MCPcopy Create free account

hub / github.com/ASLP-lab/OSUM-Pangu / functions

Functions1,198 in github.com/ASLP-lab/OSUM-Pangu

⨍ Functions 1,198◇ Types & classes 199
Method__init__
(self, config)
patches/modelling_qwen2_infer_gpu.py:217
Method__init__
(self, text_token_num: int)
patches/cumstom_stop_criteria.py:6
Method__init__
(self, text_token_num: int)
patches/cumstom_stop_criteria.py:17
Method__init__
(self, text_token_num: int, text_eos_id: int)
patches/cumstom_stop_criteria.py:30
Method__init__
(self, text_eos_id: int, speech_eos_id: int)
patches/cumstom_stop_criteria.py:53
Method__init__
(self, max_tokens)
patches/cumstom_stop_criteria.py:66
Method__init__
(self)
patches/cumstom_stop_criteria.py:77
Method__init__
Create a ContextState. Args: id: The node id, only for visualization now. A node is in [0, graph.num_nodes).
wenet/utils/context_graph.py:63
Method__init__
Initialize a ContextGraph with the given ``context_score``. A root node will be created (**NOTE:** the token of root is hardcoded to -1).
wenet/utils/context_graph.py:115
Method__init__
(self, global_step: int = 0, device: torch.device = torch.device("cpu"))
wenet/utils/executor.py:34
Method__init__
(self, optimizer, *, warmup_steps=None, wa
wenet/utils/scheduler.py:89
Method__init__
(self, optimizer, *, constant_steps=None,
wenet/utils/scheduler.py:152
Method__init__
( self, optimizer, *, warmup_steps=None, warmup_ratio=None, ho
wenet/utils/scheduler.py:217
Method__init__
( self, optimizer, *, warmup_steps=None, warmup_ratio=None, co
wenet/utils/scheduler.py:300
Method__init__
(self, optimizer, *, max_steps, min_lr=1e-
wenet/utils/scheduler.py:451
Method__init__
(self, optimizer, *, max_steps, min_lr=0,
wenet/utils/scheduler.py:478
Method__init__
(self, optimizer, *, max_steps, min_lr=0,
wenet/utils/scheduler.py:504
Method__init__
(self, optimizer, *, d_model, warmup_steps
wenet/utils/scheduler.py:565
Method__init__
From Nemo: Implementation of the Noam Hold Annealing policy from the SqueezeFormer paper. Unlike NoamAnnealing, the
wenet/utils/scheduler.py:631
Method__init__
(self, step=0.0)
wenet/utils/common.py:343
Method__init__
( self, size: int, kernel_size: int, dropout_rate: float, use_linear_a
wenet/branchformer/cgmlp.py:33
Method__init__
(self, p: List[float], modules=None)
wenet/branchformer/encoder.py:168
Method__init__
( self, size: int, attn: Optional[torch.nn.Module], cgmlp: Optional[torch.nn.M
wenet/branchformer/encoder_layer.py:41
Method__init__
(self, source, f, *args, **kw)
wenet/dataset/dataset.py:29
Method__init__
(self, shuffle=True, partition=True, split_num=1,multi_num=1)
wenet/dataset/dataset.py:54
Method__init__
(self, lists, shuffle=True, partition=True, split_num=1)
wenet/dataset/dataset.py:163
Method__init__
(self,s2t_dataset,t2s_dataset,s2s_dataset,t2t_dataset, weight_num:List[int])
wenet/dataset/dataset.py:183
Method__init__
(self, dataset: IterDataPipe, fn: Callable, input_col=None,
wenet/dataset/datapipes.py:37
Method__init__
( self, dataset: IterDataPipe, elem_length_func, bucket_boundaries: List[int],
wenet/dataset/datapipes.py:66
Method__init__
( self, dataset: IterDataPipe, key_func, window_size_func, wrapper_cla
wenet/dataset/datapipes.py:105
Method__init__
(self, dataset: IterDataPipe, window_class, wrapper_class)
wenet/dataset/datapipes.py:188
Method__init__
( self, dataset: IterDataPipe, buffer_size: int = 500, )
wenet/dataset/datapipes.py:219
Method__init__
(self, dataset: IterDataPipe, count: int = -1)
wenet/dataset/datapipes.py:262
Method__init__
(self, dataset: IterDataPipe, partition: bool = False)
wenet/dataset/datapipes.py:282
Method__init__
( self, source_datapipes: List[IterDataPipe], weights: Optional[List[float]] = None,
wenet/dataset/datapipes.py:309
Method__init__
(self, filenames, mode='r')
wenet/dataset/datapipes.py:351
Method__init__
(self, dataset: IterDataPipe)
wenet/dataset/datapipes.py:370
Method__init__
(self, filenames: str, prefetch: int = 500, partition: bool
wenet/dataset/datapipes.py:432
Method__init__
(self, filenames: str, prefetch: int = 500, partition: bool
wenet/dataset/datapipes.py:453
Method__init__
(self)
wenet/transducer/predictor.py:21
Method__init__
(self, voca_size: int, embed_size: int, output_size: int,
wenet/transducer/predictor.py:218
Method__init__
(self, voca_size: int, embed_size: int, output_size: int,
wenet/transducer/predictor.py:381
Method__init__
(self, vocab_size: int, enc_output_size: int, pred_output_s
wenet/transducer/joint.py:10
Method__init__
( self, vocab_size: int, blank: int, encoder: nn.Module, predictor: Pr
wenet/transducer/transducer.py:22
Method__init__
( self, hyp: List[torch.Tensor], score, cache: List[torch.Tensor], )
wenet/transducer/search/prefix_beam_search.py:11
Method__init__
(self, encoder, predictor, joint, ctc, blank)
wenet/transducer/search/prefix_beam_search.py:24
Method__init__
( self, input_size: int, output_size: int = 256, attention_heads: int = 4,
wenet/e_branchformer/encoder.py:36
Method__init__
( self, size: int, attn: torch.nn.Module, cgmlp: torch.nn.Module, feed
wenet/e_branchformer/encoder_layer.py:39
Method__init__
Construct an PositionwiseFeedForward object.
wenet/paraformer/layers.py:104
Method__init__
Resize input in_size to size
wenet/paraformer/layers.py:127
Method__init__
( self, input_size: int, output_size: int = 256, attention_heads: int = 4,
wenet/paraformer/layers.py:184
Method__init__
(self, hidden: int, pos_clss: torch.nn.Module)
wenet/paraformer/layers.py:305
Method__init__
(self, size: int, self_attn: Optional[torch.nn.Module], src
wenet/paraformer/layers.py:316
Method__init__
( self, vocab_size: int, encoder_output_size: int, attention_heads: int = 4,
wenet/paraformer/layers.py:383
Method__init__
(self, idim: int, odim: int, dropout_rate: float, pos_enc_class: torch.nn.Module)
wenet/paraformer/subsampling.py:10
Method__init__
Construct an MultiHeadedAttention object.
wenet/paraformer/attention.py:16
Method__init__
(self, n_head, in_feat, n_feat, dropout_ra
wenet/paraformer/attention.py:163
Method__init__
(self, vocab_size: int, encoder: BaseEncoder, decoder: Tran
wenet/paraformer/paraformer.py:112
Method__init__
(self, normalize_length=False)
wenet/paraformer/cif.py:211
Method__init__
(self, depth: int, d_model: int, dropout_rate: float = 0.1,
wenet/paraformer/embedding.py:8
Method__init__
( self, n_kv_head: int, head_dim: int, hidden_size: int, attention_hea
wenet/LLM/decoder.py:17
Method__init__
( self, vocab_size: int, decoder: DecoderOnly, special_tokens: dict, t
wenet/LLM/causallm_model.py:11
Method__init__
(self, model_dir: str, resample_rate: int = 16000)
wenet/cli/paraformer_model.py:16
Method__init__
(self)
wenet/cli/hub.py:83
Method__init__
(self, model_dir: str, gpu: int = -1, beam: int = 5,
wenet/cli/model.py:33
Method__init__
( self, vocab_size: int, encoder: TransformerEncoder, decoder: TransformerDeco
wenet/whisper/whisper_with_clap.py:32
Method__init__
( self, vocab_size: int, encoder: TransformerEncoder, decoder: TransformerDeco
wenet/whisper/whisper.py:30
Method__init__
( self, vocab_size: int, encoder: TransformerEncoder, decoder:
wenet/k2/model.py:29
Method__init__
Wrap encoder to train using W2V-BERT's style Described in: https://arxiv.org/pdf/2108.06209v2.pdf Args: encoder
wenet/ssl/w2vbert/w2vbert_model.py:18
Method__init__
(self, features_dim: int = 256, num_codebooks: int = 2, num
wenet/ssl/wav2vec2/quantizer.py:28
Method__init__
Wrap encoder to train using wav2vec2's style Args: encoder: wenet's encoder, only support conformer and tra
wenet/ssl/wav2vec2/wav2vec2_model.py:107
Method__init__
( self, encoder: torch.nn.Module, num_mel_bins: int = 80, embedding_dim: int =
wenet/ssl/bestrq/bestrq_model.py:59
Method__init__
Construct a PositionwiseFeedForward object.
wenet/squeezeformer/positionwise_feed_forward.py:34
Method__init__
(self, idim: int, odim: int, pos_enc_class: torch.nn.Module
wenet/squeezeformer/subsampling.py:39
Method__init__
(self, kernel_size: int = 5, stride: int = 2, encoder_dim:
wenet/squeezeformer/subsampling.py:182
Method__init__
(self, channel: int, out_dim: int, kernel_size: int = 1,
wenet/squeezeformer/subsampling.py:254
Method__init__
Construct an RelPositionMultiHeadedAttention object.
wenet/squeezeformer/attention.py:35
Method__init__
Construct an ConvolutionModule object. Args: channels (int): The number of channels of conv layers. kernel_size (int):
wenet/squeezeformer/convolution.py:27
Method__init__
Construct SqueezeformerEncoder Args: input_size to use_dynamic_chunk, see in Transformer BaseEncoder. encoder_dim (in
wenet/squeezeformer/encoder.py:37
Method__init__
( self, in_channels: int, out_channels: int, kernel_size: _siz
wenet/squeezeformer/conv2d.py:25
Method__init__
( self, size: int, self_attn: torch.nn.Module, feed_forward1: Optional[nn.Modu
wenet/squeezeformer/encoder_layer.py:40
Method__init__
Construct an Conv2dSubsampling4 object.
wenet/efficient_conformer/subsampling.py:34
Method__init__
Construct an RelPositionMultiHeadedAttention object.
wenet/efficient_conformer/attention.py:39
Method__init__
Construct an ConvolutionModule object. Args: channels (int): The number of channels of conv layers. kernel_size (int):
wenet/efficient_conformer/convolution.py:26
Method__init__
Construct Efficient Conformer Encoder Args: input_size to use_dynamic_chunk, see in BaseEncoder macaron_style (bool):
wenet/efficient_conformer/encoder.py:44
Method__init__
Construct an EncoderLayer object.
wenet/efficient_conformer/encoder_layer.py:44
Method__init__
Construct a PositionwiseFeedForward object.
wenet/transformer/positionwise_feed_forward.py:33
Method__init__
Construct a PositionwiseFeedForward object.
wenet/transformer/positionwise_feed_forward.py:128
Method__init__
( self, vocab_size: int, encoder_output_size: int, attention_heads: int = 4,
wenet/transformer/decoder.py:334
Method__init__
Construct an DecoderLayer object.
wenet/transformer/decoder_layer.py:44
Method__init__
( self, vocab_size: int, encoder: BaseEncoder, decoder: TransformerDecoder,
wenet/transformer/asr_model.py:38
Method__init__
(self, idim: int, odim: int, dropout_rate: float, pos_enc_class: torch.nn.Module)
wenet/transformer/subsampling.py:41
Method__init__
Construct an linear object.
wenet/transformer/subsampling.py:81
Method__init__
Construct an Conv1dSubsampling2 object.
wenet/transformer/subsampling.py:130
Method__init__
Construct an Conv2dSubsampling4 object.
wenet/transformer/subsampling.py:188
Method__init__
Construct an Conv2dSubsampling6 object.
wenet/transformer/subsampling.py:244
Method__init__
Construct an Conv2dSubsampling8 object.
wenet/transformer/subsampling.py:297
Method__init__
(self, idim: int, odim: int, dropout_rate: float,
wenet/transformer/subsampling.py:345
Method__init__
Construct an LabelSmoothingLoss object.
wenet/transformer/label_smoothing_loss.py:54
Method__init__
Construct CTC module Args: odim: dimension of outputs encoder_output_size: number of encoder projection units
wenet/transformer/ctc.py:24
← previousnext →601–700 of 1,198, ranked by callers