Skip to main content
from onyxengine.modeling import RNNConfig

config = RNNConfig(
    outputs: List[Output],
    inputs: List[Input],
    dt: float,
    sequence_length: int = 1,
    rnn_type: Literal['RNN', 'LSTM', 'GRU'] = 'LSTM',
    hidden_layers: int = 2,
    hidden_size: int = 32,
    dropout: float = 0.0,
    bias: bool = True
)
Configuration class for Recurrent Neural Network models.

​
Parameters

​
outputs
List[Output]
required
List of output feature definitions.
​
inputs
List[Input]
required
List of input feature definitions.
​
dt
float
required
Time step in seconds. Must match your dataset’s sampling rate.
​
sequence_length
int
default:"1"
Length of the input sequence. Range: 1-100.
​
rnn_type
Literal
default:"LSTM"
Type of recurrent unit. Options: 'RNN', 'LSTM', 'GRU'.
​
hidden_layers
int
default:"2"
Number of stacked RNN layers. Range: 1-10.
​
hidden_size
int
default:"32"
Number of hidden units per layer. Range: 1-1024.
​
dropout
float
default:"0.0"
Dropout rate between RNN layers. Range: 0.0-1.0.
​
bias
bool
default:"True"
Whether to include bias terms.

​
RNN Types

TypeDescriptionUse Case
'RNN'Basic recurrent unitSimple sequences, debugging
'LSTM'Long Short-Term MemoryLong-range dependencies (default)
'GRU'Gated Recurrent UnitSimilar to LSTM, fewer parameters

​
Example

from onyxengine.modeling import RNNConfig, Input, Output

outputs = [Output(name='acceleration')]
inputs = [
    Input(name='velocity', parent='acceleration', relation='derivative'),
    Input(name='position', parent='velocity', relation='derivative'),
    Input(name='control_input'),
]

config = RNNConfig(
    outputs=outputs,
    inputs=inputs,
    dt=0.01,
    sequence_length=12,
    rnn_type='LSTM',
    hidden_layers=2,
    hidden_size=64,
    dropout=0.1,
    bias=True
)

​
Architecture

Input: (batch, sequence_length, num_inputs)
  ↓ RNN/LSTM/GRU layers (with hidden state)
  ↓ Take final hidden state
  ↓ Linear(hidden_size, num_outputs)
Output: (batch, num_outputs)

​
RNNOptConfig

For hyperparameter optimization: 
from onyxengine.modeling import RNNOptConfig

rnn_opt = RNNOptConfig(
    outputs=outputs,
    inputs=inputs,
    dt=0.01,
    rnn_type={"select": ['RNN', 'LSTM', 'GRU']},
    sequence_length={"select": [4, 8, 12, 16]},
    hidden_layers={"range": [2, 4, 1]},
    hidden_size={"select": [32, 64, 128]},
    dropout={"range": [0.0, 0.4, 0.1]},
    bias=True
)