MCPcopy Create free account

hub / github.com/NVIDIA-Merlin/HugeCTR / functions

Functions3,072 in github.com/NVIDIA-Merlin/HugeCTR

↓ 1 callersFunctionmain
(args)
test/sok_perf_test/demo/run_tf.py:25
↓ 1 callersFunctionmain
(args)
test/sok_perf_test/demo/gen_data.py:32
↓ 1 callersMethodmax_nnz
HugeCTR/include/tensor2.hpp:316
↓ 1 callersFunctionmax_per_line_cpu
test/utest/core23_layer_test/masked_softmax_layer_test.cpp:39
↓ 1 callersMethodmodel_backward_per_gpu
(self)
notebooks/prototype_embedding_collection/embedding.py:320
↓ 1 callersMethodmodel_forward_per_gpu
(self, embedding_table)
notebooks/prototype_embedding_collection/embedding.py:253
↓ 1 callersFunctionmomentum_update_grad
* Momentum SGD * ------------ * v_i = beta * v_i + g_i / s * g_i = -eta * v_i */
HugeCTR/embedding_storage/optimizers.hpp:46
↓ 1 callersFunctionmulti_head_attention_3d_cpu_fused
test/utest/core23_layer_test/multi_head_attention_layer_test.cpp:318
↓ 1 callersFunctionmulti_head_attention_cpu_fused
test/utest/core23_layer_test/multi_head_attention_layer_test.cpp:299
↓ 1 callersFunctionmulti_head_attention_dgrad_3d_cpu_fused
test/utest/core23_layer_test/multi_head_attention_layer_test.cpp:420
↓ 1 callersFunctionmulti_head_attention_dgrad_cpu_noT
test/utest/core23_layer_test/multi_head_attention_layer_test.cpp:404
↓ 1 callersFunctionnaiveCoalescedReduction2d
test/utest/prims/reduce.h:93
↓ 1 callersFunctionnaiveStridedReduction2d
test/utest/prims/reduce.h:107
↓ 1 callersMethodname
HugeCTR/core23/data_type.cpp:37
↓ 1 callersFunctionnccl_allreduce
(send_recv_buffer, count, comm)
notebooks/prototype_embedding_collection/utils.py:77
↓ 1 callersFunctionnccl_recv
(recv_tensor, recv_count, peer, comm)
notebooks/prototype_embedding_collection/utils.py:73
↓ 1 callersFunctionnccl_send
(send_tensor, send_count, peer, comm)
notebooks/prototype_embedding_collection/utils.py:70
↓ 1 callersFunctionnesterov_update_grad
* Nesterov Momentum * ----------------- * w*_i = w_i + beta * v_i * g = f(w*) * v_i = beta * v_i + g_i / s * g_i = -eta * v_i */
HugeCTR/embedding_storage/optimizers.hpp:70
↓ 1 callersMethodnetwork_backward_per_gpu
(self, top_grad)
notebooks/prototype_embedding_collection/embedding.py:300
↓ 1 callersMethodnetwork_forward_per_gpu
(self, output_buffer)
notebooks/prototype_embedding_collection/embedding.py:280
↓ 1 callersMethodnnz
HugeCTR/include/sparse_tensor.hpp:120
↓ 1 callersMethodnum_bytes
HugeCTR/core23/tensor_container.hpp:143
↓ 1 callersMethodnum_gpu_per_rank
(self)
sparse_operation_kit/sparse_operation_kit/communication.py:138
↓ 1 callersMethodnum_gpus
(self)
sparse_operation_kit/sparse_operation_kit/communication.py:144
↓ 1 callersMethodnum_nodes
* Get the total number of nodes. */
HugeCTR/include/device_map.hpp:165
↓ 1 callersMethodnum_parameters_per_weight
HugeCTR/include/optimizer.hpp:156
↓ 1 callersFunctionnum_ranks
()
sparse_operation_kit/sparse_operation_kit/communication.py:200
↓ 1 callersMethodnum_ranks
(self)
sparse_operation_kit/sparse_operation_kit/communication.py:129
↓ 1 callersMethodon_allocate
HugeCTR/core23/buffer_client.cpp:34
↓ 1 callersFunctionon_multi_gpu
(args)
test/sok_perf_test/demo/run_sok.py:85
↓ 1 callersFunctionon_single_gpu
(args)
test/sok_perf_test/demo/run_sok.py:25
↓ 1 callersMethodon_subscribe
HugeCTR/core23/buffer_client.cpp:26
↓ 1 callersMethodon_unsubscribe
HugeCTR/core23/buffer_client.cpp:30
↓ 1 callersFunctionpack_dense_features
row major to `extended-col-major`
test/utest/data_reader/data_reader_parquet_test.cpp:63
↓ 1 callersFunctionparse_args
()
tools/criteo_script/preprocess_nvt.py:300
↓ 1 callersFunctionparse_args
()
samples/din/utils/4_nvt_process.py:104
↓ 1 callersFunctionparse_args
()
samples/bst/utils/4_nvt_process.py:170
↓ 1 callersFunctionparse_args
()
samples/dlrm/preprocessing/convert_to_raw.py:172
↓ 1 callersFunctionparse_args
(parser)
test/pybind_test/model_test.py:39
↓ 1 callersFunctionparse_args
(argv: List[str])
benchmarks/embedding_collection/hugectr/train.py:26
↓ 1 callersFunctionparse_config
(src_config)
tools/criteo_predict/criteo2predict.py:25
↓ 1 callersMethodpop_bucket
( self, bucket_id: int, )
samples/dlrm/sharding/planner.py:122
↓ 1 callersMethodpop_bucket
( self, bucket_id: int, )
benchmarks/embedding_collection/hugectr/sharding/planner.py:249
↓ 1 callersMethodpost_completion
HugeCTR/src/collectives/ib_proxy.cpp:43
↓ 1 callersMethodpost_set_source
HugeCTR/include/data_readers/data_reader_worker_interface.hpp:78
↓ 1 callersMethodpre_set_source
HugeCTR/include/data_readers/data_reader_worker_interface.hpp:77
↓ 1 callersMethodprint
HugeCTR/core23/logger.hpp:286
↓ 1 callersMethodprint_info
HugeCTR/embedding_storage/weight_io/data_info.hpp:118
↓ 1 callersFunctionprocess
(args)
samples/din/utils/4_nvt_process.py:38
↓ 1 callersFunctionprocess
(args)
samples/bst/utils/4_nvt_process.py:62
↓ 1 callersFunctionprocess_NVT
(args)
tools/criteo_script/preprocess_nvt.py:77
↓ 1 callersFunctionpython_concat
test/utest/core23_layer_test/concat_3d_layer_test.cpp:43
↓ 1 callersFunctionrank
()
sparse_operation_kit/sparse_operation_kit/communication.py:195
↓ 1 callersMethodread_a_batch
HugeCTR/src/pybind/model.cpp:1022
↓ 1 callersMethodread_a_batch
test/utest/embedding/sparse_embedding_hash_cpu.hpp:343
↓ 1 callersMethodread_a_batch_to_device_delay_release
HugeCTR/src/data_readers/data_reader.cpp:191
↓ 1 callersMethodread_group
HugeCTR/src/data_readers/file_source_parquet.cpp:141
↓ 1 callersFunctionread_json_file
inline to avoid build error: multiple definition
HugeCTR/include/parser.hpp:51
↓ 1 callersFunctionread_samples_for_bst
( data_file, num_samples=64, key_type="I64", slot_num=23, slot_shift=slot_shift )
test/onnx_converter_test/utils.py:228
↓ 1 callersFunctionread_samples_for_dcn
( data_file, num_samples=64, key_type="I64", slot_num=26, slot_shift=dcn_offset )
test/onnx_converter_test/utils.py:107
↓ 1 callersFunctionread_samples_for_din
( data_file, num_samples=64, key_type="I64", slot_num=23, slot_shift=slot_shift )
test/onnx_converter_test/utils.py:200
↓ 1 callersFunctionread_samples_for_mmoe
(data_file, key_slot_shift, num_samples=64, slot_num=32)
test/onnx_converter_test/utils.py:213
↓ 1 callersFunctionread_samples_for_ncf
(data_file, num_samples=64, key_type="I32", slot_num=2)
test/onnx_converter_test/utils.py:158
↓ 1 callersFunctionread_samples_for_wdl
( data_file, num_samples=64, key_type="I64", slot_num=26, slot_shift=wdl_offset )
test/onnx_converter_test/utils.py:130
↓ 1 callersMethodrelease
HugeCTR/include/data_readers/multi_hot/detail/batch_file_reader.hpp:60
↓ 1 callersMethodrelease_batch
HugeCTR/src/data_readers/multi_hot/detail/batch_file_reader.cpp:146
↓ 1 callersFunctionremove_at
benchmarks/core23/random_allocations.cpp:38
↓ 1 callersFunctionreplace_table_id
(shard_matrix, shard_strategy)
benchmarks/embedding_collection/hugectr/sharding/generate_plan.py:86
↓ 1 callersMethodreserved_size
HugeCTR/core23/buffer.hpp:58
↓ 1 callersMethodreset_accomplished_worker
HugeCTR/src/data_readers/row_group_reading_thread.cpp:109
↓ 1 callersMethodreset_read_flag
HugeCTR/src/data_readers/row_group_reading_thread.cpp:103
↓ 1 callersMethodreset_shape
HugeCTR/include/tensor2.hpp:193
↓ 1 callersMethodreset_shard_ll
(self)
samples/dlrm/sharding/planner.py:112
↓ 1 callersMethodreset_shard_ll
(self)
benchmarks/embedding_collection/hugectr/sharding/planner.py:236
↓ 1 callersMethodreset_source
HugeCTR/src/data_readers/row_group_reading_thread.cpp:186
↓ 1 callersFunctionrms_prop_update_grad
* RMSProp * ------- * g_i = g_i / s * v_i = beta * v_i + (1 - beta) * g_i^2 * g_i = -eta * g_i / (sqrt(v_i) + epsilon) */
HugeCTR/embedding_storage/optimizers.hpp:116
↓ 1 callersFunctionrowwise_sharding_plan
()
test/embedding_collection_test/dgx_a100_one_hot.py:181
↓ 1 callersFunctionrun
( src_dir, mid_dir, dst_dir, parallel_jobs=40, shuffle=False, split_category=False,
sparse_operation_kit/documents/tutorials/DLRM_Benchmark/preprocess/parquet_to_binary.py:75
↓ 1 callersFunctions3_configs_test
test/utest/io/s3_backend_test.cpp:27
↓ 1 callersFunctions3_path_test
test/utest/io/s3_backend_test.cpp:34
↓ 1 callersFunctions3_simple_read_write_test_with_dsp
test/utest/io/s3_backend_test.cpp:53
↓ 1 callersFunctions3_simple_read_write_test_with_path
test/utest/io/s3_backend_test.cpp:66
↓ 1 callersFunctionsanity_check
(shard_matrix, shard_strategy)
samples/dlrm/sharding/generate_plan.py:31
↓ 1 callersFunctionsanity_check
(shard_matrix, shard_strategy)
benchmarks/embedding_collection/hugectr/sharding/generate_plan.py:73
↓ 1 callersFunctionsave_graph_to_json
HugeCTR/src/pybind/add_dense_layer.cpp:67
↓ 1 callersMethodsave_model
(self, model_path, op_version=10, ir_version=7)
onnx_converter/hugectr2onnx/graph_builder.py:1708
↓ 1 callersFunctionsave_optimizer_to_filesysyem_dynamic
(optimizer, var, path, sok_var_info)
sparse_operation_kit/sparse_operation_kit/dump_load.py:475
↓ 1 callersFunctionsave_optimizer_to_filesysyem_static
(optimizer, var, path, sok_var_info)
sparse_operation_kit/sparse_operation_kit/dump_load.py:416
↓ 1 callersFunctionsave_table_to_filesystem_dynamic
(var, optimizer, path, have_states)
sparse_operation_kit/sparse_operation_kit/dump_load.py:660
↓ 1 callersMethodscatter_add
(self, sparse_delta, use_locking=False, name=None)
sparse_operation_kit/sparse_operation_kit/dynamic_variable.py:316
↓ 1 callersMethodscatter_sub
(self, sparse_delta, use_locking=False, name=None)
sparse_operation_kit/sparse_operation_kit/dynamic_variable.py:305
↓ 1 callersFunctionsegmented_sort
(keys_in, num_items, values_in, begin_offsets, end_offsets, num_segments)
notebooks/prototype_embedding_collection/op.py:14
↓ 1 callersFunctionsequence_mask_cpu
test/utest/core23_layer_test/sequence_mask_layer_test.cpp:54
↓ 1 callersMethodset
HugeCTR/embedding_storage/weight_io/data_info.hpp:92
↓ 1 callersFunctionset_affinity
(rank)
sparse_operation_kit/documents/tutorials/DenseDemo/run_sok_MultiWorker_mpi.py:136
↓ 1 callersFunctionset_affinity
(rank)
sparse_operation_kit/documents/tutorials/DenseDemo/run_tf.py:144
↓ 1 callersMethodset_all_from_json
HugeCTR/src/io/s3_filesystem.cpp:65
↓ 1 callersMethodset_all_from_json
HugeCTR/src/io/hadoop_filesystem.cpp:41
↓ 1 callersMethodset_ar_comm
HugeCTR/src/collectives/collective.cpp:34
↓ 1 callersMethodset_buffer
HugeCTR/core23/buffer_channel.hpp:53
← previousnext →901–1,000 of 3,072, ranked by callers