Skip to main content
Ctrl+K
Time-series-prediction 0.0.1 documentation - Home Time-series-prediction 0.0.1 documentation - Home
  • Quick start
  • Tutorials
  • Models
  • Tricks
  • API
Ctrl+K
  • Quick start
  • Tutorials
  • Models
  • Tricks
  • API
Ctrl+K

Section Navigation

  • layers
    • attention_layer
      • Attention
      • FastAttention
      • ProbAttention
      • SelfAttention
      • SparseAttention
    • autoformer_layer
      • AutoCorrelation
      • MovingAvg
      • SeriesDecomp
    • cnn_layer
      • ConvTemp
    • dense_layer
      • DenseTemp
      • FeedForwardNetwork
    • embed_layer
      • DataEmbedding
      • PatchEmbedding
      • TemporalEmbedding
      • TokenEmbedding
      • TokenRnnEmbedding
    • mask_layer
      • CausalMask
      • ProbMask
    • moe_layer
      • MoELayer
    • nbeats_layer
      • generic_model
      • seasonality_model
      • trend_model
      • GenericBlock
      • SeasonalityBlock
      • TrendBlock
    • position_layer
      • PositionalEmbedding
      • PositionalEncoding
      • RelativePositionEmbedding
      • RotaryPositionEmbedding
    • rwkv_layer
      • ChannelMixing
      • TimeMixing
    • unet_layer
      • ConvbrLayer
      • ReBlock
      • SeBlock
    • util_layer
      • CreateDecoderFeature
      • ShapeLayer
      • ZerosLayer
  • models
    • auto_config
      • AutoConfig
    • auto_model
      • AutoModel
      • AutoModelForAnomaly
      • AutoModelForClassification
      • AutoModelForPrediction
      • AutoModelForSegmentation
      • AutoModelForUncertainty
    • autoformer
      • AutoFormer
      • AutoFormerConfig
      • Decoder
      • DecoderLayer
      • Encoder
      • EncoderLayer
    • base
      • flatten_dict
      • BaseConfig
      • BaseModel
    • bert
      • Bert
      • BertConfig
    • deep_ar
      • DeepAR
      • DeepARConfig
    • diffusion
      • Diffusion
      • DiffusionConfig
      • NoiseScheduler
      • TransformerBlock
    • dlinear
      • DLinear
      • DLinearConfig
    • gpt
      • GPT
      • GPTConfig
    • informer
      • Decoder
      • DecoderLayer
      • DistilConv
      • Encoder
      • EncoderLayer
      • Informer
      • InformerConfig
    • nbeats
      • NBeats
      • NBeatsConfig
    • patches_tst
      • PatchTST
      • PatchTSTConfig
      • TransformerBlock
    • pfn
    • rnn
      • Encoder
      • RNN
      • RNN2
      • RNNConfig
    • rwkv
      • RWKV
      • RWKVBlock
      • RWKVConfig
    • seq2seq
      • DecoderV1
      • DecoderV2
      • Encoder
      • Seq2seq
      • Seq2seqConfig
    • tcn
      • Encoder
      • TCN
      • TCNConfig
    • tft
      • TFTransformer
      • TFTransformerConfig
    • tide
      • DenseEncoderBlock
      • Tide
      • TideConfig
    • timesfm
      • TimesFm
      • TimesFmConfig
      • TransformerBlock
    • transformer
      • Decoder
      • DecoderLayer
      • Encoder
      • Transformer
      • TransformerBlock
      • TransformerConfig
    • unet
      • Decoder
      • Encoder
      • Unet
      • UnetConfig
    • wavenet
      • DecoderV1
      • DecoderV2
      • Encoder
      • WaveNet
      • WaveNetConfig
  • trainer
    • BaseTrainer
    • KerasTrainer
    • Seq2seqKerasTrainer
    • Trainer
  • API
  • layers
  • embed_layer

embed_layer#

Layer for transformer

Classes

DataEmbedding(*args, **kwargs)

Data Embedding layer that combines value embeddings with optional positional embeddings.

PatchEmbedding(*args, **kwargs)

TemporalEmbedding(*args, **kwargs)

TokenEmbedding(*args, **kwargs)

A layer that performs token embedding, equivalent to a dense layer for time series data.

TokenRnnEmbedding(*args, **kwargs)

previous

FeedForwardNetwork

next

DataEmbedding

This Page

  • Show Source

© Copyright 2020, Longxing Tan.

Created using Sphinx 8.1.3.

Built with the PyData Sphinx Theme 0.16.1.