Functions — github.com/bytedance/ByteTransformer

CodeHub MCPcopy Create free account

hub / github.com/bytedance/ByteTransformer / functions

Functions317 in github.com/bytedance/ByteTransformer

⨍ Functions 317◇ Types & classes 134

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_iterator.h:564

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/predicated_tile_iterator.h:120

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/predicated_tile_iterator.h:477

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/predicated_tile_iterator.h:736

MethodPredicatedTileAccessIterator

Constructs a TileIterator from its precomputed state, threadblock offset, and thread ID

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_access_iterator.h:675

MethodPredicatedTileAccessIterator

Constructs a TileIterator from its precomputed state, threadblock offset, and thread ID

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_access_iterator.h:864

FunctionPredicatedTileAccessIterator operator++

Increment and return an instance to self.

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_access_iterator.h:532

MethodPredicatedTileAccessIteratorPredicates

Constructs a TileIterator from its precomputed state, threadblock offset, and thread ID

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_access_iterator.h:219

MethodPredicatedTileIterator

Constructs a TileIterator from its precomputed state, threadblock offset, and thread ID

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_iterator.h:144

MethodPredicatedTileIterator

Constructs a TileIterator from its precomputed state, threadblock offset, and thread ID

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_iterator.h:390

MethodPredicatedTileIterator

Constructs a TileIterator from its precomputed state, threadblock offset, and thread ID

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_iterator.h:588

MethodPredicatedTileIterator

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/predicated_tile_iterator.h:212

MethodSoftmaxEpilogue

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/softmax_epilogue.h:103

MethodSoftmaxFusedEpilogue

Constructs the function object, possibly loading from pointers in host memory

bytetransformer/include/cutlass_attention_operators.h:614

FunctionTensorCoord get_access_offset

Returns offset of access in tensor

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_access_iterator.h:576

FunctionTensorCoord thread_offset

Returns the thread offset

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_access_iterator.h:570

FunctionTransformerEncoder

bytetransformer/th_op/ths_op_f.cc:28

Functionact_fun

bytetransformer/include/common.h:70

Functionadd_bias_act

bytetransformer/include/gemm_bias_act.h:332

Functionadd_bias_half_input_layernorm_kernel_launcher

bytetransformer/include/layernorm.h:710

Functionadd_bias_half_input_layernorm_restore_output_kernel_launcher

bytetransformer/include/layernorm.h:728

Functionadd_bias_half_input_layernorm_restore_output_v2

bytetransformer/include/layernorm.h:580

Functionadd_bias_half_input_layernorm_v2

bytetransformer/include/layernorm.h:530

Functionadd_bias_half_input_out_layernorm_kernel_launcher

bytetransformer/include/layernorm.h:746

Functionadd_bias_half_input_out_layernorm_v2

bytetransformer/include/layernorm.h:659

Functionadd_bias_input_layernorm_kernel_launcher

bytetransformer/include/layernorm.h:411

Functionadd_bias_input_layernorm_restore_output_kernel_launcher

bytetransformer/include/layernorm.h:431

Functionadd_bias_input_layernorm_restore_output_v2

bytetransformer/include/layernorm.h:299

Functionadd_bias_input_layernorm_v2

bytetransformer/include/layernorm.h:254

Functionadd_bias_input_out_layernorm_kernel_launcher

bytetransformer/include/layernorm.h:505

Functionadd_bias_input_out_layernorm_v2

bytetransformer/include/layernorm.h:456

Functionadd_pointer_offset

Adds a pointer offset in units of Element

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/predicated_tile_iterator.h:254

Methodadd_pointer_offset

Adds a pointer offset in units of Element

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_access_iterator.h:710

Methodadd_pointer_offset

Adds a pointer offset in units of Element

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_iterator.h:170

Methodadd_pointer_offset

Adds a pointer offset in units of Element

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_iterator.h:414

Methodadd_pointer_offset

Adds a pointer offset in units of Element

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/predicated_tile_iterator.h:861

Functionadd_tile_offset

Advances an iterator along logical dimensions of matrix in units of whole tiles

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_access_iterator.h:449

Methodadd_tile_offset

Advances an iterator along logical dimensions of matrix in units of whole tiles

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_access_iterator.h:717

Methodapply_output_operator_

Helper to invoke the output functor over each vector of output

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/epilogue.h:448

Methodapply_output_operator_source_not_needed_

Helper to invoke the output functor over each vector of output

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/epilogue.h:480

FunctionblockReduceMax

bytetransformer/include/reduce.h:97

FunctionblockReduceSum

bytetransformer/include/reduce.h:40

Functionbuild_sequence_length_padding_offset_kernelLauncher

bytetransformer/include/remove_padding.h:89

Methodcal_bufsize

bytetransformer/include/bert_transformer.h:193

Methodcal_bufsize

bytetransformer/include/cutlass_attention.h:124

Methodcalc_thread_column_offset

bytetransformer/include/cutlass_attention_operators.h:940

Methodcalc_thread_contiguous_offset

bytetransformer/include/cutlass_attention_operators.h:774

Methodcalc_thread_row_offset

bytetransformer/include/cutlass_attention_operators.h:448

Methodcalc_thread_row_offset

bytetransformer/include/cutlass_attention_operators.h:936

Methodcalc_thread_strided_offset

bytetransformer/include/cutlass_attention_operators.h:770

Methodcan_implement

Determines whether the GEMM can execute the given problem.

cutlass_contrib/include/cutlass/contrib/gemm/device/gemm_grouped.h:74

Methodcan_implement

Determines whether kernel satisfies alignment

cutlass_contrib/include/cutlass/contrib/gemm/kernel/gemm_grouped.h:417

bytetransformer/include/common.h:131

Methodcheck_model_type_supported

bytetransformer/include/cutlass_attention.h:128

Methodcheck_seqlen_supported

bytetransformer/include/cutlass_attention.h:132

< Efficiently disables all accesses guarded by mask

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/predicated_tile_iterator.h:510

< Efficiently disables all accesses guarded by mask

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/predicated_tile_iterator.h:763

Functionclear_mask

Clears the predicate set efficiently

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_access_iterator.h:540

Functionclear_mask

< Efficiently disables all accesses guarded by mask

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/predicated_tile_iterator.h:419

Methodclear_mask

Clears the predicate set efficiently

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_access_iterator.h:754

Methodclear_mask

Clears the predicate set efficiently

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_access_iterator.h:943

Methodclear_mask

Clears the predicate set efficiently

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_iterator.h:205

Methodclear_mask

Clears the predicate set efficiently

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_iterator.h:443

Methodclear_mask

< Efficiently disables all accesses guarded by mask

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/predicated_tile_iterator.h:934

FunctioncompressBertInput_kernelLauncher

bytetransformer/include/remove_padding.h:114

Functioncompress_bert_input

bytetransformer/include/remove_padding.h:105

Methodcompute_predicates_

Computes predicates based on internally tracked per-thread offset.

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_access_iterator.h:387

Methodcompute_source_needed_

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/softmax_epilogue.h:117

Methodcompute_source_needed_

Streams the result to global memory

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/epilogue.h:364

Methodcompute_source_no_needed_

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/softmax_epilogue.h:162

Methodcompute_source_not_needed_

Streams the result to global memory

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/epilogue.h:256

< CUTLASS_HOST_DEVICE enables all accesses guarded by mask

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/predicated_tile_iterator.h:518

< CUTLASS_HOST_DEVICE enables all accesses guarded by mask

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/predicated_tile_iterator.h:771

Functionenable_mask

Clears the predicate set efficiently

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_access_iterator.h:546

Functionenable_mask

< Efficiently enables all accesses guarded by mask

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/predicated_tile_iterator.h:424

Methodenable_mask

Clears the predicate set efficiently

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_access_iterator.h:760

Methodenable_mask

Clears the predicate set efficiently

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_access_iterator.h:949

Methodenable_mask

Clears the predicate set efficiently

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_iterator.h:211

Methodenable_mask

Clears the predicate set efficiently

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_iterator.h:449

Methodenable_mask

< Efficiently enables all accesses guarded by mask

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/predicated_tile_iterator.h:663

Methodenable_mask

< Efficiently enables all accesses guarded by mask

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/predicated_tile_iterator.h:939

Functionextent_row

Extent of the matrix in rows

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/predicated_tile_iterator.h:380

bytetransformer/th_op/bert_transformer_ext.h:80

Methodforwardv2

bytetransformer/th_op/ths_op.h:84

Returns a pointer

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_access_iterator.h:481

Returns a pointer

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_access_iterator.h:723

Methodget_extra_workspace_size

cutlass_contrib/include/cutlass/contrib/gemm/kernel/gemm_grouped.h:425

Methodget_grid_shape

Computes the grid shape

cutlass_contrib/include/cutlass/contrib/gemm/device/gemm_grouped.h:85

Functionget_mask

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_access_iterator.h:558

Functionget_mask

< Sets the mask

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/predicated_tile_iterator.h:429

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_access_iterator.h:772

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_access_iterator.h:961

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_iterator.h:223

cutlass_contrib/include/cutlass/contrib/transform/threadblock/predicated_tile_iterator.h:461

< Sets the mask

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/predicated_tile_iterator.h:668

< Sets the mask

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/predicated_tile_iterator.h:944

Functionget_ptr

bytetransformer/th_op/util.h:50

Methodget_workspace_size

Gets the workspace size

cutlass_contrib/include/cutlass/contrib/gemm/device/gemm_grouped.h:79

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/epilogue.h:215

cutlass_contrib/include/cutlass/contrib/epilogue/threadblock/epilogue.h:344

← previous next →101–200 of 317, ranked by callers