MCPcopy Create free account

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

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

⨍ Functions 3,072◇ Types & classes 981↳ Endpoints 3
↓ 3 callersMethod__init__
(self, arch, activation="relu", out_activation=None, **kwargs)
sparse_operation_kit/example/sok_train_hps_inference/sok_train.py:87
↓ 3 callersFunctionalltoall
(*args, **kwargs)
sparse_operation_kit/sparse_operation_kit/communication.py:230
↓ 3 callersFunctionbuild_map
(df, col_name)
samples/din/utils/2_remap_id.py:32
↓ 3 callersFunctionbuild_map
(df, col_name)
samples/bst/utils/2_remap_id.py:32
↓ 3 callersFunctioncheck_correctness
test/utest/core23_layer_test/gru_layer_test.cpp:58
↓ 3 callersMethodclean
HugeCTR/src/io/file_loader.cpp:96
↓ 3 callersMethodcommunicate
HugeCTR/embedding/operators/communication.cpp:52
↓ 3 callersMethodcommunication
(self, send_tensors, send_counts, recv_tensors, recv_counts)
notebooks/prototype_embedding_collection/op.py:668
↓ 3 callersFunctioncompare_element
test/utest/embedding/embedding_test_utils.hpp:30
↓ 3 callersMethodconfigure_default
HugeCTR/src/io/s3_filesystem.cpp:59
↓ 3 callersFunctioncopy_gpu_common
HugeCTR/core23/low_level_primitives.cpp:42
↓ 3 callersMethodcopy_weights_for_evaluation
HugeCTR/src/pybind/model.cpp:1260
↓ 3 callersFunctioncpu_mm
test/utest/core23_layer_test/mlp_test.cpp:60
↓ 3 callersFunctioncpu_mm
test/utest/core23_layer_test/fully_connected_layer_half_test.cpp:28
↓ 3 callersMethoddata_type
HugeCTR/core23/details/tensor_impl.hpp:52
↓ 3 callersMethoddevice
HugeCTR/core23/details/tensor_impl.hpp:53
↓ 3 callersMethoddump_embedding_weight
HugeCTR/embedding_storage/weight_io/parameter_IO.cpp:262
↓ 3 callersMethoddump_metadata
HugeCTR/embedding_storage/weight_io/parameter_IO.cpp:179
↓ 3 callersFunctionexport
Abbreviated as ``sok.export``. Export the indices and value tensor from the given variable. Parameters ---------- var: sok.Dyna
sparse_operation_kit/sparse_operation_kit/dynamic_variable.py:465
↓ 3 callersMethodforward_per_gpu
HugeCTR/embedding/data_parallel_embedding.cpp:249
↓ 3 callersFunctionfully_connected_layer_test
test/utest/core23_layer_test/fully_connected_layer_test.cpp:65
↓ 3 callersMethodget
(self, flatten_ids, length)
test/sok_perf_test/demo/model.py:114
↓ 3 callersMethodget_cost_per_lookup
(self, hotness_array: np.array, ev_sizes_array: np.array)
benchmarks/embedding_collection/hugectr/sharding/planner.py:352
↓ 3 callersMethodget_dimensions
HugeCTR/src/data_readers/multi_hot/async_data_reader.cpp:458
↓ 3 callersMethodget_dtype_size
(cls, tf_data_type)
sparse_operation_kit/sparse_operation_kit/dump_load.py:99
↓ 3 callersMethodget_embedding_variables
(self)
sparse_operation_kit/example/sok_train_hps_inference/sok_train.py:243
↓ 3 callersMethodget_file_size
HugeCTR/embedding_storage/weight_io/fs_interface.cpp:70
↓ 3 callersMethodget_file_stats
HugeCTR/include/data_readers/metadata.hpp:83
↓ 3 callersFunctionget_mean
test/utest/core23_layer_test/prelu_dice_layer_test.cpp:36
↓ 3 callersMethodget_metadata_status
HugeCTR/include/data_readers/metadata.hpp:98
↓ 3 callersMethodget_num_of_files
HugeCTR/include/data_readers/file_list.hpp:119
↓ 3 callersFunctionget_num_table
test/utest/embedding_collection/configuration.hpp:182
↓ 3 callersMethodget_num_values
HugeCTR/include/data_readers/csr23.hpp:149
↓ 3 callersMethodget_raw_metrics_all
HugeCTR/src/core23_network.cpp:111
↓ 3 callersMethodget_rowoffset_tensor
HugeCTR/include/sparse_tensor.hpp:116
↓ 3 callersMethodget_size_in_bytes
HugeCTR/include/tensor2.hpp:181
↓ 3 callersFunctionget_table_group_strategy
HugeCTR/include/embeddings/embedding_collection.hpp:66
↓ 3 callersFunctionget_table_name
HugeCTR/include/embeddings/embedding_collection.hpp:70
↓ 3 callersFunctionget_table_place_strategy
HugeCTR/include/embeddings/embedding_collection.hpp:64
↓ 3 callersFunctionget_tensor_shape
HugeCTR/src/pybind/model_compile.cpp:60
↓ 3 callersFunctionhugectr2onnx_ncf_test
( batch_size, num_batches, data_file, graph_config, dense_model, sparse_models, on
test/onnx_converter_test/hugectr2onnx_ncf_test.py:24
↓ 3 callersMethodinit_buf
HugeCTR/src/collectives/ib_proxy.cpp:357
↓ 3 callersMethodinit_params
HugeCTR/src/layer.cpp:30
↓ 3 callersFunctionisDynamicVariable
(variable)
sparse_operation_kit/sparse_operation_kit/lookup.py:167
↓ 3 callersFunctionisEmbeddingVariable
(variable)
sparse_operation_kit/sparse_operation_kit/lookup.py:171
↓ 3 callersMethodis_open
HugeCTR/include/data_readers/file_source.hpp:99
↓ 3 callersMethodlookup_id_in_group
HugeCTR/embedding/common.hpp:241
↓ 3 callersMethodnext_source
* Start a new file to read. mmap() parquet to memory * @return `Success`, `FileCannotOpen` or `UnspecificError` */ counter_ always points to next fi
HugeCTR/src/data_readers/file_source_parquet.cpp:112
↓ 3 callersMethodon_training_end
HugeCTR/include/pybind/training_callback.hpp:101
↓ 3 callersFunctionprint_logs
()
sparse_operation_kit/documents/tutorials/DLRM/utils.py:228
↓ 3 callersFunctionprint_logs
()
test/sok_perf_test/dlrm/utils.py:228
↓ 3 callersMethodread_from
HugeCTR/embedding_storage/weight_io/fs_interface.cpp:46
↓ 3 callersMethodrows
sparse_operation_kit/kit_src/variable/kernels/dummy_var.cc:49
↓ 3 callersMethodrun_graph
HugeCTR/src/pipeline.cpp:149
↓ 3 callersMethodscope
(self)
test/sok_perf_test/dlrm/utils.py:138
↓ 3 callersMethodsearch_algorithm
HugeCTR/src/core23_network.cpp:248
↓ 3 callersMethodset
HugeCTR/core23/shape.hpp:48
↓ 3 callersMethodset_block_size
HugeCTR/src/io/hadoop_filesystem.cpp:31
↓ 3 callersMethodset_current_batch_size
HugeCTR/include/metrics.hpp:59
↓ 3 callersMethodset_device
HugeCTR/core23/device_guard.cpp:33
↓ 3 callersMethodset_learning_rate
HugeCTR/include/pybind/model.hpp:436
↓ 3 callersMethodsummary
(self)
sparse_operation_kit/example/sok_train_hps_inference/hps_use_sok_weight_inference.py:146
↓ 3 callersFunctiontest_impl
test/utest/core23/device_test.cpp:29
↓ 3 callersMethodupdate_params
HugeCTR/src/core23_network.cpp:230
↓ 3 callersMethodwait_until_writeable
HugeCTR/src/data_readers/row_group_reading_thread.cpp:66
↓ 3 callersMethodworld_size
HugeCTR/core23/mpi_init_service.hpp:33
↓ 2 callersFunctionGetDeviceTypeName
HugeCTR/core23/device_type.cpp:32
↓ 2 callersFunctionGetWeightBufferChannel
HugeCTR/src/network_buffer_channels.cpp:52
↓ 2 callersFunctionSetNetworkBufferChannel
namespace
HugeCTR/src/network_buffer_channels.cpp:32
↓ 2 callersMethod_get_categorical_feature_type
(self, size)
sparse_operation_kit/documents/tutorials/DLRM_Benchmark/dataset.py:268
↓ 2 callersMethod_read_metadata
(self, f)
samples/dlrm/preprocessing/convert_to_raw.py:87
↓ 2 callersFunction_replica_loss
(labels, logits)
sparse_operation_kit/documents/tutorials/DLRM/main.py:169
↓ 2 callersFunction_replica_loss
(labels, logits)
test/sok_perf_test/dlrm/main.py:171
↓ 2 callersFunction_replica_loss
(labels, logits)
test/sok_perf_test/demo/run_sok.py:55
↓ 2 callersFunction_step
(samples, labels)
sparse_operation_kit/documents/tutorials/DLRM_Benchmark/trainer.py:15
↓ 2 callersFunction_step
(samples, labels)
sparse_operation_kit/SOK_DLRM_Benchmark/trainer.py:16
↓ 2 callersFunctionactivate_optimizer_state
(optimizer, load_vars)
sparse_operation_kit/sparse_operation_kit/dump_load.py:383
↓ 2 callersFunctionadd_dense_layer_impl
HugeCTR/src/pybind/add_dense_layer_helpers.cpp:122
↓ 2 callersMethodallocatable
HugeCTR/core23/buffer.hpp:51
↓ 2 callersMethodallocate
HugeCTR/core23/details/pool_cuda_allocator.cpp:67
↓ 2 callersMethodallocate
HugeCTR/include/general_buffer2.hpp:28
↓ 2 callersFunctionallreduce
(*args, **kwargs)
sparse_operation_kit/sparse_operation_kit/communication.py:235
↓ 2 callersMethodallreduce
HugeCTR/src/exchange_wgrad.cpp:49
↓ 2 callersMethodassign_add
(self, *args, **kwargs)
sparse_operation_kit/sparse_operation_kit/dynamic_variable.py:394
↓ 2 callersMethodbackward
(self, top_grad, grad_key, grad_key_offset, grad_ev, grad_ev_offset, grad_id_space_list, do_allreduce: bool)
notebooks/prototype_embedding_collection/main.py:39
↓ 2 callersMethodbackward
* The first stage of backward propagation of embedding layer, * which computes the wgrad by the dgrad from the top layer. */
HugeCTR/include/embeddings/localized_slot_sparse_embedding_hash.hpp:289
↓ 2 callersFunctionbinary_cross_entropy_loss
test/utest/loss/loss_test.cpp:127
↓ 2 callersMethodcache_ddl_output
HugeCTR/src/embeddings/embedding_collection.cpp:303
↓ 2 callersFunctioncal_workspace_size_per_gpu_from_vocabulary_size_per_gpu
( vocabulary_size_per_gpu, emb_vec_size, num_gpus, optimizer, optimizer_update_type )
tools/embedding_workspace_calculator/cal_vocabulary_size_per_gpu_and_workspace_size_per_gpu.py:20
↓ 2 callersMethodcheck_dense
test/utest/data_reader/split_batch_test.cpp:117
↓ 2 callersFunctioncheck_device
HugeCTR/src/pybind/model.cpp:39
↓ 2 callersMethodcheck_label
test/utest/data_reader/split_batch_test.cpp:105
↓ 2 callersMethodcheck_sparse
test/utest/data_reader/split_batch_test.cpp:135
↓ 2 callersFunctioncheck_table_name_correct
HugeCTR/include/embeddings/embedding_collection.hpp:216
↓ 2 callersFunctioncheck_weight_file_valid
(key_path, weight_path, optimizer_state_paths)
sparse_operation_kit/sparse_operation_kit/dump_load.py:775
↓ 2 callersFunctioncheck_zero_line
(arr)
sparse_operation_kit/sparse_operation_kit/test/function_test/tf2/lookup/lookup_sparse_hkv_low_frequency_test.py:113
↓ 2 callersFunctioncompute_strides
HugeCTR/core23/details/tensor_helpers.hpp:15
↓ 2 callersFunctionconvert_bytes_to_string_list
(bytes_list, data_length, signed=False)
sparse_operation_kit/sparse_operation_kit/dump_load.py:262
↓ 2 callersFunctionconvert_sok_weights_for_hps
(sok_file_path, dtype, dim, output_path)
sparse_operation_kit/example/sok_train_hps_inference/sok_train.py:331
↓ 2 callersMethodcopy
HugeCTR/src/io/s3_filesystem.cpp:252
← previousnext →301–400 of 3,072, ranked by callers