Functions — github.com/LeelaChessZero/lc0

CodeHub MCPcopy Create free account

hub / github.com/LeelaChessZero/lc0 / functions

Functions2,664 in github.com/LeelaChessZero/lc0

⨍ Functions 2,664◇ Types & classes 523

↓ 2 callersFunctionLoadWeights

src/neural/loader.cc:213

↓ 2 callersMethodLoadWeights0

src/neural/backends/cuda/layers.cc:1180

↓ 2 callersMethodLoadWeights1

src/neural/backends/cuda/layers.cc:1207

↓ 2 callersMethodLogPlusOne

src/neural/xla/hlo_builder.cc:360

↓ 2 callersMethodLookupForwardFieldType

(self, object_stack)

scripts/compile_proto.py:191

↓ 2 callersMethodMakeNotTerminal

src/search/classic/node.cc:319

↓ 2 callersFunctionMakePolicyMap

src/neural/onnx/converter.cc:802

↓ 2 callersMethodMatrixMultiply

src/neural/backends/dx/shader_wrapper.cc:456

↓ 2 callersFunctionMoveAsPackedInt

src/neural/encoder.cc:588

↓ 2 callersMethodNewComputation

src/neural/backends/network_tf_cc.cc:499

↓ 2 callersMethodNewGame

src/engine.cc:228

↓ 2 callersMethodNotifyComplete

src/neural/backends/network_demux.cc:73

↓ 2 callersMethodNotifyReady

src/neural/backends/network_mux.cc:72

↓ 2 callersFunctionNumBits

src/neural/backends/network_trivial.cc:405

↓ 2 callersMethodOnSearchDone

src/search/classic/stoppers/smooth.cc:629

↓ 2 callersFunctionOnnxShapeToHloTensorType

Converts an ONNX shape to an XLA shape, replacing the batch dimension with the provided batch size.

src/neural/xla/onnx2hlo.cc:392

↓ 2 callersFunctionOnnxTensorToHloTensorType

Type is not a field of the ONNX tensor, so this function extracts the shape and converts it (discarding the data).

src/neural/xla/onnx2hlo.cc:415

↓ 2 callersFunctionOnnxTensorToXlaLiteral

Converts an ONNX tensor to an XLA literal (which is a shape and a data).

src/neural/xla/onnx2hlo.cc:425

↓ 2 callersMethodPerformConv

src/neural/backends/dx/layers_dx.cc:364

↓ 2 callersMethodPerformGemm

src/neural/backends/dx/layers_dx.cc:246

↓ 2 callersFunctionPolicyMap

src/neural/backends/sycl/common_kernels.dp.cpp:1001

↓ 2 callersFunctionPopulateIntrinsicStoppers

Parameters needed for selfplay and uci, but not benchmark nor infinite mode.

src/search/classic/stoppers/common.cc:99

↓ 2 callersMethodPopulateParams

src/chess/uciloop.cc:259

↓ 2 callersFunctionPrettyPrintHlo

src/neural/xla/print_hlo.cc:431

↓ 2 callersFunctionProcessFileInternal

src/trainingdata/rescorer.cc:1129

↓ 2 callersFunctionProcessFiles

src/trainingdata/rescorer.cc:1193

↓ 2 callersMethodRaiseError

src/utils/optionsdict.cc:145

↓ 2 callersMethodReadChunk

src/trainingdata/reader.cc:130

↓ 2 callersFunctionReadFixed

src/utils/protomessage.cc:30

↓ 2 callersFunctionReadIdentifierPath

Reads qualified identifier a.b.d into ['a', 'b', 'd'] list

scripts/compile_proto.py:150

↓ 2 callersMethodRegisterUciResponder

src/engine.cc:266

↓ 2 callersMethodRunBlocking

src/selfplay/tournament.cc:593

↓ 2 callersFunctionSe_Fp16_NHWC

src/neural/backends/sycl/fp16_kernels.dp.cpp:139

↓ 2 callersMethodSelect

src/neural/xla/hlo_builder.cc:379

↓ 2 callersMethodSendId

src/chess/uciloop.cc:274

↓ 2 callersMethodSet

src/utils/optionsdict.h:344

↓ 2 callersMethodSetLowNode

src/search/dag_classic/node.cc:423

↓ 2 callersMethodSetNNEval

src/search/dag_classic/node.h:417

↓ 2 callersMethodSetPosition

src/engine.cc:215

↓ 2 callersMethodSetRepetition

src/search/dag_classic/node.h:314

↓ 2 callersMethodShouldStop

src/search/classic/stoppers/smooth.cc:584

↓ 2 callersFunctionShowNetworkWeightsInfo

src/tools/describenet.cc:212

↓ 2 callersMethodSigmoid

src/neural/onnx/builder.cc:366

↓ 2 callersMethodSlice

src/neural/xla/hlo_builder.cc:446

↓ 2 callersMethodSoftmax

src/neural/onnx/builder.cc:240

↓ 2 callersMethodSplit

src/neural/onnx/builder.cc:302

↓ 2 callersMethodSqrt

src/neural/xla/hlo_builder.cc:352

↓ 2 callersMethodStartSearch

src/search/classic/wrapper.cc:111

↓ 2 callersMethodStartThreads

src/search/classic/search.cc:894

↓ 2 callersFunctionSteadyClockToSystemClock

src/utils/logging.cc:96

↓ 2 callersFunctionStringsEqualIgnoreCase

src/utils/string.cc:93

↓ 2 callersMethodTanh

src/neural/onnx/builder.cc:234

↓ 2 callersMethodTanh

src/neural/xla/hlo_builder.cc:348

↓ 2 callersMethodTestAndReset

src/utils/optionsdict.h:153

↓ 2 callersFunctionTransposeTensor

src/utils/transpose.h:41

↓ 2 callersMethodTrimTreeAtHead

src/search/classic/node.cc:483

↓ 2 callersMethodTryStartScoreUpdate

src/search/dag_classic/node.cc:321

↓ 2 callersMethodTryStartScoreUpdate

src/search/classic/node.cc:348

↓ 2 callersMethodUnindent

scripts/pybind/writer.py:36

↓ 2 callersMethodUnpin

Unpins the element given key and value. Use of HashedKeyCacheLock is recommended to automate this pin management.

src/utils/cache.h:133

↓ 2 callersMethodUpdateChildrenParents

src/search/classic/node.cc:394

↓ 2 callersMethodUpdateEstimatedRemainingTimeMs

src/search/classic/stoppers/timemgr.cc:37

↓ 2 callersMethodWait

src/search/dag_classic/search.cc:1060

↓ 2 callersMethodWait

src/search/classic/search.cc:1056

↓ 2 callersMethodWait

src/selfplay/tournament.cc:609

↓ 2 callersMethodWaitSearch

src/search/classic/wrapper.cc:64

↓ 2 callersMethodWriteChunk

src/trainingdata/writer.cc:69

↓ 2 callersFunctionWriteFixed

src/utils/protomessage.cc:41

↓ 2 callersFunctionWriteStringToFile

src/utils/files.cc:62

↓ 2 callersFunctionXlaTypeToPjrtType

src/neural/backends/xla/xla_runner.cc:74

↓ 2 callersMethod__init__

(self, name, *argv, cpp_name=None, disable

scripts/pybind/__init__.py:53

↓ 2 callersMethod_generate_functions

scripts/pybind/__init__.py:37

↓ 2 callersMethod_generate_params

scripts/pybind/functions.py:161

↓ 2 callersFunctionactivationString

src/neural/backends/metal/network_metal.cc:63

↓ 2 callersFunctioncalc_key_from_pcs

Produce a 64-bit material key corresponding to the material combination defined by pcs[16], where pcs[1], ..., pcs[6] are the number of white pawns, .

src/syzygy/syzygy.cc:449

↓ 2 callersMethodcapacity

src/utils/atomic_vector.h:65

↓ 2 callersMethodcpp_type

scripts/pybind/retval.py:71

↓ 2 callersFunctioncpuTranspose

src/neural/backends/sycl/layers.cc.dp.cpp:158

↓ 2 callersFunctioncpuTranspose

src/neural/backends/cuda/layers.cc:428

↓ 2 callersMethoddestructor_name

scripts/pybind/__init__.py:127

↓ 2 callersFunctionfrom_le_u32

src/syzygy/syzygy.cc:699

↓ 2 callersMethodfunction_list_name

scripts/pybind/__init__.py:130

↓ 2 callersMethodfunction_meth_flags

scripts/pybind/functions.py:96

↓ 2 callersFunctiongaviota_tb_probe_hard

src/trainingdata/rescorer.cc:251

↓ 2 callersMethodgetMaxMatchSize

src/neural/backends/opencl/OpenCL.h:93

↓ 2 callersMethodget_device_name

src/neural/backends/opencl/OpenCL.cc:401

↓ 2 callersMethodhas_mlh

src/neural/network.h:107

↓ 2 callersMethodholds_value

Returns whether lock holds any value.

src/utils/cache.h:303

↓ 2 callersFunctioninit_enc_info

src/syzygy/syzygy.cc:615

↓ 2 callersMethodisNHWC

src/neural/backends/cuda/layers.h:55

↓ 2 callersMethodis_castling

src/chess/types.h:170

↓ 2 callersMethodis_null

TODO remove this once UciReponder starts using std::optional for ponder.

src/chess/types.h:173

↓ 2 callersMethodis_pinned

src/chess/board.h:47

↓ 2 callersMethodis_promotion

src/chess/types.h:166

↓ 2 callersFunctionmatrixMul_gpu_serial_4x4x6

src/neural/backends/dx/shaders/WinogradCommon.h:122

↓ 2 callersFunctionmatrixMul_gpu_serial_4x6x6

src/neural/backends/dx/shaders/WinogradCommon.h:107

↓ 2 callersMethodmax_single_move_time_fraction

Fraction of a total available move time that is allowed to use for a single move.

src/search/classic/stoppers/smooth.cc:71

↓ 2 callersMethodmax_tree_reuse

Do not allow tree reuse expectation to go above this value.

src/search/classic/stoppers/smooth.cc:52

↓ 2 callersMethodmin_smartpruning_timeuse

Do not allow timeuse estimation to fall below this.

src/search/classic/stoppers/smooth.cc:64

↓ 2 callersMethodmutable_data

TODO replace with std::span once C++20 is available.

src/neural/xla/xla_tensor.h:118

← previous next →601–700 of 2,664, ranked by callers