L2Normalize

Normalizes along dimension axis using an L2 norm.

Abstract Signature:

L2Normalize(x: Tensor, axis: int, epsilon: float = 1e-12)

PyTorchJAX (Core)KerasTensorFlowApple MLXFlax NNXPaxML / Praxis

PyTorch

API: torch.nn.functional.normalize

Strategy: Direct Mapping

Official Docs ↗

JAX (Core)

API: —

Strategy: Macro '{x} / jnp.sqrt(jnp.maximum(jnp.sum({x}**2, axis={axis}, keepdims=True), {epsilon}))'

Keras

API: keras.utils.normalize

Strategy: Direct Mapping

Official Docs ↗

TensorFlow

API: tf.math.l2_normalize

Strategy: Direct Mapping

Official Docs ↗

Apple MLX

API: —

Strategy: Macro '{x} / mx.sqrt(mx.maximum(mx.sum({x}**2, axis={axis}, keepdims=True), {epsilon}))'

Flax NNX

API: —

Strategy: Macro '{x} / jnp.sqrt(jnp.maximum(jnp.sum({x}**2, axis={axis}, keepdims=True), {epsilon}))'

PaxML / Praxis

API: —

Strategy: Macro '{x} / jnp.sqrt(jnp.maximum(jnp.sum({x}**2, axis={axis}, keepdims=True), {epsilon}))'