytes.
Align256
| The data blocks processed by the | SIMD instructions (e.g. SSE2), | which must always be before any | other variables in the class, in | order to be properly aligned to 16 | bytes.
ArrayBase
AveragePoolFp320p
Bookkeeper
CPUGeneratorImpl
CPUOp
CUDAStream
CompressedSparseColumn
ConstEigenArrayMap
Context
CpuId
| Identification of an Intel CPU. | | Supports CPUID feature flags (EAX=1) | and extended features (EAX=7, ECX=0). | | Values from http://www.intel.com/content/www/us/en/processors/processor-identification-cpuid-instruction-note.html |
CuBlasStatus
CuRandGenerator
CuRandStatus
CublasHandle
CudaDeviceProp
CudaEvent
CudaGuard
CudaStream
CudnnActivationDescriptor
CudnnConvolutionBwdDataAlgo
CudnnConvolutionBwdFilterAlgo
CudnnConvolutionDescriptor
CudnnConvolutionFwdAlgo
CudnnDropoutDescriptor
CudnnFilterDescriptor
CudnnHandle
CudnnLRNDescriptor
CudnnPoolingDescriptor
CudnnPoolingMode
CudnnRNNDescriptor
CudnnStatus
CudnnTensorDescriptor
CudnnTensorFormat
DNNLowPOp
Database
DatabaseMode
Dim3
EArrXXt
EArrXf
EArrXi
ERArrXXf
EigenArrayMap
FixedDivisor
FixedDivisorI32
| Utility class for quickly calculating | quotients and remainders for a known integer | divisor | | Works for any positive divisor, 1 to | INT_MAX. One 64-bit multiplication and one | 64-bit shift is used to calculate the result.
FixedType
FunctionSchema
Fuser
GPUContext
HIPContext
IAlgo
IAttr
IDEEPConvolutionForwardParams
IDEEPScale
IDEEPTensor
IDEEPTensorDescriptor
IDEEPTensorDims
IDType
IFormat
ILowpKind
IProp
IScale
ITensor
ITensorDescriptor
ITensorDims
IValue
IntrusivePtr
LabelType
List
MaxPoolFp320p
MemoryType
NNGraph_EdgeRef
NNGraph_NodeRef
NNGraph_SubgraphType
NoOpFunctor
OnnxBackend
OnnxBackendGraphMap
OnnxBackendID
OnnxExporter
OnnxGraph
OnnxModelProto
OnnxSharedPtrBackendGraphInfo
OnnxStatus
OnnxTensorDescriptorV1
OnnxTypeProto
OnnxValueInfoProto
OnnxifiLibrary
OperatorHandle
OperatorName
PackBMatrix
PackWeightMatrixForGConv
PackWeightMatrixForGConvB
PackedDepthWiseConvMatrix
PatternMatchType
PerGpuCudnnStates
QnnpOperator
RequantizationParams
ScalingParamType
SerializationFormat
SigAction
SignalHandler
SignalHandlerInfo
SimpleQueue
SkipOutputCopy
Stack
SwrContext
TensorCoreEngine
Test
TestMatchGraph_NodeRef
TestWithParam
Types
ZipArchive
ZmqContext

ZmqMessage

ZmqSocket

Enums§

CudnnDataType
SignalHandlerAction

Constants§

kSignalHandlers

Traits§

GetSigned
Invoke
Named

Functions§

compare_div_mod
dummy_thread
edit_distance
edit_distance_helper
ends_with
fmix32
| Finalization mix - force all bits of | a hash block to avalanche |
fmix64
fork_and_pipe
get
get_cast_data_type
get_cpu_id
getblock32
| Block read - if your platform needs to | do endian-swapping or can only handle | aligned reads, do the conversion here |
getblock64
handle_signal
hookup_handler
is_pointer_aligned
murmur_hash3_x64_128
murmur_hash3_x86_32
murmur_hash3_x86_128
rotl32
| MurmurHash3 was written by Austin Appleby, and | is placed in the public domain. The author | hereby disclaims copyright to this source code. | | Note - The x86 and x64 versions do not | produce the same results, as the algorithms are | optimized for their respective platforms. You | can still compile and run any of them on any | platform, but your performance with the | non-native version will be less than optimal.
rotl64
split
starts_with
sum_into
Performs acc[i] += sum_j toSum_j[i] pointwise
to
trim
unhook_handler
Set the signal handlers to the default.
wipe_cache

Type Aliases§

Array2d
Array2f
Array3d
Array3f
ArrayXd
ArrayXf
Vector2d
Vector2f
Vector3d
Vector3f
VectorXd
VectorXf