Bias

Simple learnable bias layer.

Abstract Signature:

Bias(dims: int)

PyTorch

API: torch.nn.modules.rnn.RNNCellBase.bias
Strategy: Direct Mapping
Official Docs ↗

Keras

API: keras.layers.Bias
Strategy: Macro 'lambda x: x + keras.Variable(shape=({dims},))'

Apple MLX

API: mlx.nn.layers.convolution_transpose.ConvTranspose3d.bias
Strategy: Direct Mapping
Official Docs ↗

Flax NNX

API: nnx.nn.normalization.InstanceNorm.bias
Strategy: Direct Mapping
Official Docs ↗

PaxML / Praxis

API: praxis.layers.Bias
Strategy: Direct Mapping
Official Docs ↗