Function candle_nn::ops::rms_norm_slow

pub fn rms_norm_slow(x: &Tensor, alpha: &Tensor, eps: f32) -> Result<Tensor>