enum dnnl_prop_kind_t

Overview

Kinds of propagation. More…

#include <dnnl_types.h>

enum dnnl_prop_kind_t
{
    dnnl_prop_kind_undef   = 0,
    dnnl_forward_training  = 64,
    dnnl_forward_inference = 96,
    dnnl_forward_scoring   = dnnl_forward_inference,
    dnnl_forward           = dnnl_forward_training,
    dnnl_backward          = 128,
    dnnl_backward_data     = 160,
    dnnl_backward_weights  = 192,
    dnnl_backward_bias     = 193,
};

Detailed Documentation

Kinds of propagation.

Enum Values

dnnl_prop_kind_undef

Undefined propagation type.

dnnl_forward_training

Forward data propagation (training mode).

In this mode primitives perform computations necessary for subsequent backward propagation.

dnnl_forward_inference

Forward data propagation (inference mode).

In this mode primitives perform only computations that are necessary for inference and omit computations that are necessary only for backward propagation.

dnnl_forward_scoring

Forward data propagation (alias for dnnl_forward_inference).

dnnl_forward

Forward data propagation (alias for dnnl_forward_training).

dnnl_backward

Backward propagation (with respect to all parameters).

dnnl_backward_data

Backward data propagation.

dnnl_backward_weights

Backward weights propagation.

dnnl_backward_bias

Backward bias propagation.