Functions — github.com/OpenBMB/CPM-Bee

hub / github.com/OpenBMB/CPM-Bee / functions

Functions302 in github.com/OpenBMB/CPM-Bee

⨍ Functions 302◇ Types & classes 69

↓ 14 callersFunction_dataset_identity

(c: _MixedDatasetConfig)

src/cpm_live/training_tasks/bee/pretrain.py:248

↓ 12 callersMethodwrite

( self, time: float, iteration: int, loss: float, lr: float, l

src/cpm_live/utils/log.py:59

↓ 11 callersMethodread

(self, size=-1)

src/cpm_live/dataset/distributed_dataset.py:143

↓ 10 callersMethoddumps

(self)

src/cpm_live/dataset/distributed_dataset.py:76

↓ 10 callersMethodloads

(self, data: str)

src/cpm_live/dataset/distributed_dataset.py:79

↓ 9 callersMethodload

(self, fp: io.TextIOWrapper)

src/cpm_live/dataset/distributed_dataset.py:87

↓ 8 callersMethodprojection

Projection based on embedding's weight. For example, embedding map vocab_size to embed_size, than projection map embed_size back to vocab_siz

src/cpm_live/layers/embedding.py:55

↓ 7 callersFunction_read_info_list

(meta_path: str)

src/cpm_live/dataset/distributed_dataset.py:92

↓ 7 callersMethodget

(self)

src/cpm_live/training_tasks/bee/pretrain.py:1026

↓ 6 callersMethodclose

(self)

src/cpm_live/dataset/distributed_dataset.py:179

↓ 6 callersMethodinference

( self, input: torch.Tensor, # (batch, seqlen) length: torch.Tensor, # (batch)

src/cpm_live/models/ant.py:156

↓ 5 callersFunction_convert_to_abs_path

(transform_path: str)

src/cpm_live/training_tasks/bee/pretrain.py:757

↓ 5 callersMethod_update_states

(self, fast_skip: bool = True)

src/cpm_live/dataset/distributed_dataset.py:263

↓ 5 callersMethodadd

Add a new hypothesis to the list.

src/cpm_live/generation/generation_utils.py:87

↓ 5 callersFunctionadd_mem_time

(info, mem_usage, tim_usage)

src/finetune_cpm_bee.py:109

↓ 5 callersMethodload_state_dict

(self, d)

src/cpm_live/dataset/distributed_dataset.py:67

↓ 5 callersMethodtell

(self)

src/cpm_live/dataset/distributed_dataset.py:155

↓ 5 callersMethodwrite

Write a piece of data into dataset. Args: data (Any): Serialization will be done using pickle. Example: >>>

src/cpm_live/dataset/distributed_dataset.py:632

↓ 4 callersMethod_get_block_file

(self, block_id: int)

src/cpm_live/dataset/distributed_dataset.py:369

↓ 4 callersFunction_write_info_list

(meta_path: str, info: List[FileInfo])

src/cpm_live/dataset/distributed_dataset.py:108

↓ 4 callersFunctionadd_mem_time

(info, mem_usage, tim_usage)

src/pretrain_cpm_bee.py:109

↓ 4 callersMethodopen_file

(self, fname, mode, offset=None)

src/cpm_live/dataset/distributed_dataset.py:187

↓ 3 callersFunction_build_sample_weights

(config: List[_MixedDatasetConfig])

src/cpm_live/training_tasks/bee/pretrain.py:763

↓ 3 callersFunction_filtered_range

( begin: int, end: int, rank: int, world_size: int, filter_set: Optional[Set[int]] = None )

src/cpm_live/dataset/distributed_dataset.py:123

↓ 3 callersFunction_random_string

()

src/cpm_live/dataset/distributed_dataset.py:30

↓ 3 callersFunctionapply_repetition_penalty

( logits, batch_size, num_beams, prev_output_tokens, repetition_penalty, start_idx=Non

src/cpm_live/generation/generation_utils.py:34

↓ 3 callersFunctionbuild_dataset

Open the dataset in write mode and returns a writer. Args: path (str): Path to dataset. dbname (str): The name of the file to whi

src/cpm_live/dataset/distributed_dataset.py:750

↓ 3 callersFunctionconvert_data_to_id

( tokenizer: CPMBeeTokenizer, data: Any, prev_ext_states: Optional[_PrevExtTableStates] = None,

src/cpm_live/training_tasks/bee/pretrain.py:101

↓ 3 callersMethodencode

Encode a string into ids.

src/cpm_live/tokenizers/ant.py:142

↓ 3 callersMethodfrom_json_file

(cls, json_file: Union[str, os.PathLike], **args)

src/cpm_live/utils/config.py:54

↓ 3 callersFunctiongradient_shrink

(x: torch.Tensor, alpha: float = 0.1)

src/cpm_live/utils/gradient_shrink.py:15

↓ 3 callersMethodstate_dict

(self)

src/cpm_live/dataset/distributed_dataset.py:56

↓ 2 callersMethod__init__

( self, dim_model: int, dim_ff: int, dtype=torch.half, eps: float = 1e

src/cpm_live/layers/blocks.py:113

↓ 2 callersMethod__init__

( self, dim, base=10000, distance_scale: Union[int, float] = 1, dtype:

src/cpm_live/layers/position_embedding.py:220

↓ 2 callersMethod__init__

( self, dim_model: int, dim_ff: int, dtype=torch.half, eps: float = 1e

src/cpm_live/native_layers/blocks.py:112

↓ 2 callersMethod__init__

( self, dim, base=10000, distance_scale: Union[int, float] = 1, dtype:

src/cpm_live/native_layers/position_embedding.py:213

↓ 2 callersMethod_open_file

(self, fname, offset)

src/cpm_live/dataset/distributed_dataset.py:536

↓ 2 callersMethodadd_data

(self, config: _MixedDatasetConfig)

src/cpm_live/training_tasks/bee/pretrain.py:625

↓ 2 callersFunctionallgather_objects

(obj)

src/cpm_live/utils/object.py:6

↓ 2 callersMethodapply_transform

( self, data: CPMBeeInputType, transform: Union[Dict[str, Any], Callable[[CPMBeeInputT

src/cpm_live/training_tasks/bee/pretrain.py:281

↓ 2 callersMethodbackward

(ctx, grad_output)

src/cpm_live/utils/gradient_shrink.py:11

↓ 2 callersMethodchanged

(self)

src/cpm_live/training_tasks/bee/pretrain.py:706

↓ 2 callersMethodclose

(self)

src/cpm_live/training_tasks/bee/pretrain.py:971

↓ 2 callersMethoddata_to_id

( self, data: Any, prev_ext_states: Optional[_PrevExtTableStates] = None, shuf

src/cpm_live/training_tasks/bee/pretrain.py:358

↓ 2 callersMethoddecode

Decode ids into a string.

src/cpm_live/tokenizers/ant.py:146

↓ 2 callersMethodgenerate

(self, text_list, **kwargs)

src/cpm_live/generation/ant.py:41

↓ 2 callersFunctionget_args

(pretrain: bool = False, finetune: bool = False)

src/cpm_live/arguments.py:237

↓ 2 callersMethodget_config

(self)

src/cpm_live/training_tasks/bee/pretrain.py:733

↓ 2 callersMethodget_log_name

(self, now: Optional[datetime.datetime] = None)

src/cpm_live/utils/log.py:54

↓ 2 callersMethodget_log_time

(self)

src/cpm_live/utils/log.py:51

↓ 2 callersMethodis_done

If there are enough hypotheses and that none of the hypotheses being generated can become better than the worst one in the heap, then

src/cpm_live/generation/generation_utils.py:102

↓ 2 callersMethodload_state_dict

(self, data: OrderedDict, strict: bool = False)

src/cpm_live/training_tasks/bee/pretrain.py:1017

↓ 2 callersMethodpack_batch

(self, force: bool = False)

src/cpm_live/training_tasks/bee/pretrain.py:516

↓ 2 callersFunctionpad

(orig_items, key, padding_value=0, padding_side="left")

src/cpm_live/utils/data_utils.py:4

↓ 2 callersFunctionrel_to_bucket

(n_up: int, n_down: int, max_depth: int = 8)

src/cpm_live/training_tasks/bee/pretrain.py:92

↓ 2 callersMethodseek

(self, offset, whence=0)

src/cpm_live/dataset/distributed_dataset.py:167

↓ 2 callersMethodstart

(self)

src/cpm_live/training_tasks/bee/pretrain.py:982

↓ 1 callersMethod__batch_iter

(self)

src/cpm_live/training_tasks/bee/finetune.py:39

↓ 1 callersMethod__get_item_data

(self, raw_data)

src/cpm_live/training_tasks/ant/pretrain.py:35

↓ 1 callersMethod__init__

( self, file_name: str = "", block_begin: int = 0, block_end: int = 0,

src/cpm_live/dataset/distributed_dataset.py:38

↓ 1 callersMethod__init__

( self, vocab_size: int, embedding_size: int, dtype: torch.dtype = torch.half,

src/cpm_live/layers/embedding.py:25

↓ 1 callersMethod__init__

( self, dim_model: int, dim_ff: int, dtype=torch.half, dropout_p: Opti

src/cpm_live/layers/feedforward.py:79

↓ 1 callersMethod__init__

(self, config: CPMAntConfig)

src/cpm_live/models/ant.py:66

↓ 1 callersMethod__init__

(self, config: CPMBeeConfig)

src/cpm_live/models/bee.py:73

↓ 1 callersMethod__init__

( self, vocab_size: int, embedding_size: int, dtype: torch.dtype = torch.half,

src/cpm_live/native_layers/embedding.py:24

↓ 1 callersMethod__init__

( self, dim_model: int, dim_ff: int, dtype=torch.half, dropout_p: Opti

src/cpm_live/native_layers/feedforward.py:77

↓ 1 callersMethod__str__

(self)

src/cpm_live/tokenizers/bee.py:42

↓ 1 callersFunction_build_dict_tree

(data: CPMBeeInputType, depth: int, need_predict: bool)

src/cpm_live/training_tasks/bee/pretrain.py:118

↓ 1 callersFunction_build_segment_rel

(node: _DictTree)

src/cpm_live/training_tasks/bee/pretrain.py:158

↓ 1 callersMethod_convert_to_tensors

(self, data: Any, in_context_samples: List[Any] = [])

src/cpm_live/generation/bee.py:18

↓ 1 callersMethod_decode

(self, model_inputs, **kwargs)

src/cpm_live/generation/ant.py:47

↓ 1 callersMethod_decode

(self, model_inputs, other_info, **kwargs)

src/cpm_live/generation/bee.py:250

↓ 1 callersMethod_dict_from_json_file

(cls, json_file: Union[str, os.PathLike], **args)

src/cpm_live/utils/config.py:59

↓ 1 callersMethod_ensure_transform_function

( self, module_name: str, transform_script_path: str )

src/cpm_live/training_tasks/bee/pretrain.py:368

↓ 1 callersMethod_get_file_path

(self, fname)

src/cpm_live/dataset/distributed_dataset.py:533

↓ 1 callersFunction_get_logger

()

src/cpm_live/utils/log.py:11

↓ 1 callersMethod_get_next_block

(self)

src/cpm_live/dataset/distributed_dataset.py:389

↓ 1 callersMethod_mask_file

(self, f: FileInfo)

src/cpm_live/dataset/distributed_dataset.py:362

↓ 1 callersMethod_position_bucket

( self, relative_position, bidirectional=True, num_buckets=32, max_distance=128 )

src/cpm_live/layers/position_embedding.py:104

↓ 1 callersMethod_position_bucket

(self, relative_position, num_buckets=32, max_distance=128)

src/cpm_live/layers/position_embedding.py:196

↓ 1 callersMethod_position_bucket

( self, relative_position, bidirectional=True, num_buckets=32, max_distance=128 )

src/cpm_live/native_layers/position_embedding.py:100

↓ 1 callersMethod_position_bucket

(self, relative_position, num_buckets=32, max_distance=128)

src/cpm_live/native_layers/position_embedding.py:189

↓ 1 callersMethod_prepare_new_epoch

(self)

src/cpm_live/dataset/distributed_dataset.py:374

↓ 1 callersMethod_process_list

(self, data_list: List[Any])

src/cpm_live/generation/bee.py:124

↓ 1 callersMethod_process_texts

(self, text_list)

src/cpm_live/generation/ant.py:33

↓ 1 callersMethod_segment_relative_position_bucket

(self, query_segment, key_segment)

src/cpm_live/layers/position_embedding.py:101

↓ 1 callersMethod_segment_relative_position_bucket

(self, query_segment, key_segment)

src/cpm_live/native_layers/position_embedding.py:97

↓ 1 callersMethod_state_dict

(self)

src/cpm_live/dataset/distributed_dataset.py:399

↓ 1 callersMethod_write_log

(self, iteration: int, data: Any, result: List[Dict[str, Any]])

src/pretrain_cpm_bee.py:144

↓ 1 callersFunctionadd_finetune_args

(parser: argparse.ArgumentParser)

src/cpm_live/arguments.py:197

↓ 1 callersFunctionadd_model_config_args

Model arguments

src/cpm_live/arguments.py:19

↓ 1 callersFunctionadd_pretrain_args

(parser: argparse.ArgumentParser)

src/cpm_live/arguments.py:179

↓ 1 callersFunctionadd_training_args

Training arguments.

src/cpm_live/arguments.py:32

↓ 1 callersMethodbuild_instance

(self, config: _MixedDatasetConfig)

src/cpm_live/training_tasks/bee/pretrain.py:403

↓ 1 callersMethoddeserialize

(self, data: bytes)

src/cpm_live/dataset/serializer.py:27

↓ 1 callersMethoddump

(self, fp: io.TextIOWrapper)

src/cpm_live/dataset/distributed_dataset.py:83

↓ 1 callersMethodescape

(text: str)

src/cpm_live/tokenizers/bee.py:179

↓ 1 callersFunctionevaluation

(model, args, tokenizer, loss_func)

src/finetune_cpm_bee.py:116

↓ 1 callersFunctionfinetune

( args, tokenizer: CPMBeeTokenizer, model: CPMBee, optimizer: bmt.optim.AdamOffloadOptimizer,

src/finetune_cpm_bee.py:188

↓ 1 callersFunctionget_args

()

src/preprocess_dataset.py:9

next →1–100 of 302, ranked by callers