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
↓ 4 callersMethoddequantize_test_helper
(self, weight_type, quant_type)
fastertransformer/tests/gemm_dequantize/th_gemm_dequantize.py:22
↓ 4 callersMethoddequantize_test_helper
(self, weight_type, quant_type)
fastertransformer/tests/gemm_dequantize/single_test.py:23
↓ 4 callersMethoddequantize_test_helper
(self, weight_type, quant_type)
fastertransformer/tests/moe/th_moe_unit_tests.py:157
↓ 4 callersFunctiondevice_breakpoint
Triggers a breakpoint on the device
fastertransformer/3rdparty/cutlass/include/cutlass/arch/arch.h:89
↓ 4 callersFunctiondevice_mem_alloc
(size)
fastertransformer/3rdparty/cutlass/tools/library/scripts/pycutlass/src/pycutlass/memory_manager.py:64
↓ 4 callersFunctiondiv_up
fastertransformer/src/fastertransformer/kernels/llama/decoder_masked_groupedquery_attention/decoder_masked_groupedquery_attention_template.hpp:1038
↓ 4 callersMethoddownload_wikicorpus
(self, language)
fastertransformer/examples/pytorch/bert/bert-quantization-sparsity/data/Downloader.py:68
↓ 4 callersFunctiondropout
Perform dropout. Args: input_tensor: float Tensor. dropout_prob: Python float. The probability of dropping out a value (NOT of *keepi
fastertransformer/examples/tensorflow/bert/tensorflow_bert/bert/modeling.py:344
↓ 4 callersFunctiondropout
Perform dropout. Args: input_tensor: float Tensor. dropout_prob: Python float. The probability of dropping out a value (NOT of *keepi
fastertransformer/examples/tensorflow/bert/bert-quantization/modeling.py:367
↓ 4 callersMethodelapsed_msecs
engine/common/timer.h:47
↓ 4 callersMethodembed_text
Embed a batch of tokens using the model
fastertransformer/3rdparty/Megatron-LM/megatron/model/biencoder_model.py:135
↓ 4 callersMethodemit
(self)
fastertransformer/3rdparty/cutlass/tools/library/scripts/pycutlass/src/pycutlass/gemm_operation.py:538
↓ 4 callersMethodenabled
fastertransformer/src/fastertransformer/models/t5/T5AdapterWeight.h:39
↓ 4 callersMethodepilogue
fastertransformer/src/fastertransformer/kernels/abqgemm/mma_any/aq_wmma_kernel.h:489
↓ 4 callersFunctionfind_chars_or_comment
Return pointer to first char (of chars) or inline comment in given string, or pointer to null at end of string if neither found. Inline comment mus
fastertransformer/3rdparty/INIReader.h:137
↓ 4 callersFunctionfinetune
Main finetune function used across all tasks.
fastertransformer/3rdparty/Megatron-LM/tasks/vision/finetune_utils.py:241
↓ 4 callersFunctionfix_query_key_value_ordering
Fix up query/key/value matrix ordering if checkpoint version is smaller than 2.0
fastertransformer/3rdparty/Megatron-LM/megatron/checkpointing.py:268
↓ 4 callersMethodflops
(self)
fastertransformer/examples/pytorch/swin/Swin-Transformer-Quantization/SwinTransformer/models/swin_mlp.py:461
↓ 4 callersMethodforward
# Args. max_input_length: int, maximum input context length. step: int, the current step index. ite: int
fastertransformer/examples/pytorch/gpt/utils/gpt_decoder.py:656
↓ 4 callersMethodforward_features
(self, x)
fastertransformer/examples/pytorch/swin/Swin-Transformer-Quantization/SwinTransformer/models/swin_mlp.py:442
↓ 4 callersMethodfrom_config
(cls, config: GptInitModelParameters, **kwargs)
fastertransformer/examples/pytorch/gpt/utils/gpt_decoder.py:840
↓ 4 callersMethodfrom_json_file
Constructs a `BertConfig` from a json file of parameters.
fastertransformer/examples/tensorflow/bert/tensorflow_bert/bert/modeling.py:91
↓ 4 callersMethodfrom_opt
Alternate constructor. Args: model (onmt.modules.NMTModel): See :func:`__init__()`. fields (dict[str, torchtext.data.
fastertransformer/examples/pytorch/decoding/utils/translator.py:183
↓ 4 callersFunctionft_bert
Run the bert transformer layer by FasterTransformer. Args: inputs: A tf.Tensor with shape [batch_size, seq_len, hidden_dimension].
fastertransformer/examples/tensorflow/bert/utils/bert.py:344
↓ 4 callersFunctionft_decoding
Run the decoding with beam search by TensorFlow. Args: memory_tensor: A tf.tensor with shape [batch_size * beam_width, max(memor
fastertransformer/examples/tensorflow/decoding/utils/ft_decoding.py:43
↓ 4 callersFunctionfuse_qkv_weight
(q, k, v)
fastertransformer/examples/pytorch/gpt/utils/huggingface_opt_convert.py:149
↓ 4 callersFunctiongen_relative_pos_bias
fastertransformer/src/fastertransformer/th_op/swin/SwinOp.cc:138
↓ 4 callersFunctiongeneral_detokenize
(string)
algorithm/lm_eval/utils.py:62
↓ 4 callersFunctiongenerateWeightSize
fastertransformer/examples/cpp/swin/functions.h:73
↓ 4 callersFunctiongetCusparseOperation
fastertransformer/src/fastertransformer/utils/gemm.cc:1086
↓ 4 callersMethodgetDModel
fastertransformer/src/fastertransformer/models/t5/T5Encoder.h:157
↓ 4 callersMethodgetMaxSeqLen
fastertransformer/src/fastertransformer/models/multi_gpu_gpt/ParallelGptWeight.h:69
↓ 4 callersFunctiongetSetDevice
fastertransformer/src/fastertransformer/utils/cuda_utils.cc:333
↓ 4 callersFunctiongetTensorView
(tensor: np.ndarray, problem_size: cutlass.gemm.GemmCoord, operand: str, layout: str)
fastertransformer/3rdparty/cutlass/tools/library/scripts/pycutlass/src/pycutlass/test/gemm_testbed.py:99
↓ 4 callersFunctionget_checkpoint_version
()
fastertransformer/3rdparty/Megatron-LM/megatron/checkpointing.py:40
↓ 4 callersMethodget_checkpoints_paths
Injects tensor/pipeline model parallel ranks into the filepath. Does nothing if not using model parallelism.
fastertransformer/examples/pytorch/nemo.py:129
↓ 4 callersMethodget_cmd_line_arguments
fastertransformer/3rdparty/cutlass/tools/util/include/cutlass/util/command_line.h:178
↓ 4 callersFunctionget_forward_backward_func
()
fastertransformer/3rdparty/Megatron-LM/megatron/schedules.py:32
↓ 4 callersFunctionget_ft_quant_type
fastertransformer/src/fastertransformer/th_op/common/WeightOnlyQuantOps.cc:40
↓ 4 callersFunctionget_gpu_memory
()
algorithm/parallel_utils.py:56
↓ 4 callersFunctionget_group_world_size_rank
()
fastertransformer/3rdparty/Megatron-LM/pretrain_ict.py:49
↓ 4 callersFunctionget_indexed_dataset_
(data_prefix, data_impl, skip_warmup)
fastertransformer/3rdparty/Megatron-LM/megatron/data/dataset_utils.py:585
↓ 4 callersFunctionget_loaders
( name, nsamples=128, seed=0, seqlen=2048, model='', )
algorithm/datautils.py:169
↓ 4 callersFunctionget_ltor_masks_and_position_ids
Build masks and position id for left to right model.
fastertransformer/3rdparty/Megatron-LM/megatron/utils.py:147
↓ 4 callersFunctionget_shape_list
(tensor, expected_rank=None, name=None)
fastertransformer/examples/tensorflow/bert/utils/bert.py:276
↓ 4 callersMethodget_tensor_ref
(extent: 'tuple[int]', device_ptr: cuda.CUdeviceptr, layout: cutlass.layout)
fastertransformer/3rdparty/cutlass/tools/library/scripts/pycutlass/src/pycutlass/reduction_operation.py:107
↓ 4 callersMethodget_workspace_size
Gets the workspace size
fastertransformer/3rdparty/cutlass/examples/43_dual_gemm/device/dual_gemm.h:304
↓ 4 callersMethodget_zipfile
(path)
fastertransformer/examples/pytorch/swin/Swin-Transformer-Quantization/SwinTransformer/data/zipreader.py:31
↓ 4 callersMethodgood
Returns true if the TensorRef is non-null
fastertransformer/3rdparty/cutlass/include/cutlass/tensor_ref.h:262
↓ 4 callersFunctiongraph_fn_builder
(model_fn)
fastertransformer/examples/tensorflow/bert/tensorflow_bert/profile_transformer_inference.py:130
↓ 4 callersMethodhas_adapters
fastertransformer/src/fastertransformer/models/t5/T5Decoder.h:125
↓ 4 callersMethodinitialize
fastertransformer/src/fastertransformer/kernels/abqgemm/mma_any/aq_bmma_op.h:62
↓ 4 callersFunctionint_result_cross_check
(name, tf_result, op_result, shape)
fastertransformer/examples/tensorflow/decoder/utils/common.py:281
↓ 4 callersMethodinverse
fastertransformer/3rdparty/cutlass/include/cutlass/layout/tensor_op_multiplicand_sm80.h:142
↓ 4 callersFunctioninvokeAddBias
fastertransformer/src/fastertransformer/kernels/activation_kernels.h:160
↓ 4 callersFunctioninvokeGeneralAddResidualT5PreLayerNorm
fastertransformer/src/fastertransformer/kernels/layernorm_kernels.h:178
↓ 4 callersMethodisFuseBatchGemm
fastertransformer/src/fastertransformer/utils/cublasMMWrapper.cc:625
↓ 4 callersMethodis_complex
(self)
fastertransformer/3rdparty/cutlass/tools/library/scripts/rank_k_operation.py:42
↓ 4 callersMethodis_complex
(self)
fastertransformer/3rdparty/cutlass/tools/library/scripts/gemm_operation.py:40
↓ 4 callersMethodis_complex
(self)
fastertransformer/3rdparty/cutlass/tools/library/scripts/trmm_operation.py:41
↓ 4 callersMethodis_complex
(self)
fastertransformer/3rdparty/cutlass/tools/library/scripts/symm_operation.py:44
↓ 4 callersMethodis_complex
(self)
fastertransformer/3rdparty/cutlass/tools/library/scripts/rank_2k_operation.py:44
↓ 4 callersFunctionis_non_str_iterable
(obj)
algorithm/lm_eval/metrics.py:157
↓ 4 callersFunctionisnan
fastertransformer/3rdparty/cutlass/include/cutlass/half.h:510
↓ 4 callersFunctionkl_loss
(x, y)
fastertransformer/examples/tensorflow/bert/bert-quantization/run_squad.py:522
↓ 4 callersMethodlast_problem
(cls, doc)
algorithm/lm_eval/tasks/race.py:102
↓ 4 callersFunctionload
Loads a fragment from memory at the location pointed to by the iterator.
fastertransformer/3rdparty/cutlass/include/cutlass/gemm/warp/mma_tensor_op_tile_iterator.h:378
↓ 4 callersMethodload
Loads a fragment from memory at the location pointed to by the iterator.
fastertransformer/3rdparty/cutlass/include/cutlass/gemm/warp/mma_complex_tensor_op_tile_iterator_sm80.h:969
↓ 4 callersMethodload
Loads a fragment from memory at the location pointed to by the iterator.
fastertransformer/3rdparty/cutlass/include/cutlass/gemm/warp/mma_tensor_op_tile_iterator_sm70.h:1312
↓ 4 callersMethodload
Loads a fragment from memory at the location pointed to by the iterator.
fastertransformer/3rdparty/cutlass/include/cutlass/gemm/warp/mma_tensor_op_tile_iterator.h:2527
↓ 4 callersFunctionloadMatrixSync
engine/mma_any/bmma.h:160
↓ 4 callersFunctionloadMatrixSync
algorithm/ABQ/csrc/mma_any/bmma.h:111
↓ 4 callersFunctionloadMatrixSync
fastertransformer/src/fastertransformer/kernels/abqgemm/mma_any/bmma.h:97
↓ 4 callersMethodload_from
(self, weights)
fastertransformer/examples/pytorch/vit/ViT-quantization/vit_int8.py:420
↓ 4 callersMethodload_state_dict
Customized load.
fastertransformer/3rdparty/Megatron-LM/megatron/model/language_model.py:224
↓ 4 callersFunctionload_with_byte_offset
Loads a fragment from memory with additional logical offset
fastertransformer/3rdparty/cutlass/include/cutlass/gemm/warp/mma_tensor_op_tile_iterator_sparse.h:283
↓ 4 callersFunctionload_with_byte_offset
Loads a fragment from memory
fastertransformer/3rdparty/cutlass/include/cutlass/epilogue/threadblock/predicated_tile_iterator.h:303
↓ 4 callersFunctionload_with_pointer_offset
Loads a fragment from memory
fastertransformer/3rdparty/cutlass/include/cutlass/conv/threadblock/predicated_scale_bias_vector_iterator.h:185
↓ 4 callersFunctionload_with_pointer_offset
Loads a fragment from memory
fastertransformer/3rdparty/cutlass/include/cutlass/epilogue/threadblock/shared_load_iterator.h:163
↓ 4 callersMethodload_with_pointer_offset
Loads a fragment from memory with additional logical offset
fastertransformer/3rdparty/cutlass/include/cutlass/gemm/warp/mma_complex_tensor_op_tile_iterator_sm80.h:975
↓ 4 callersMethodload_with_pointer_offset
Loads a fragment from memory with additional logical offset
fastertransformer/3rdparty/cutlass/include/cutlass/gemm/warp/mma_tensor_op_tile_iterator_sm70.h:1318
↓ 4 callersMethodload_with_pointer_offset
Loads a fragment from memory with additional logical offset
fastertransformer/3rdparty/cutlass/include/cutlass/gemm/warp/mma_tensor_op_tile_iterator.h:2533
↓ 4 callersMethodload_with_pointer_offset
Loads a fragment from memory
fastertransformer/3rdparty/cutlass/include/cutlass/transform/threadblock/regular_tile_iterator_tensor_op.h:315
↓ 4 callersFunctionlog2_ceil
fastertransformer/3rdparty/Megatron-LM/megatron/fused_kernels/scaled_masked_softmax.h:50
↓ 4 callersFunctionnormalize_answer
Lower text and remove punctuation, articles and extra whitespace.
fastertransformer/examples/tensorflow/bert/tensorflow_bert/squad_evaluate-v1.1.py:25
↓ 4 callersFunctionnormalize_answer
Lower text and remove punctuation, articles and extra whitespace.
fastertransformer/examples/tensorflow/bert/tensorflow_bert/squad_evaluate_v1_1.py:25
↓ 4 callersFunctionnp2th
(weights, is_conv=False)
fastertransformer/examples/pytorch/vit/VisionTransformerWeightLoader.py:32
↓ 4 callersMethodnum_bytes
fastertransformer/examples/cpp/xlnet/cnpy.h:61
↓ 4 callersMethodnumpy_type
(type)
fastertransformer/3rdparty/cutlass/tools/library/scripts/pycutlass/src/pycutlass/test/gemm_grouped_testbed.py:76
↓ 4 callersMethodnumpy_type
(type)
fastertransformer/3rdparty/cutlass/tools/library/scripts/pycutlass/src/pycutlass/test/conv2d_testbed.py:193
↓ 4 callersFunctionpost_attention
Post-attention processing.
fastertransformer/examples/tensorflow/xlnet/modeling.py:106
↓ 4 callersFunctionprintMatrix
fastertransformer/tests/unittests/fp8_gemm_test/worker.cpp:191
↓ 4 callersFunctionprint_fraction
(a, b, counters, desc)
fastertransformer/examples/pytorch/vit/ViT-quantization/quant_utils.py:243
↓ 4 callersFunctionprint_fraction
(a, b, counters, desc)
fastertransformer/examples/pytorch/bert/bert-quantization-sparsity/quant_utils.py:243
↓ 4 callersFunctionprint_fraction
(a, b, counters, desc)
fastertransformer/examples/pytorch/swin/Swin-Transformer-Quantization/quant_utils.py:212
↓ 4 callersMethodproduct
Returns the product of all elements
fastertransformer/3rdparty/cutlass/include/cutlass/coord.h:311
↓ 4 callersMethodquant_x1
(self, x1)
algorithm/quantize/int_matmul.py:45
↓ 4 callersFunctionread_longs
(f, n)
fastertransformer/3rdparty/Megatron-LM/megatron/data/indexed_dataset.py:82
↓ 4 callersFunctionread_squad_examples
Read a SQuAD json file into a list of SquadExample.
fastertransformer/examples/tensorflow/bert/bert-quantization/utils/create_squad_data.py:152
↓ 4 callersFunctionread_start_ids
fastertransformer/examples/cpp/multi_gpu_gpt/gpt_example_utils.cc:28
← previousnext →901–1,000 of 14,558, ranked by callers