paddlespeech.t2s.modules.transformer 包

子模块

• paddlespeech.t2s.modules.transformer.attention 模块
  • 遗留相对位置多头注意力
    • LegacyRelPositionMultiHeadedAttention.forward()
    • LegacyRelPositionMultiHeadedAttention.rel_shift()
  • MultiHeadedAttention
    • MultiHeadedAttention.forward()
    • MultiHeadedAttention.forward_attention()
    • MultiHeadedAttention.forward_qkv()
  • RelPositionMultiHeadedAttention
    • RelPositionMultiHeadedAttention.forward()
    • RelPositionMultiHeadedAttention.rel_shift()
• paddlespeech.t2s.modules.transformer.decoder 模块
  • 解码器
    • Decoder.batch_score()
    • Decoder.forward()
    • 解码器.前进一步()
    • Decoder.score()
• paddlespeech.t2s.modules.transformer.decoder_layer 模块
  • DecoderLayer
    • 解码器层.forward()
• paddlespeech.t2s.modules.transformer.embedding 模块
  • LegacyRelPositionalEncoding
    • LegacyRelPositionalEncoding.extend_pe()
    • LegacyRelPositionalEncoding.forward()
  • PositionalEncoding
    • PositionalEncoding.extend_pe()
    • 位置编码.forward()
  • RelPositionalEncoding
    • RelPositionalEncoding.extend_pe()
    • RelPositionalEncoding.forward()
  • ScaledPositionalEncoding
    • ScaledPositionalEncoding.forward()
    • ScaledPositionalEncoding.reset_parameters()
• paddlespeech.t2s.modules.transformer.encoder 模块
  • BaseEncoder
    • BaseEncoder.forward()
    • BaseEncoder.get_embed()
    • BaseEncoder.get_encoder_selfattn_layer()
    • BaseEncoder.get_pos_enc_class()
    • BaseEncoder.get_positionwise_layer()
  • CNNDecoder
    • CNNDecoder.forward()
  • CNNPostnet
    • CNNPostnet.forward()
  • ConformerEncoder
    • ConformerEncoder.forward()
  • Conv1dResidualBlock
    • Conv1dResidualBlock.forward()
  • TransformerEncoder
    • TransformerEncoder.forward()
    • TransformerEncoder.forward_one_step()
• paddlespeech.t2s.modules.transformer.encoder_layer 模块
  • EncoderLayer
    • EncoderLayer.forward()
• paddlespeech.t2s.modules.transformer.lightconv 模块
  • LightweightConvolution
    • LightweightConvolution.forward()
• paddlespeech.t2s.modules.transformer.mask 模块
  • subsequent_mask()
  • target_mask()
• paddlespeech.t2s.modules.transformer.multi_layer_conv 模块
  • Conv1dLinear
    • Conv1dLinear.forward()
  • MultiLayeredConv1d
    • MultiLayeredConv1d.forward()
• paddlespeech.t2s.modules.transformer.positionwise_feed_forward 模块
  • PositionwiseFeedForward
    • PositionwiseFeedForward.forward()
• paddlespeech.t2s.modules.transformer.repeat 模块
  • MultiSequential
    • MultiSequential.forward()
  • repeat()
• paddlespeech.t2s.modules.transformer.subsampling 模块
  • Conv2dSubsampling
    • Conv2dSubsampling.forward()