LogCoshLoss

Logarithm of the hyperbolic cosine of the prediction error.

Abstract Signature:

LogCoshLoss(y_true: Tensor, y_pred: Tensor)

PyTorch

API: —
Strategy: Macro 'torch.log(torch.cosh({y_pred} - {y_true}))'

JAX (Core)

API: optax.log_cosh
Strategy: Direct Mapping
Official Docs ↗

Keras

API: keras.losses.log_cosh
Strategy: Direct Mapping
Official Docs ↗

TensorFlow

API: tf.keras.losses.log_cosh
Strategy: Direct Mapping
Official Docs ↗

Apple MLX

API: mlx.nn.losses.log_cosh_loss
Strategy: Direct Mapping
Official Docs ↗

Flax NNX

API: optax.log_cosh
Strategy: Direct Mapping
Official Docs ↗

PaxML / Praxis

API: optax.log_cosh
Strategy: Direct Mapping
Official Docs ↗