Trait collenchyma_nn::Softmax [−] [src]

pub trait Softmax<F>: NN<F> {
    fn softmax(&self, x: &mut SharedTensor<F>, result: &mut SharedTensor<F>) -> Result<(), Error>;
    fn softmax_plain(&self, x: &SharedTensor<F>, result: &mut SharedTensor<F>) -> Result<(), Error>;
    fn softmax_grad(&self, x: &mut SharedTensor<F>, x_diff: &mut SharedTensor<F>, result_diff: &mut SharedTensor<F>) -> Result<(), Error>;
    fn softmax_grad_plain(&self, x: &SharedTensor<F>, x_diff: &SharedTensor<F>, result_diff: &mut SharedTensor<F>) -> Result<(), Error>;
}

Provides the functionality for a Backend to support Softmax operations.

Required Methods

fn softmax(&self, x: &mut SharedTensor<F>, result: &mut SharedTensor<F>) -> Result<(), Error>

Computes a Softmax over the input Tensor x with complete memory management.

Saves the result to result.

For a no-memory managed version see softmax_plain.

fn softmax_plain(&self, x: &SharedTensor<F>, result: &mut SharedTensor<F>) -> Result<(), Error>

Computes the softmax over the input Tensor x without any memory management.

Saves the result to result.

Attention:
For a correct computation result, you need to manage the memory allocation and synchronization yourself.
For a memory managed version see softmax.

fn softmax_grad(&self, x: &mut SharedTensor<F>, x_diff: &mut SharedTensor<F>, result_diff: &mut SharedTensor<F>) -> Result<(), Error>

Computes the gradient of a Softmax over the input Tensor x with complete memory management.

Saves the result to result_diff.

For a no-memory managed version see softmax_grad_plain.

fn softmax_grad_plain(&self, x: &SharedTensor<F>, x_diff: &SharedTensor<F>, result_diff: &mut SharedTensor<F>) -> Result<(), Error>

Computes the gradient of a softmax over the input Tensor x without any memory management.

Saves the result to result_diff.

Attention:
For a correct computation result, you need to manage the memory allocation and synchronization yourself.
For a memory managed version see softmax_grad.

Implementors