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 callersMethodget_target
(self, soft_target)
tests/speech_recognition/asr_test_base.py:535
↓ 1 callersMethodget_target_weights
(self, targets, net_output)
fairseq/models/wav2vec.py:204
↓ 1 callersMethodget_targets
(self, sample, net_output, expand_steps=True)
fairseq/models/wav2vec.py:200
↓ 1 callersFunctionget_valid_stats
(trainer, args, extra_meters=None)
train.py:249
↓ 1 callersFunctionget_valid_stats
(trainer, args, extra_meters=None)
fairseq_cli/train.py:249
↓ 1 callersFunctionget_world_size
()
fairseq/distributed_utils.py:115
↓ 1 callersFunctiongopen
()
scripts/count_docs.py:25
↓ 1 callersMethodhandle_ooms
c10d accumulates/syncs gradients between gpus during backward pass. In case of OOMs, gpus may fail to sync, so we manually iterate
fairseq/trainer.py:450
↓ 1 callersFunctionhttp_get
(url, temp_file)
fairseq/file_utils.py:216
↓ 1 callersMethodhub_models
(cls)
fairseq/models/fconv.py:45
↓ 1 callersMethodhub_models
(cls)
fairseq/models/fairseq_model.py:178
↓ 1 callersMethodhub_models
(cls)
fairseq/models/roberta/model.py:33
↓ 1 callersFunctionimport_user_module
(args)
fairseq/utils.py:276
↓ 1 callersMethodin_proj_qkv
(self, query)
fairseq/modules/multihead_attention.py:283
↓ 1 callersMethodinfer_conv_output_dim
(self, in_channels, input_dim)
examples/speech_recognition/models/vggtransformer.py:381
↓ 1 callersMethodinit_meters
(self, args)
fairseq/trainer.py:60
↓ 1 callersFunctionintra_ref
(refs)
examples/translation_moe/score.py:172
↓ 1 callersFunctionis_batch_full
(num_tokens)
fairseq/data/data_utils.py:205
↓ 1 callersFunctionis_better
(a, b)
fairseq/checkpoint_utils.py:32
↓ 1 callersFunctionis_master
(args)
fairseq/distributed_utils.py:18
↓ 1 callersFunctionitem
(tensor)
fairseq/utils.py:217
↓ 1 callersFunctionlabel_smoothed_nll_loss
(lprobs, target, epsilon, ignore_index=None, reduce=True)
fairseq/criterions/label_smoothed_cross_entropy.py:13
↓ 1 callersFunctionlast_n_checkpoints
(files, n)
scripts/rm_pt.py:32
↓ 1 callersFunctionlast_n_checkpoints
(paths, n, update_based, upper_bound=None)
scripts/average_checkpoints.py:70
↓ 1 callersMethodlayer_mb
(self, layer_branches, x, encoder_padding_mask)
fairseq/models/transformer_mb/transformer_mb.py:239
↓ 1 callersMethodlayer_mb
(self, layer_branches, layer_args, layer_kwargs)
fairseq/models/transformer_mb/transformer_mb.py:418
↓ 1 callersMethodlengths_to_attn_mask
create attention mask according to sequence lengths and transformer context Args: - input_lengths: (B, )-shape I
examples/speech_recognition/models/vggtransformer.py:488
↓ 1 callersMethodload_dataset
Load a given dataset split. Args: split (str): name of the split (e.g., train, valid, test)
examples/roberta/wsc/wsc_task.py:80
↓ 1 callersMethodload_dataset
Load a given dataset split. Args: split (str): name of the split (e.g., train, valid, test)
examples/speech_recognition/tasks/speech_recognition.py:96
↓ 1 callersMethodload_dictionary
Load the dictionary from the filename Args: filename (str): the filename
fairseq/tasks/sentence_ranking.py:59
↓ 1 callersMethodload_dictionary
Load the dictionary from the filename Args: filename (str): the filename
examples/roberta/commonsense_qa/commonsense_qa_task.py:49
↓ 1 callersFunctionload_model_ensemble_and_task
(filenames, arg_overrides=None, task=None)
fairseq/checkpoint_utils.py:159
↓ 1 callersFunctionload_ref
(path)
examples/translation_moe/score.py:73
↓ 1 callersMethodload_state_dict
Load an optimizer state dict. In general we should prefer the configuration of the existing optimizer instance (e.g., learning rate)
fairseq/optim/fp16_optimizer.py:125
↓ 1 callersMethodload_state_dict
Load an optimizer state dict. In general we should prefer the configuration of the existing optimizer instance (e.g., learning rate)
fairseq/optim/fp16_optimizer.py:285
↓ 1 callersFunctionload_sys
(paths)
examples/translation_moe/score.py:52
↓ 1 callersFunctionmain
(args, override_args=None)
validate.py:14
↓ 1 callersFunctionmain
(args)
interactive.py:57
↓ 1 callersFunctionmain
(parsed_args)
eval_lm.py:46
↓ 1 callersFunctionmain
()
score.py:35
↓ 1 callersFunctionmain
(args)
generate.py:16
↓ 1 callersFunctionmain
(args)
preprocess.py:23
↓ 1 callersFunctionmain
(args)
fairseq_cli/interactive.py:57
↓ 1 callersFunctionmain
(parsed_args)
fairseq_cli/eval_lm.py:46
↓ 1 callersFunctionmain
()
fairseq_cli/score.py:35
↓ 1 callersFunctionmain
(args)
fairseq_cli/generate.py:16
↓ 1 callersFunctionmain
(args)
fairseq_cli/preprocess.py:23
↓ 1 callersFunctionmain
()
scripts/read_binarized.py:25
↓ 1 callersFunctionmain
()
scripts/rm_pt.py:42
↓ 1 callersFunctionmain
()
scripts/compare_namespaces.py:7
↓ 1 callersFunctionmain
()
scripts/split_train_valid_docs.py:16
↓ 1 callersFunctionmain
()
scripts/average_checkpoints.py:91
↓ 1 callersFunctionmain
()
scripts/read_checkpoint.py:18
↓ 1 callersFunctionmain
()
scripts/count_docs.py:19
↓ 1 callersFunctionmain
()
scripts/spm_decode.py:15
↓ 1 callersFunctionmain
()
scripts/init-mbt-ckpt.py:23
↓ 1 callersFunctionmain
()
scripts/build_sym_alignment.py:29
↓ 1 callersFunctionmain
(args)
scripts/wav2vec_manifest.py:30
↓ 1 callersFunctionmain
()
scripts/shard_docs.py:17
↓ 1 callersFunctionmain
()
scripts/spm_encode.py:17
↓ 1 callersFunctionmain
()
examples/translation_moe/score.py:23
↓ 1 callersFunctionmain
Helper script to extract paragraphs questions and answers from RACE datasets.
examples/roberta/preprocess_RACE.py:60
↓ 1 callersFunctionmain
Helper script to encode raw text with the GPT-2 BPE using multiple processes. The encoder.json and vocab.bpe files can be obtained here:
examples/roberta/multiprocessing_bpe_encoder.py:18
↓ 1 callersFunctionmain
(args)
examples/speech_recognition/infer.py:137
↓ 1 callersFunctionmain
()
examples/speech_recognition/datasets/asr_prep_json.py:39
↓ 1 callersFunctionmake_batches
(lines, args, task, max_positions, encode_fn)
interactive.py:36
↓ 1 callersFunctionmake_batches
(lines, args, task, max_positions, encode_fn)
fairseq_cli/interactive.py:36
↓ 1 callersFunctionmake_binary_dataset
(vocab, input_prefix, output_prefix, lang, num_workers)
preprocess.py:97
↓ 1 callersFunctionmake_binary_dataset
(vocab, input_prefix, output_prefix, lang, num_workers)
fairseq_cli/preprocess.py:97
↓ 1 callersMethodmake_generation_fast_
Optimize model for faster generation.
fairseq/models/fairseq_model.py:98
↓ 1 callersFunctionmap_value_update
(d1, d2)
fairseq/utils.py:241
↓ 1 callersMethodmax_decoder_positions
(self)
fairseq/sequence_generator.py:532
↓ 1 callersMethodmax_positions
Maximum input length supported by the encoder.
fairseq/models/fconv.py:281
↓ 1 callersMethodmax_positions
Maximum input length supported by the encoder.
fairseq/models/lightconv.py:271
↓ 1 callersMethodmax_positions
Maximum input length supported by the encoder.
fairseq/models/fconv_self_att.py:269
↓ 1 callersMethodmax_positions
Maximum input length supported by the encoder.
fairseq/models/transformer.py:285
↓ 1 callersMethodmax_positions
Maximum input length supported by the encoder.
fairseq/models/transformer_mb/transformer_mb.py:307
↓ 1 callersFunctionmerge
(src, tgt, hypos, log_probs, path)
examples/translation_moe/score.py:94
↓ 1 callersFunctionmerge_result
(worker_result)
preprocess.py:102
↓ 1 callersMethodmeteor_score
(self)
fairseq/bleu.py:122
↓ 1 callersFunctionmock_trainer
(epoch, num_updates, iterations_in_epoch)
tests/test_train.py:16
↓ 1 callersFunctionmulti_ref
(refs, hypos)
examples/translation_moe/score.py:134
↓ 1 callersMethodmultiply_grads
Multiplies grads by a constant *c*.
fairseq/optim/bmuf.py:80
↓ 1 callersMethodno_sync
A context manager to disable gradient synchronization.
fairseq/legacy_distributed_data_parallel.py:78
↓ 1 callersMethodnoising
(self, x, lengths, max_shuffle_distance=None)
fairseq/data/noising.py:151
↓ 1 callersFunctionoptimize_models
Optimize ensemble for generation
examples/speech_recognition/infer.py:123
↓ 1 callersMethodordered_indices
(self)
fairseq/data/sort_dataset.py:21
↓ 1 callersMethodordered_indices
Ordered indices for batching. Here we call the underlying dataset's ordered_indices() so that we get the same random ordering
fairseq/data/multi_corpus_sampled_dataset.py:56
↓ 1 callersMethodoutput_layer
Project features to the default output size, e.g., vocabulary size. Args: features (Tensor): features returned by *extra
fairseq/models/fairseq_decoder.py:45
↓ 1 callersMethodoutput_layer
Project features to the default output size (typically vocabulary size).
fairseq/models/fairseq_model.py:380
↓ 1 callersMethodoutput_layer
Project features to the vocabulary size.
fairseq/models/transformer.py:455
↓ 1 callersMethodoutput_layer
(self, features, **unused)
fairseq/models/roberta/model.py:296
↓ 1 callersMethodoutput_layer
Project features to the vocabulary size.
fairseq/models/transformer_mb/transformer_mb.py:515
↓ 1 callersFunctionparse_args
()
scripts/read_checkpoint.py:9
↓ 1 callersMethodparse_transformer_context
transformer_context can be the following: - None; indicates no context is used, i.e., transformer can access full conte
examples/speech_recognition/models/vggtransformer.py:402
↓ 1 callersMethodparse_transformer_sampling
parsing transformer sampling configuration Args: - transformer_sampling, accepted input: * None, indicat
examples/speech_recognition/models/vggtransformer.py:436
↓ 1 callersMethodpov
Percent of valid code.
fairseq/bleu.py:268
↓ 1 callersMethodprefetch
(self, indices)
fairseq/data/noising.py:313
↓ 1 callersMethodprepare_for_onnx_export_
Make model exportable via ONNX trace.
fairseq/models/fairseq_model.py:131
↓ 1 callersFunctionprepare_result_files
(args)
examples/speech_recognition/infer.py:105
← previousnext →601–700 of 2,038, ranked by callers