Run the cell on embedded inputs.

Run the input projection and then the cell.

Run the cell and output projection on inputs, starting from state.

LSTM cell with layer normalization and recurrent dropout.

Run this multi-layer cell on inputs, starting from state.

Run one step of LSTM. Args: inputs: input Tensor, 2D, batch x num_units. state: if `state_is_tuple` is False, this must be a state Te

Run one step of LSTM. Args: inputs: input Tensor, 2D, batch x num_units. state: state Tensor, 2D, batch x state_size. Returns:

Run one step of LSTM. Args: inputs: input Tensor, 2D, [batch, feature_size]. state: Tensor or tuple of Tensors, 2D, [batch, state_siz

Run one step of LSTM. Args: inputs: input Tensor, 2D, [batch, num_units]. state: tuple of Tensors, 2D, [batch, state_size]. Retu

Long short-term memory cell with attention (LSTMA).

LSTM cell with layer normalization and recurrent dropout.

LSTM cell with layer normalization and recurrent dropout.

Run one step of NAS Cell. Args: inputs: input Tensor, 2D, batch x num_units. state: This must be a tuple of state Tensors, both `2-D`

Run one step of UGRNN. Args: inputs: input Tensor, 2D, batch x input size. state: state Tensor, 2D, batch x num units. Returns:

Run one step of the Intersection RNN. Args: inputs: input Tensor, 2D, batch x input size. state: state Tensor, 2D, batch x num units.

Phased LSTM Cell. Args: inputs: A tuple of 2 Tensor. The first Tensor has shape [batch, 1], and type float32 or float64.

Run one step of G-LSTM. Args: inputs: input Tensor, 2D, [batch x num_units]. state: this must be a tuple of state Tensors, both `2-D`

fsr_data : one of LSMDCData.build_data(), [[video_features], [sentences], [roots]] return per iter : [[feature]*batch_size, [sentences]*batch

The dictionary size.

Number of units in this cell (output dimension).

Number of units in this cell (output dimension).

either Fill pad to video to have same length or if too long interpolate to max_length Pad in Left. video = [pad,..., pad, frm1, frm2, ...

Creates a dynamic bidirectional recurrent neural network. Stacks several bidirectional rnn layers. The combined forward and backward layer output

Creates a bidirectional recurrent neural network. Stacks several bidirectional rnn layers. The combined forward and backward layer outputs are us