FusedRmsNormOps

Optional interface that TensorOps implementations can provide to support fused RMS normalization without intermediate tensor allocations.

Functions

abstract fun <T : DType, V> fusedRmsNorm(input: Tensor<T, V>, weight: Tensor<T, V>, eps: Float): Tensor<T, V>?

Compute fused RMS normalization: outputi = (inputi / rms(input)) * weighti