MCPcopy Create free account

hub / github.com/HA-Transformer/MAT / functions

Functions2,038 in github.com/HA-Transformer/MAT

⨍ Functions 2,038◇ Types & classes 299
↓ 1 callersFunctionbase_wav2vec_architecture
(args)
fairseq/models/wav2vec.py:438
↓ 1 callersFunctionbert_base_architecture
(args)
fairseq/models/masked_lm.py:298
↓ 1 callersMethodbinarize
(self, sentence: str)
fairseq/hub_utils.py:175
↓ 1 callersFunctionbleu_ltrim
left trim (remove pad)
fairseq/clib/libbleu/libbleu.cpp:29
↓ 1 callersFunctionbleu_rtrim
right trim remove (eos)
fairseq/clib/libbleu/libbleu.cpp:40
↓ 1 callersFunctionbleu_zero_init
fairseq/clib/libbleu/libbleu.cpp:101
↓ 1 callersMethodbpe
(self, token)
fairseq/data/encoders/gpt2_bpe_utils.py:65
↓ 1 callersFunctionbuffered_arange
(max)
fairseq/utils.py:187
↓ 1 callersMethodbuffered_future_mask
(self, tensor)
fairseq/models/transformer.py:472
↓ 1 callersMethodbuffered_future_mask
(self, tensor)
examples/speech_recognition/models/vggtransformer.py:686
↓ 1 callersFunctionbuffered_read
(input, buffer_size)
interactive.py:23
↓ 1 callersFunctionbuffered_read
(input, buffer_size)
fairseq_cli/interactive.py:23
↓ 1 callersMethodbuild_criterion
(cls, args, task)
fairseq/criterions/composite_loss.py:35
↓ 1 callersMethodbuild_dataset_for_inference
(self, sample_json)
examples/roberta/wsc/wsc_task.py:207
↓ 1 callersMethodbuild_decoder
(cls, args, tgt_dict, embed_tokens)
fairseq/models/transformer.py:182
↓ 1 callersMethodbuild_decoder
(cls, args, tgt_dict, embed_tokens)
fairseq/models/transformer_mb/transformer_mb.py:185
↓ 1 callersMethodbuild_decoder
(cls, args, task)
examples/speech_recognition/models/vggtransformer.py:128
↓ 1 callersMethodbuild_encoder
(cls, args, src_dict, embed_tokens)
fairseq/models/transformer.py:178
↓ 1 callersMethodbuild_encoder
(cls, args, src_dict, embed_tokens)
fairseq/models/transformer_mb/transformer_mb.py:181
↓ 1 callersMethodbuild_encoder
(cls, args, task)
examples/speech_recognition/models/vggtransformer.py:118
↓ 1 callersMethodbuild_underlying_criterion
(args, task)
fairseq/criterions/composite_loss.py:26
↓ 1 callersFunctionbytes_to_unicode
Returns list of utf-8 byte and a corresponding list of unicode strings. The reversible bpe codes work on unicode strings. This means you
fairseq/data/encoders/gpt2_bpe_utils.py:13
↓ 1 callersFunctioncached_path
Given something that might be a URL (or might be a local path), determine which. If it's a URL, download the file and cache it, and retur
fairseq/file_utils.py:134
↓ 1 callersFunctioncalc_mean_invstddev
(feature)
examples/speech_recognition/data/data_utils.py:9
↓ 1 callersFunctioncli_main
()
validate.py:88
↓ 1 callersFunctioncli_main
()
interactive.py:186
↓ 1 callersFunctioncli_main
()
train.py:287
↓ 1 callersFunctioncli_main
()
eval_lm.py:216
↓ 1 callersFunctioncli_main
()
generate.py:204
↓ 1 callersFunctioncli_main
()
preprocess.py:264
↓ 1 callersFunctioncli_main
()
fairseq_cli/interactive.py:186
↓ 1 callersFunctioncli_main
()
fairseq_cli/train.py:287
↓ 1 callersFunctioncli_main
()
fairseq_cli/eval_lm.py:216
↓ 1 callersFunctioncli_main
()
fairseq_cli/generate.py:204
↓ 1 callersFunctioncli_main
()
fairseq_cli/preprocess.py:264
↓ 1 callersFunctioncli_main
()
examples/speech_recognition/infer.py:235
↓ 1 callersMethodclip_grad_norm
Clips gradient norm.
fairseq/optim/bmuf.py:84
↓ 1 callersMethodclip_grad_norm
Clips gradient norm and updates dynamic loss scaler.
fairseq/optim/fp16_optimizer.py:172
↓ 1 callersFunctioncollate
(samples, pad_idx, eos_idx)
fairseq/data/monolingual_dataset.py:12
↓ 1 callersMethodcollater
(self, samples)
fairseq/data/id_dataset.py:19
↓ 1 callersMethodcollater
Generate a mini-batch for this dataset. To convert this into a regular mini-batch we use the following logic: 1.
fairseq/data/multi_corpus_sampled_dataset.py:97
↓ 1 callersMethodcollater
(self, samples)
fairseq/data/audio/raw_audio_dataset.py:64
↓ 1 callersMethodcollater
Merge a list of samples to form a mini-batch. Args: samples (List[int]): sample indices to collate Returns:
examples/speech_recognition/data/asr_dataset.py:85
↓ 1 callersFunctioncollect_filtered
Similar to :func:`filter` but collects filtered elements in ``filtered``. Args: function (callable): function that returns ``False``
fairseq/data/data_utils.py:108
↓ 1 callersMethodcompute_fixed_attention_subset
(self, word_index, tgt_len)
fairseq/modules/sparse_multihead_attention.py:60
↓ 1 callersMethodcompute_loss
(self, model, net_output, sample, reduce=True)
fairseq/criterions/label_smoothed_cross_entropy.py:68
↓ 1 callersMethodcompute_loss
(self, model, net_output, target, reduction, log_probs)
examples/speech_recognition/criterions/cross_entropy_acc.py:22
↓ 1 callersFunctionconvert_state_dict_type
(state_dict, ttype=torch.FloatTensor)
fairseq/checkpoint_utils.py:207
↓ 1 callersMethodcopy_labels
(self)
scripts/wav2vec_featurize.py:171
↓ 1 callersFunctioncopy_tensor
(src, dst)
fairseq/data/data_utils.py:32
↓ 1 callersFunctiondecode
(l)
scripts/spm_decode.py:27
↓ 1 callersMethoddecode
(self, x: str)
fairseq/data/encoders/gpt2_bpe.py:42
↓ 1 callersFunctiondecode_fn
(x)
interactive.py:114
↓ 1 callersFunctiondecode_fn
(x)
fairseq_cli/interactive.py:114
↓ 1 callersFunctiondistributed_main
(i, args, start_rank=0)
train.py:280
↓ 1 callersFunctiondistributed_main
(i, args, start_rank=0)
fairseq_cli/train.py:280
↓ 1 callersFunctiondummy_dictionary
(vocab_size, prefix='token_')
tests/utils.py:20
↓ 1 callersFunctionencode
(l)
scripts/spm_encode.py:39
↓ 1 callersMethodencode
(self, sentence: str)
fairseq/hub_utils.py:145
↓ 1 callersFunctionencode_fn
(x)
interactive.py:107
↓ 1 callersFunctionencode_fn
(x)
fairseq_cli/interactive.py:107
↓ 1 callersMethodend_of_epoch
Returns whether the most recent epoch iterator has been exhausted
fairseq/data/iterators.py:185
↓ 1 callersFunctioneval_str_list
(x, type=float)
fairseq/options.py:59
↓ 1 callersFunctionevery_n_checkpoints
(files, n)
scripts/rm_pt.py:37
↓ 1 callersMethodextract_features
Returns: tuple: - the decoder's features of shape `(batch, tgt_len, embed_dim)` - a dictionary wi
fairseq/models/fairseq_decoder.py:36
↓ 1 callersMethodextract_features
Similar to *forward* but only return features. Returns: tuple: - the decoder's features of shape `(batch
fairseq/models/transformer.py:397
↓ 1 callersMethodextract_features
(self, src_tokens, return_all_hiddens=False, **unused)
fairseq/models/roberta/model.py:289
↓ 1 callersMethodextract_features
Similar to *forward* but only return features. Returns: tuple: - the decoder's features of shape `(batch
fairseq/models/transformer_mb/transformer_mb.py:441
↓ 1 callersMethodfc1
(self, x)
fairseq/models/transformer_mb/transformer_mb_layer.py:264
↓ 1 callersMethodfc1fc2
(self, x)
fairseq/models/transformer_mb/transformer_mb_layer.py:113
↓ 1 callersMethodfc1fc2
(self, x)
fairseq/models/transformer_mb/transformer_mb_layer.py:272
↓ 1 callersMethodfc2
(self, x)
fairseq/models/transformer_mb/transformer_mb_layer.py:268
↓ 1 callersFunctionfind_token
(sentence, start_pos)
examples/roberta/wsc/wsc_utils.py:52
↓ 1 callersMethodforward
( self, x: torch.Tensor )
fairseq/modules/highway.py:42
↓ 1 callersMethodforward
(self, src_tokens, src_lengths)
fairseq/models/fconv_self_att.py:198
↓ 1 callersMethodforward_decoder
(self, tokens, encoder_outs, temperature=1.)
fairseq/sequence_generator.py:542
↓ 1 callersMethodforward_encoder
(self, encoder_input)
fairseq/sequence_generator.py:536
↓ 1 callersMethodgenerate
(self, tokens: torch.LongTensor, beam: int = 5, verbose: bool = False, **kwargs)
fairseq/hub_utils.py:110
↓ 1 callersFunctionget_asr_dataset_from_json
Parse data json and create dataset. See scripts/asr_prep_json.py which pack json from raw files Json example: { "utts": {
examples/speech_recognition/tasks/speech_recognition.py:16
↓ 1 callersFunctionget_dataset_itr
(args, task)
examples/speech_recognition/infer.py:61
↓ 1 callersFunctionget_default_group
()
fairseq/distributed_utils.py:119
↓ 1 callersFunctionget_detokenizer
()
examples/roberta/wsc/wsc_utils.py:77
↓ 1 callersMethodget_encoder_langtok
(self, src_lang, tgt_lang)
fairseq/tasks/multilingual_translation.py:158
↓ 1 callersFunctionget_examples
Extract paragraph and question-answer list from each json file
examples/roberta/preprocess_RACE.py:21
↓ 1 callersFunctionget_from_cache
Given a URL, look for the corresponding dataset in the local cache. If it's not there, download it. Then return the path to the cached file.
fairseq/file_utils.py:230
↓ 1 callersFunctionget_generation_parser
(interactive=False, default_task='translation')
fairseq/options.py:31
↓ 1 callersMethodget_log_prob
Computes the log probabilities for all the words of the vocabulary, given a 2D tensor of hidden vectors.
fairseq/modules/adaptive_softmax.py:170
↓ 1 callersMethodget_logging_output
(self, sample, target, lprobs, loss)
examples/speech_recognition/criterions/cross_entropy_acc.py:45
↓ 1 callersMethodget_logits
(self, net_output)
fairseq/models/wav2vec.py:196
↓ 1 callersMethodget_lr
(self)
fairseq/optim/fp16_optimizer.py:113
↓ 1 callersMethodget_next_lr
(self, epoch)
fairseq/optim/lr_scheduler/polynomial_decay_schedule.py:40
↓ 1 callersMethodget_next_lr
(self, epoch)
fairseq/optim/lr_scheduler/fixed_schedule.py:37
↓ 1 callersMethodget_normalized_probs
(self, net_output, log_probs, sample=None)
examples/speech_recognition/models/vggtransformer.py:148
↓ 1 callersFunctionget_parser
()
score.py:18
↓ 1 callersFunctionget_parser
()
fairseq_cli/score.py:18
↓ 1 callersFunctionget_parser
()
scripts/read_binarized.py:12
↓ 1 callersFunctionget_parser
()
scripts/wav2vec_manifest.py:17
↓ 1 callersFunctionget_rank
()
fairseq/distributed_utils.py:111
↓ 1 callersFunctionget_spacy_nlp
()
examples/roberta/wsc/wsc_utils.py:84
↓ 1 callersMethodget_src_tokens
correct_prediction: True if the net_output (src_tokens) should predict the correct target aggregate: True if the
tests/speech_recognition/asr_test_base.py:516
← previousnext →501–600 of 2,038, ranked by callers