MCPcopy Create free account

hub / github.com/bytedance/ABQ-LLM / functions

Functions14,558 in github.com/bytedance/ABQ-LLM

⨍ Functions 14,558◇ Types & classes 4,979
↓ 3 callersFunctionbasic_moe_fc
(activations, expert_for_row, weights, scales, biases)
fastertransformer/tests/moe/th_moe_unit_tests.py:23
↓ 3 callersFunctionbatch_igemm_config
fastertransformer/src/fastertransformer/utils/gemm_test/encoder_igemm_func.cc:982
↓ 3 callersFunctionbias_dropout_add
(x, bias, residual, prob, training)
fastertransformer/3rdparty/Megatron-LM/megatron/model/transformer.py:371
↓ 3 callersFunctionbiencoder_model_provider
Build the model.
fastertransformer/3rdparty/Megatron-LM/megatron/model/biencoder_model.py:36
↓ 3 callersFunctionbroadcast_from_last_to_first_pipeline_stage
Broadcast tensor values from last stage into the first stage.
fastertransformer/3rdparty/Megatron-LM/megatron/text_generation/communication.py:96
↓ 3 callersFunctionbroadcast_tensor
Given size and type of a tensor on all ranks and the tensor value only on a specific rank, broadcast from that rank to all other ranks.
fastertransformer/3rdparty/Megatron-LM/megatron/text_generation/communication.py:157
↓ 3 callersFunctionbuild_dataset
(index, name)
fastertransformer/3rdparty/Megatron-LM/megatron/data/dataset_utils.py:517
↓ 3 callersFunctionbuild_dataset
(index, name)
fastertransformer/3rdparty/Megatron-LM/megatron/data/gpt_dataset.py:105
↓ 3 callersFunctionbuild_dataset
(is_train, config)
fastertransformer/examples/pytorch/vit/ViT-quantization/data.py:111
↓ 3 callersFunctionbuild_pretraining_data_loader
Buld dataloader given an input dataset.
fastertransformer/3rdparty/Megatron-LM/megatron/data/data_samplers.py:27
↓ 3 callersFunctionbuild_tokens_types_paddings_from_ids
Build token types and paddings, trim if needed, and pad if needed.
fastertransformer/3rdparty/Megatron-LM/tasks/orqa/supervised/data.py:70
↓ 3 callersFunctionbuild_train_valid_test_datasets
(data_prefix, data_impl, splits_string, train_valid_test_num_samples,
fastertransformer/3rdparty/Megatron-LM/megatron/data/dataset_utils.py:416
↓ 3 callersFunctioncalc_params_l2_norm
Calculate l2 norm of parameters
fastertransformer/3rdparty/Megatron-LM/megatron/utils.py:49
↓ 3 callersMethodcan_implement
Determines whether kernel satisfies alignment
fastertransformer/src/fastertransformer/cutlass_extensions/include/cutlass_extensions/gemm/kernel/fpA_intB_gemm.h:228
↓ 3 callersFunctioncast_to_float
fastertransformer/src/fastertransformer/kernels/decoder_masked_multihead_attention/decoder_masked_multihead_attention_template.hpp:877
↓ 3 callersFunctioncast_to_float
fastertransformer/src/fastertransformer/kernels/llama/decoder_masked_groupedquery_attention/decoder_masked_groupedquery_attention_template.hpp:877
↓ 3 callersFunctioncheck
(group, world_size, rank)
fastertransformer/3rdparty/Megatron-LM/megatron/mpu/tests/test_initialize.py:36
↓ 3 callersFunctioncheck_adlr_autoresume_termination
Check for autoresume signal and exit if it is received.
fastertransformer/3rdparty/Megatron-LM/megatron/utils.py:128
↓ 3 callersFunctioncheck_and_append_tensor_for_gather
(group, rank, world_size, input_)
fastertransformer/3rdparty/Megatron-LM/tasks/orqa/supervised/finetune.py:37
↓ 3 callersFunctioncheck_and_clean_text
(args, words, ngrams, text, start_position, \ text_buf_ngram_free, text_buf, local_ngram)
fastertransformer/3rdparty/Megatron-LM/tools/openwebtext/filter_ngrams.py:64
↓ 3 callersMethodcheck_output
(self, result)
fastertransformer/examples/tensorflow/bert/tensorflow_bert/bert/modeling_test.py:114
↓ 3 callersMethodclean
Clean a string (e.g. a training set) by removing all ngrams previously registered as contaminants. Returns a list of clean chunks, or empty if
algorithm/lm_eval/decontamination/janitor.py:159
↓ 3 callersMethodclear
Clear the embedding data structures to save memory. The metadata ends up getting used, and is also much smaller in dimensiona
fastertransformer/3rdparty/Megatron-LM/megatron/data/realm_index.py:42
↓ 3 callersMethodcls
(self)
fastertransformer/3rdparty/Megatron-LM/megatron/tokenizer/tokenizer.py:104
↓ 3 callersMethodcommit
(self)
algorithm/lm_eval/decontamination/archiver.py:39
↓ 3 callersFunctionconv1x1
(cin, cout, stride=1, bias=False)
fastertransformer/examples/pytorch/vit/ViT-quantization/ViT-pytorch/models/modeling_resnet.py:50
↓ 3 callersFunctionconv2d_fixed_channel_problemsizes
(channels)
fastertransformer/3rdparty/cutlass/tools/library/scripts/pycutlass/test/conv/conv2d_fprop_fixed_channels_f16nhwc_f16nhwc_f16nhwc_tensor_op_f32_sm80.py:6
↓ 3 callersMethodconvert_choice
(choice)
algorithm/lm_eval/tasks/superglue.py:218
↓ 3 callersFunctionconvert_examples_to_features
Loads a data file into a list of `InputBatch`s.
fastertransformer/examples/tensorflow/bert/bert-quantization/extract_features.py:210
↓ 3 callersFunctionconvert_examples_to_features
Loads a data file into a list of `InputBatch`s.
fastertransformer/examples/pytorch/bert/bert-quantization-sparsity/run_swag.py:139
↓ 3 callersFunctionconvert_parameter_name
(name: str)
fastertransformer/examples/pytorch/gpt/utils/huggingface_bloom_convert.py:130
↓ 3 callersFunctionconvert_to_unicode
Converts `text` to Unicode (if it's not already), assuming utf-8 input.
fastertransformer/3rdparty/Megatron-LM/megatron/tokenizer/bert_tokenization.py:78
↓ 3 callersFunctionconvert_to_unicode
Converts `text` to Unicode (if it's not already), assuming utf-8 input.
fastertransformer/examples/tensorflow/bert/tensorflow_bert/bert/tokenization.py:78
↓ 3 callersFunctionconvert_to_unicode
Converts `text` to Unicode (if it's not already), assuming utf-8 input.
fastertransformer/examples/tensorflow/bert/bert-quantization/tokenization.py:92
↓ 3 callersFunctioncreate_model
Creates a classification model.
fastertransformer/examples/tensorflow/bert/bert-quantization/run_squad.py:308
↓ 3 callersMethodcuLaunchKernel
fastertransformer/3rdparty/common/cudaDriverWrapper.cpp:132
↓ 3 callersMethodcuda
(self, device=None)
fastertransformer/examples/pytorch/gpt/utils/gpt_decoder.py:948
↓ 3 callersMethodcuda
(self)
fastertransformer/examples/pytorch/llama2/llamav2.py:423
↓ 3 callersMethodcuda
(self)
fastertransformer/examples/pytorch/baichuan2/baichuanv2.py:419
↓ 3 callersFunctioncustomMatmulRun
fastertransformer/src/fastertransformer/utils/gemm_test/gemm_func.cc:156
↓ 3 callersFunctioncyclic_iter
(iter)
fastertransformer/3rdparty/Megatron-LM/megatron/training.py:843
↓ 3 callersMethoddecode
(self, token_ids)
fastertransformer/examples/pytorch/gpt/utils/tokenizer.py:67
↓ 3 callersFunctiondecode_output
(outputs, model_dir, input_ids, actual_seq_len)
fastertransformer/examples/pytorch/longformer/longformer_qa.py:96
↓ 3 callersFunctiondetach
(tensor)
fastertransformer/3rdparty/Megatron-LM/megatron/data/realm_index.py:13
↓ 3 callersMethoddetokenize
(self, text)
algorithm/lm_eval/tasks/mutual.py:70
↓ 3 callersMethoddevice
(self)
algorithm/lm_eval/models/gpt2.py:103
↓ 3 callersMethoddownload
(self)
fastertransformer/examples/pytorch/bert/bert-quantization-sparsity/data/Downloader.py:28
↓ 3 callersMethoddownload_glue
(self, task_name)
fastertransformer/examples/pytorch/bert/bert-quantization-sparsity/data/Downloader.py:80
↓ 3 callersMethoddtype
(self)
fastertransformer/3rdparty/Megatron-LM/megatron/data/indexed_dataset.py:427
↓ 3 callersMethodelapsed
Calculate the elapsed time.
fastertransformer/3rdparty/Megatron-LM/megatron/global_vars.py:224
↓ 3 callersMethodelapsed_time_in_sec
(self, tag='__default')
fastertransformer/examples/pytorch/gpt/bloom_lambada.py:108
↓ 3 callersMethodenable
Sets all predicates to true
fastertransformer/3rdparty/cutlass/include/cutlass/predicate_vector.h:444
↓ 3 callersMethodenable
< CUTLASS_HOST_DEVICE enables all accesses guarded by mask
fastertransformer/3rdparty/cutlass/include/cutlass/epilogue/threadblock/predicated_tile_iterator_v2.h:156
↓ 3 callersMethodenable
< CUTLASS_HOST_DEVICE enables all accesses guarded by mask
fastertransformer/3rdparty/cutlass/include/cutlass/epilogue/threadblock/predicated_tile_iterator.h:161
↓ 3 callersMethodencode
(self, text)
fastertransformer/3rdparty/Megatron-LM/megatron/tokenizer/gpt2_tokenization.py:280
↓ 3 callersMethodencode
fastertransformer/examples/pytorch/gpt/utils/tokenizer.py:56
↓ 3 callersFunctionestimate_best_config_from_occupancies
fastertransformer/src/fastertransformer/kernels/cutlass_kernels/cutlass_heuristic.cc:128
↓ 3 callersMethodexecute
fastertransformer/3rdparty/cutlass/include/cutlass/transform/thread/unary_op.h:54
↓ 3 callersFunctionexpand_dims
(x, v, axis)
fastertransformer/examples/tensorflow/bert/bert-quantization/ft-tensorflow-quantization/ft_tensorflow_quantization/python/utils/utils.py:23
↓ 3 callersFunctionfast_min
fastertransformer/3rdparty/cutlass/include/cutlass/fast_math.h:600
↓ 3 callersFunctionfile_based_convert_examples_to_features
Convert a set of `InputExample`s to a TFRecord file.
fastertransformer/examples/tensorflow/bert/tensorflow_bert/bert/run_classifier.py:479
↓ 3 callersFunctionfile_based_convert_examples_to_features
Convert a set of `InputExample`s to a TFRecord file.
fastertransformer/examples/tensorflow/bert/bert-quantization/utils/create_glue_data.py:450
↓ 3 callersFunctionfile_based_input_fn_builder
Creates an `input_fn` closure to be passed to TPUEstimator.
fastertransformer/examples/tensorflow/bert/tensorflow_bert/bert/run_classifier.py:509
↓ 3 callersFunctionfloat_from_int8
fastertransformer/src/fastertransformer/kernels/decoder_masked_multihead_attention/decoder_masked_multihead_attention_template.hpp:943
↓ 3 callersFunctionfloat_from_int8
fastertransformer/src/fastertransformer/kernels/llama/decoder_masked_groupedquery_attention/decoder_masked_groupedquery_attention_template.hpp:943
↓ 3 callersFunctionfloat_to_int8_rn_host
be consistent with FasterTransformer
fastertransformer/src/fastertransformer/utils/convert_data_type.h:22
↓ 3 callersMethodflops
(self)
fastertransformer/examples/pytorch/swin/Swin-Transformer-Quantization/SwinTransformer/models/swin_transformer_moe.py:590
↓ 3 callersMethodflops
(self)
fastertransformer/examples/pytorch/swin/Swin-Transformer-Quantization/SwinTransformer/models/swin_transformer.py:428
↓ 3 callersMethodflops
(self)
fastertransformer/examples/pytorch/swin/Swin-Transformer-Quantization/SwinTransformer/models/swin_transformer_v2.py:439
↓ 3 callersMethodflops
(self)
fastertransformer/examples/pytorch/swin/Swin-Transformer-Quantization/models/swin_transformer_v2.py:498
↓ 3 callersMethodformat_answer
(answer, label)
algorithm/lm_eval/tasks/superglue.py:251
↓ 3 callersMethodforward
fastertransformer/src/fastertransformer/th_op/vit/ViTINT8Op.cc:104
↓ 3 callersMethodforward
fastertransformer/src/fastertransformer/th_op/longformer/LongformerEncoderOp.cc:58
↓ 3 callersMethodforward
fastertransformer/src/fastertransformer/th_op/swin/SwinOp.cc:128
↓ 3 callersMethodforward
fastertransformer/src/fastertransformer/th_op/swin/SwinINT8Op.cc:131
↓ 3 callersMethodfrom_json_file
Constructs a `BertConfig` from a json file of parameters.
fastertransformer/examples/tensorflow/bert/bert-quantization/modeling.py:96
↓ 3 callersFunctionfuse3
(qq, qk, qv)
fastertransformer/examples/pytorch/vit/ViT-quantization/quant_utils.py:194
↓ 3 callersFunctionfuse3
(qq, qk, qv)
fastertransformer/examples/pytorch/bert/bert-quantization-sparsity/quant_utils.py:194
↓ 3 callersFunctionfuse3
(qq, qk, qv)
fastertransformer/examples/pytorch/swin/Swin-Transformer-Quantization/quant_utils.py:165
↓ 3 callersMethodfuse3
(self, qq, qk, qv)
fastertransformer/examples/tensorflow/bert/bert-quantization/run_squad.py:286
↓ 3 callersMethodgating_softmax_test_helper
(self, dtype, rtol=1e-05, atol=1e-08)
fastertransformer/tests/moe/th_moe_unit_tests.py:58
↓ 3 callersMethodgemm_bias_act
fastertransformer/src/fastertransformer/kernels/cutlass_kernels/fpA_intB_gemm/fpA_intB_gemm_template.h:511
↓ 3 callersFunctiongenerate_and_post_process
Run inference and post-process outputs, i.e., detokenize, move to cpu and convert to list.
fastertransformer/3rdparty/Megatron-LM/megatron/text_generation/api.py:30
↓ 3 callersMethodget
Returns a pointer
fastertransformer/3rdparty/cutlass/include/cutlass/transform/threadblock/regular_tile_access_iterator_tensor_op.h:690
↓ 3 callersFunctiongetCublasComputeType
fastertransformer/src/fastertransformer/utils/gemm.cc:1037
↓ 3 callersMethodgetFinishBuffer
fastertransformer/src/fastertransformer/models/gptj/GptJ.cc:1237
↓ 3 callersFunctiongetGemmOpString
fastertransformer/src/fastertransformer/utils/gemm.cc:1074
↓ 3 callersFunctiongetParams
fastertransformer/3rdparty/cutlass/include/cutlass/conv/threadblock/conv2d_fprop_activation_tile_access_iterator_analytic.h:177
↓ 3 callersMethodgetSpAlgo
fastertransformer/src/fastertransformer/utils/cublasAlgoMap.cc:204
↓ 3 callersMethodgetWeightsPtr
fastertransformer/src/fastertransformer/models/longformer/LongformerEncoder.cc:134
↓ 3 callersMethodgetXMMAKernels
fastertransformer/3rdparty/trt_fused_multihead_attention/fused_multihead_attention.h:411
↓ 3 callersFunctiongetXPos
(item)
fastertransformer/3rdparty/cutlass/docs/search/search.js:24
↓ 3 callersFunctiongetYPos
(item)
fastertransformer/3rdparty/cutlass/docs/search/search.js:38
↓ 3 callersMethodget_answer_option
(cls, problem)
algorithm/lm_eval/tasks/race.py:97
↓ 3 callersMethodget_batch_idx
Gets the batch index
fastertransformer/3rdparty/cutlass/include/cutlass/gemm/threadblock/threadblock_swizzle.h:312
↓ 3 callersFunctionget_candidate_configs
fastertransformer/src/fastertransformer/kernels/cutlass_kernels/cutlass_heuristic.cc:110
↓ 3 callersMethodget_cmd_line_argument_ranges
* Returns a list of ranges specified for a given commandline parameter * --<flag>=<key:value>,<key:value>* */
fastertransformer/3rdparty/cutlass/tools/util/include/cutlass/util/command_line.h:217
↓ 3 callersMethodget_dev_examples
Gets a collection of `InputExample`s for the dev set.
fastertransformer/examples/pytorch/bert/bert-quantization-sparsity/processors/glue.py:61
↓ 3 callersMethodget_labels
Gets the list of labels for this data set.
fastertransformer/examples/tensorflow/bert/tensorflow_bert/bert/run_classifier.py:192
↓ 3 callersMethodget_loss_scale
FP32 optimizer does not do any scaling.
fastertransformer/3rdparty/Megatron-LM/megatron/optimizer/optimizer.py:509
← previousnext →1,101–1,200 of 14,558, ranked by callers