Code
Hub
Workspaces
Following
Trending
Connect
MCP
copy
Create free account
hub
/
github.com/bytedance/ABQ-LLM
/ functions
Functions
14,558 in github.com/bytedance/ABQ-LLM
⨍
Functions
14,558
◇
Types & classes
4,979
↓ 4 callers
Function
rel_attn_core
Core relative positional attention operations.
fastertransformer/examples/tensorflow/xlnet/modeling.py:141
↓ 4 callers
Method
reset
Clears the HostTensor allocation to size/capacity = 0
fastertransformer/3rdparty/cutlass/tools/util/include/cutlass/util/host_tensor_planar_complex.h:155
↓ 4 callers
Function
reshape_to_matrix
Reshapes a >= rank 2 tensor to a rank 2 tensor (i.e., a matrix).
fastertransformer/examples/tensorflow/bert/tensorflow_bert/my_modeling.py:942
↓ 4 callers
Function
round_up
Returns the smallest value in the half-open range [a, a+b) that is a multiple of b
fastertransformer/3rdparty/cutlass/include/cutlass/fast_math.h:175
↓ 4 callers
Function
rstrip
Strip whitespace chars off end of given string, in place. Return s. */
fastertransformer/3rdparty/INIReader.h:118
↓ 4 callers
Method
run
Executes one test
fastertransformer/3rdparty/cutlass/test/unit/conv/device/conv2d_with_reduction_testbed.h:209
↓ 4 callers
Method
run
Executes one test
fastertransformer/3rdparty/cutlass/test/unit/conv/device/conv2d_with_broadcast_testbed.h:280
↓ 4 callers
Function
run_torch
(model, images, mark)
fastertransformer/examples/pytorch/swin/infer_swintransformer_op.py:173
↓ 4 callers
Function
save_checkpoint_and_time
(iteration, model, optimizer, lr_scheduler)
fastertransformer/3rdparty/Megatron-LM/megatron/training.py:649
↓ 4 callers
Method
serialize
(self)
fastertransformer/examples/pytorch/bert/bert-quantization-sparsity/data/GooglePretrainedWeightDownloader.py:145
↓ 4 callers
Method
setT5StructureDiff
fastertransformer/src/fastertransformer/models/t5/T5EncoderWeight.cc:392
↓ 4 callers
Method
set_default_quant_desc_input
Args: value: An instance of :func:`QuantDescriptor <quantization.QuantDescriptor>`
fastertransformer/examples/tensorflow/bert/bert-quantization/ft-tensorflow-quantization/ft_tensorflow_quantization/python/layers/utils.py:32
↓ 4 callers
Function
set_iteration_index
Overrides the internal iteration index
fastertransformer/3rdparty/cutlass/include/cutlass/transform/threadblock/predicated_tile_access_iterator.h:474
↓ 4 callers
Method
set_k_partition
Functionally required for serial reduction in the epilogue
fastertransformer/3rdparty/cutlass/examples/43_dual_gemm/thread/left_silu_and_mul.h:104
↓ 4 callers
Method
set_predicates
fastertransformer/3rdparty/cutlass/include/cutlass/transform/threadblock/predicated_tile_access_iterator.h:185
↓ 4 callers
Method
set_states
Set the rng states. For efficiency purposes, we do not check the size of seed for compatibility.
fastertransformer/3rdparty/Megatron-LM/megatron/mpu/random.py:208
↓ 4 callers
Function
sigmoid_op
(x: np.ndarray)
fastertransformer/3rdparty/cutlass/tools/library/scripts/pycutlass/src/pycutlass/epilogue.py:385
↓ 4 callers
Function
smooth_fc_fc_temporary
(fc1, fc2, scales,shifts=None, self_attn=None)
algorithm/models/transformation.py:44
↓ 4 callers
Function
smooth_q_k_temporary
(q_proj, k_proj, scales, self_attn=None)
algorithm/models/transformation.py:69
↓ 4 callers
Method
split_zip_style_path
(path)
fastertransformer/examples/pytorch/swin/Swin-Transformer-Quantization/SwinTransformer/data/zipreader.py:39
↓ 4 callers
Method
state_dict_for_save_checkpoint
For easy load.
fastertransformer/3rdparty/Megatron-LM/megatron/model/language_model.py:207
↓ 4 callers
Method
stop
engine/common/timer.h:41
↓ 4 callers
Function
store
Stores a fragment to memory
fastertransformer/3rdparty/cutlass/include/cutlass/gemm/warp/mma_tensor_op_tile_iterator.h:3809
↓ 4 callers
Function
store_with_byte_offset
Stores a fragment to memory
fastertransformer/3rdparty/cutlass/include/cutlass/epilogue/threadblock/predicated_tile_iterator.h:378
↓ 4 callers
Method
store_with_pointer_offset
Store a fragment to memory
fastertransformer/3rdparty/cutlass/include/cutlass/transform/threadblock/regular_tile_iterator_tensor_op.h:327
↓ 4 callers
Function
summary
()
fastertransformer/examples/pytorch/gpt/utils/profiler.py:54
↓ 4 callers
Method
swapped_matrices
Returns arguments for the transposed matrices
fastertransformer/3rdparty/cutlass/include/cutlass/gemm/kernel/trmm_universal.h:175
↓ 4 callers
Function
tile
fastertransformer/tests/unittests/gtest_utils.h:122
↓ 4 callers
Method
tile_count
fastertransformer/3rdparty/cutlass/include/cutlass/gemm/kernel/gemm_grouped_problem_visitor.h:65
↓ 4 callers
Method
to_bfloat16
(self)
fastertransformer/examples/pytorch/bert/utils/encoder.py:216
↓ 4 callers
Method
to_bfloat16
(self)
fastertransformer/examples/pytorch/t5/utils/ft_decoding.py:421
↓ 4 callers
Method
to_bfloat16
(self)
fastertransformer/examples/pytorch/t5/utils/ft_encoder.py:355
↓ 4 callers
Method
to_cuda
(self)
fastertransformer/examples/pytorch/bert/utils/encoder.py:196
↓ 4 callers
Method
to_dict
Serializes this instance to a Python dictionary.
fastertransformer/examples/pytorch/bert/bert-quantization-sparsity/modeling.py:295
↓ 4 callers
Method
to_float
(self)
fastertransformer/examples/pytorch/t5/utils/ft_encoder.py:343
↓ 4 callers
Method
to_half
(self)
fastertransformer/examples/pytorch/bert/utils/encoder.py:210
↓ 4 callers
Function
train_loop
(args, model, optimizer, step, num_steps)
fastertransformer/examples/pytorch/bert/bert-quantization-sparsity/apex_sparsity/test/toy_problem.py:31
↓ 4 callers
Function
truncate_number
(number, threshold=1e-3)
algorithm/models/models_utils.py:26
↓ 4 callers
Function
update_num_microbatches
(consumed_samples, consistency_check=True)
fastertransformer/3rdparty/Megatron-LM/megatron/global_vars.py:52
↓ 4 callers
Function
vec_from_smem_transpose
fastertransformer/src/fastertransformer/kernels/decoder_masked_multihead_attention_utils.h:1598
↓ 4 callers
Function
weight_quantize
fastertransformer/src/fastertransformer/th_op/bert/WeightQuantizeOp.cc:44
↓ 4 callers
Function
whitespace_tokenize
Runs basic whitespace cleaning and splitting on a piece of text.
fastertransformer/examples/pytorch/bert/bert-quantization-sparsity/tokenization.py:86
↓ 4 callers
Function
write_longs
(f, a)
fastertransformer/3rdparty/Megatron-LM/megatron/data/indexed_dataset.py:88
↓ 4 callers
Function
write_smem_transpose
fastertransformer/src/fastertransformer/kernels/decoder_masked_multihead_attention_utils.h:1741
↓ 4 callers
Method
writer
(cls, path, dtype)
fastertransformer/3rdparty/Megatron-LM/megatron/data/indexed_dataset.py:340
↓ 3 callers
Function
F
(K,J)
fastertransformer/3rdparty/cutlass/docs/jquery.js:68
↓ 3 callers
Function
GemmComplex
fastertransformer/3rdparty/cutlass/tools/util/include/cutlass/util/reference/host/gemm_complex.h:71
↓ 3 callers
Method
LlamaOp
fastertransformer/src/fastertransformer/th_op/llama/LlamaOp.cc:23
↓ 3 callers
Function
M
()
fastertransformer/3rdparty/cutlass/docs/jquery.js:68
↓ 3 callers
Function
S
(bv)
fastertransformer/3rdparty/cutlass/docs/jquery.js:16
↓ 3 callers
Function
TensorFillRandomSparseMeta
< Layout function
fastertransformer/3rdparty/cutlass/tools/util/include/cutlass/util/reference/host/tensor_fill.h:1359
↓ 3 callers
Function
TensorNormDiff
fastertransformer/3rdparty/cutlass/tools/util/include/cutlass/util/reference/host/tensor_reduce.h:188
↓ 3 callers
Function
TensorTransformReduce
fastertransformer/3rdparty/cutlass/tools/util/include/cutlass/util/reference/host/tensor_reduce.h:57
↓ 3 callers
Function
TensorTransformReduce
fastertransformer/3rdparty/cutlass/tools/util/include/cutlass/util/reference/device/tensor_reduce.h:205
↓ 3 callers
Method
__init__
(self, init_method, output_layer_init_method)
fastertransformer/3rdparty/Megatron-LM/megatron/model/transformer.py:53
↓ 3 callers
Method
__init__
(self)
fastertransformer/examples/tensorflow/decoder/utils/common.py:298
↓ 3 callers
Method
__init__
Constructs a InputExample. Args: guid: Unique id for the example. text_a: string. The untokenized text of the first sequen
fastertransformer/examples/tensorflow/xlnet/convertInput.py:73
↓ 3 callers
Method
__init__
(self, optimizer, last_epoch=-1)
fastertransformer/examples/pytorch/vit/ViT-quantization/ViT-pytorch/utils/scheduler.py:11
↓ 3 callers
Function
__nv_fp8_e4m3
fastertransformer/tests/unittests/fp8_gemm_test/2022_03_21__fp8_stride_batch_example/include/cuda_fp8.hpp:706
↓ 3 callers
Method
_collate_data
(self, set)
algorithm/lm_eval/tasks/race.py:54
↓ 3 callers
Function
_copy_tokenizer_file_if_defined
(key_name, tokenizer_file_path, saved_dir)
fastertransformer/examples/pytorch/gpt/utils/nemo_ckpt_convert.py:469
↓ 3 callers
Function
_copy_tokenizer_file_if_defined
(key_name, tokenizer_file_path, saved_dir)
fastertransformer/examples/pytorch/t5/utils/nemo_t5_ckpt_convert.py:1165
↓ 3 callers
Method
_create_examples
Creates examples for the training and dev sets.
fastertransformer/examples/tensorflow/bert/tensorflow_bert/bert/run_classifier.py:278
↓ 3 callers
Method
_create_examples
Creates examples for the training and dev sets.
fastertransformer/examples/tensorflow/bert/tensorflow_bert/bert/run_classifier.py:318
↓ 3 callers
Method
_create_examples
Creates examples for the training and dev sets.
fastertransformer/examples/tensorflow/bert/tensorflow_bert/bert/run_classifier.py:358
↓ 3 callers
Method
_create_examples
Creates examples for the training and dev sets.
fastertransformer/examples/tensorflow/bert/bert-quantization/utils/create_glue_data.py:217
↓ 3 callers
Method
_create_examples
Creates examples for the training and dev sets.
fastertransformer/examples/tensorflow/bert/bert-quantization/utils/create_glue_data.py:257
↓ 3 callers
Method
_create_examples
Creates examples for the training and dev sets.
fastertransformer/examples/tensorflow/bert/bert-quantization/utils/create_glue_data.py:297
↓ 3 callers
Method
_create_examples
Creates examples for the training and dev sets.
fastertransformer/examples/tensorflow/xlnet/convertInput.py:164
↓ 3 callers
Function
_cudaGetErrorEnum
debug tools ********************************* */
fastertransformer/src/fastertransformer/utils/cuda_utils.h:74
↓ 3 callers
Function
_gather
Gather tensors and concatinate along the last dimension.
fastertransformer/3rdparty/Megatron-LM/megatron/mpu/mappings.py:54
↓ 3 callers
Function
_initialize_affine_weight_cpu
Initialize affine weight for model parallel. Build the master weight on all processes and scatter the relevant chunk.
fastertransformer/3rdparty/Megatron-LM/megatron/mpu/layers.py:93
↓ 3 callers
Function
_initialize_affine_weight_gpu
Initialize affine weight for model parallel on GPU.
fastertransformer/3rdparty/Megatron-LM/megatron/mpu/layers.py:80
↓ 3 callers
Function
_is_cuda_contiguous
Check if a tensor is not none, is cuda, and is contiguous.
fastertransformer/3rdparty/Megatron-LM/megatron/text_generation/communication.py:65
↓ 3 callers
Method
_map
(self, func)
fastertransformer/examples/pytorch/gptneox/utils/gptneox.py:113
↓ 3 callers
Method
_map
(self, func)
fastertransformer/examples/pytorch/gpt/utils/gpt.py:220
↓ 3 callers
Function
_normalize
(text)
fastertransformer/3rdparty/Megatron-LM/tasks/orqa/unsupervised/qa_utils.py:176
↓ 3 callers
Method
_read_tsv
Reads a tab separated value file.
fastertransformer/examples/tensorflow/xlnet/convertInput.py:115
↓ 3 callers
Function
_reduce
All-reduce the input tensor across model parallel group.
fastertransformer/3rdparty/Megatron-LM/megatron/mpu/mappings.py:22
↓ 3 callers
Function
_sacreformat
Format refs and preds for sacrebleu corpus calculation. It is very particular
algorithm/lm_eval/metrics.py:161
↓ 3 callers
Method
_set_mips_index
Create a Faiss Flat index with inner product as the metric to search against
fastertransformer/3rdparty/Megatron-LM/megatron/data/realm_index.py:130
↓ 3 callers
Function
_split
Split the tensor along its last dimension and keep the corresponding slice.
fastertransformer/3rdparty/Megatron-LM/megatron/mpu/mappings.py:35
↓ 3 callers
Method
_split_heads
Split the last dimension into (num_heads, head_dim), results share same memory storage as `fused_qkv` Args: fused_qk
algorithm/models/int_falcon_layer.py:65
↓ 3 callers
Function
_update_config_entry
(key, file_pattern)
fastertransformer/examples/pytorch/gpt/utils/nemo_ckpt_convert.py:449
↓ 3 callers
Function
_update_config_entry
(key, file_pattern)
fastertransformer/examples/pytorch/t5/utils/nemo_t5_ckpt_convert.py:1145
↓ 3 callers
Function
aK
(e)
fastertransformer/3rdparty/cutlass/docs/jquery.js:23
↓ 3 callers
Method
accumulator_type
(self)
fastertransformer/3rdparty/cutlass/tools/library/scripts/rank_k_operation.py:56
↓ 3 callers
Method
accumulator_type
(self)
fastertransformer/3rdparty/cutlass/tools/library/scripts/trmm_operation.py:56
↓ 3 callers
Method
accumulator_type
(self)
fastertransformer/3rdparty/cutlass/tools/library/scripts/conv2d_operation.py:43
↓ 3 callers
Method
accumulator_type
(self)
fastertransformer/3rdparty/cutlass/tools/library/scripts/symm_operation.py:58
↓ 3 callers
Method
accumulator_type
(self)
fastertransformer/3rdparty/cutlass/tools/library/scripts/rank_2k_operation.py:58
↓ 3 callers
Method
accumulator_type
(self)
fastertransformer/3rdparty/cutlass/tools/library/scripts/pycutlass/src/pycutlass/conv2d_operation.py:532
↓ 3 callers
Function
add_bias_and_interleave_quantized_tensor_inplace
fastertransformer/src/fastertransformer/kernels/cutlass_kernels/cutlass_preprocessors.cc:424
↓ 3 callers
Function
add_pointer_offset
Adds a pointer offset in units of Element
fastertransformer/3rdparty/cutlass/examples/42_fused_multi_head_attention/iterators/predicated_tile_access_iterator_residual_last.h:292
↓ 3 callers
Method
add_pointer_offset
Adds a pointer offset in units of Element
fastertransformer/3rdparty/cutlass/include/cutlass/transform/threadblock/regular_tile_access_iterator_tensor_op.h:684
↓ 3 callers
Method
add_pointer_offset
Adds a pointer offset in units of Element
fastertransformer/src/fastertransformer/cutlass_extensions/include/cutlass_extensions/epilogue/threadblock/epilogue_tensor_op_int32.h:207
↓ 3 callers
Function
add_special_tokens_to_tokenizer
(tokenizer)
fastertransformer/examples/pytorch/tokenizer.py:16
↓ 3 callers
Method
add_tile_offset
Adds a tile offset
fastertransformer/3rdparty/cutlass/include/cutlass/transform/threadblock/regular_tile_access_iterator_tensor_op.h:696
↓ 3 callers
Method
add_token
(self, token)
fastertransformer/3rdparty/Megatron-LM/megatron/tokenizer/tokenizer.py:163
↓ 3 callers
Method
asdict
(self)
fastertransformer/examples/pytorch/gpt/utils/gpt.py:747
← previous
next →
1,001–1,100 of 14,558, ranked by callers