Seq2Seq-Framework-with-Mechanisms

A Seq2Seq framework based on pytorch which implements a Seq2Seq model using rnn with below privileges:

  • Attention Mechanism
    • Encoder-Decoder Attention
    • Inter-Decoder Attention
  • Pointer Mechanism
  • Coverage Mechanism
  • Teacher Forcing
    • Raw
    • Scheduling
  • Different RNN Types
    • RNN
    • LSTM
    • GRU
  • Different Decoding Mechanisms
    • Beam Search
    • Sampling Methods
    • Temprature
  • Different Training Methods
    • MLE
    • RL