.. index:: pair: example; cnn_training_bf16.cpp
.. _doxid-cnn_training_bf16_8cpp-example:
cnn_training_bf16.cpp
=====================
This C++ API example demonstrates how to build an AlexNet model training using the bfloat16 data type. Annotated version: :ref:`CNN bf16 training example <doxid-cnn_training_bf16_cpp>`
This C++ API example demonstrates how to build an AlexNet model training using the bfloat16 data type. Annotated version: :ref:`CNN bf16 training example <doxid-cnn_training_bf16_cpp>`
.. ref-code-block:: cpp
/*******************************************************************************
* Copyright 2019-2022 Intel Corporation
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*******************************************************************************/
#include <cassert>
#include <cmath>
#include <iostream>
#include <stdexcept>
#include "oneapi/dnnl/dnnl.hpp"
#include "example_utils.hpp"
using namespace :ref:`dnnl <doxid-namespacednnl>`;
void simple_net(:ref:`engine::kind <doxid-structdnnl_1_1engine_1a2635da16314dcbdb9bd9ea431316bb1a>` engine_kind) {
using :ref:`tag <doxid-structdnnl_1_1memory_1a8e71077ed6a5f7fb7b3e6e1a5a2ecf3f>` = :ref:`memory::format_tag <doxid-structdnnl_1_1memory_1a8e71077ed6a5f7fb7b3e6e1a5a2ecf3f>`;
using :ref:`dt <doxid-structdnnl_1_1memory_1a8e83474ec3a50e08e37af76c8c075dce>` = :ref:`memory::data_type <doxid-structdnnl_1_1memory_1a8e83474ec3a50e08e37af76c8c075dce>`;
auto eng = :ref:`engine <doxid-structdnnl_1_1engine>`(engine_kind, 0);
:ref:`stream <doxid-structdnnl_1_1stream>` s(eng);
// Vector of primitives and their execute arguments
std::vector<primitive> net_fwd, net_bwd;
std::vector<std::unordered_map<int, memory>> net_fwd_args, net_bwd_args;
const int batch = 32;
// float data type is used for user data
std::vector<float> net_src(batch * 3 * 227 * 227);
// initializing non-zero values for src
for (size_t i = 0; i < net_src.size(); ++i)
net_src[i] = sinf((float)i);
// AlexNet: conv
// {batch, 3, 227, 227} (x) {96, 3, 11, 11} -> {batch, 96, 55, 55}
// strides: {4, 4}
:ref:`memory::dims <doxid-structdnnl_1_1memory_1afdd20764d58c0b517d5a31276672aeb8>` conv_src_tz = {batch, 3, 227, 227};
:ref:`memory::dims <doxid-structdnnl_1_1memory_1afdd20764d58c0b517d5a31276672aeb8>` conv_weights_tz = {96, 3, 11, 11};
:ref:`memory::dims <doxid-structdnnl_1_1memory_1afdd20764d58c0b517d5a31276672aeb8>` conv_bias_tz = {96};
:ref:`memory::dims <doxid-structdnnl_1_1memory_1afdd20764d58c0b517d5a31276672aeb8>` conv_dst_tz = {batch, 96, 55, 55};
:ref:`memory::dims <doxid-structdnnl_1_1memory_1afdd20764d58c0b517d5a31276672aeb8>` conv_strides = {4, 4};
:ref:`memory::dims <doxid-structdnnl_1_1memory_1afdd20764d58c0b517d5a31276672aeb8>` conv_padding = {0, 0};
// float data type is used for user data
std::vector<float> conv_weights(product(conv_weights_tz));
std::vector<float> conv_bias(product(conv_bias_tz));
// initializing non-zero values for weights and bias
for (size_t i = 0; i < conv_weights.size(); ++i)
conv_weights[i] = sinf((float)i);
for (size_t i = 0; i < conv_bias.size(); ++i)
conv_bias[i] = sinf((float)i);
// create memory for user data
auto conv_user_src_memory
= :ref:`memory <doxid-structdnnl_1_1memory>`({{conv_src_tz}, dt::f32, tag::nchw}, eng);
write_to_dnnl_memory(net_src.data(), conv_user_src_memory);
auto conv_user_weights_memory
= :ref:`memory <doxid-structdnnl_1_1memory>`({{conv_weights_tz}, dt::f32, tag::oihw}, eng);
write_to_dnnl_memory(conv_weights.data(), conv_user_weights_memory);
auto conv_user_bias_memory = :ref:`memory <doxid-structdnnl_1_1memory>`({{conv_bias_tz}, dt::f32, tag::x}, eng);
write_to_dnnl_memory(conv_bias.data(), conv_user_bias_memory);
// create memory descriptors for bfloat16 convolution data w/ no specified
// format tag(`any`)
// tag `any` lets a primitive(convolution in this case)
// chose the memory format preferred for best performance.
auto conv_src_md = :ref:`memory::desc <doxid-structdnnl_1_1memory_1_1desc>`({conv_src_tz}, :ref:`dt::bf16 <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725afe2904d9fb3b0f4a81c92b03dec11424>`, :ref:`tag::any <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725a100b8cad7cf2a56f6df78f171f97a1ec>`);
auto conv_weights_md = :ref:`memory::desc <doxid-structdnnl_1_1memory_1_1desc>`({conv_weights_tz}, :ref:`dt::bf16 <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725afe2904d9fb3b0f4a81c92b03dec11424>`, :ref:`tag::any <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725a100b8cad7cf2a56f6df78f171f97a1ec>`);
auto conv_dst_md = :ref:`memory::desc <doxid-structdnnl_1_1memory_1_1desc>`({conv_dst_tz}, :ref:`dt::bf16 <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725afe2904d9fb3b0f4a81c92b03dec11424>`, :ref:`tag::any <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725a100b8cad7cf2a56f6df78f171f97a1ec>`);
// here bias data type is set to bf16.
// additionally, f32 data type is supported for bf16 convolution.
auto conv_bias_md = :ref:`memory::desc <doxid-structdnnl_1_1memory_1_1desc>`({conv_bias_tz}, :ref:`dt::bf16 <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725afe2904d9fb3b0f4a81c92b03dec11424>`, :ref:`tag::any <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725a100b8cad7cf2a56f6df78f171f97a1ec>`);
// create a convolution primitive descriptor
// check if bf16 convolution is supported
try {
:ref:`convolution_forward::primitive_desc <doxid-structdnnl_1_1convolution__forward_1_1primitive__desc>`(eng, :ref:`prop_kind::forward <doxid-group__dnnl__api__attributes_1ggac7db48f6583aa9903e54c2a39d65438fa965dbaac085fc891bfbbd4f9d145bbc8>`,
:ref:`algorithm::convolution_direct <doxid-group__dnnl__api__attributes_1gga00377dd4982333e42e8ae1d09a309640a5028ad8f818a45333a8a0eefad35c5c0>`, conv_src_md, conv_weights_md,
conv_bias_md, conv_dst_md, conv_strides, conv_padding,
conv_padding);
} catch (:ref:`error <doxid-structdnnl_1_1error>` &e) {
if (e.status == :ref:`dnnl_unimplemented <doxid-group__dnnl__api__utils_1ggad24f9ded06e34d3ee71e7fc4b408d57aa3a8579e8afc4e23344cd3115b0e81de1>`)
throw example_allows_unimplemented {
"No bf16 convolution implementation is available for this "
"platform.\n"
"Please refer to the developer guide for details."};
// on any other error just re-throw
throw;
}
auto conv_pd = :ref:`convolution_forward::primitive_desc <doxid-structdnnl_1_1convolution__forward_1_1primitive__desc>`(eng, :ref:`prop_kind::forward <doxid-group__dnnl__api__attributes_1ggac7db48f6583aa9903e54c2a39d65438fa965dbaac085fc891bfbbd4f9d145bbc8>`,
:ref:`algorithm::convolution_direct <doxid-group__dnnl__api__attributes_1gga00377dd4982333e42e8ae1d09a309640a5028ad8f818a45333a8a0eefad35c5c0>`, conv_src_md, conv_weights_md,
conv_bias_md, conv_dst_md, conv_strides, conv_padding,
conv_padding);
// create reorder primitives between user input and conv src if needed
auto conv_src_memory = conv_user_src_memory;
if (conv_pd.src_desc() != conv_user_src_memory.get_desc()) {
conv_src_memory = :ref:`memory <doxid-structdnnl_1_1memory>`(conv_pd.src_desc(), eng);
net_fwd.push_back(:ref:`reorder <doxid-structdnnl_1_1reorder>`(conv_user_src_memory, conv_src_memory));
net_fwd_args.push_back({{:ref:`DNNL_ARG_FROM <doxid-group__dnnl__api__primitives__common_1ga953b34f004a8222b04e21851487c611a>`, conv_user_src_memory},
{:ref:`DNNL_ARG_TO <doxid-group__dnnl__api__primitives__common_1gaf700c3396987b450413c8df5d78bafd9>`, conv_src_memory}});
}
auto conv_weights_memory = conv_user_weights_memory;
if (conv_pd.weights_desc() != conv_user_weights_memory.get_desc()) {
conv_weights_memory = :ref:`memory <doxid-structdnnl_1_1memory>`(conv_pd.weights_desc(), eng);
net_fwd.push_back(
:ref:`reorder <doxid-structdnnl_1_1reorder>`(conv_user_weights_memory, conv_weights_memory));
net_fwd_args.push_back({{:ref:`DNNL_ARG_FROM <doxid-group__dnnl__api__primitives__common_1ga953b34f004a8222b04e21851487c611a>`, conv_user_weights_memory},
{:ref:`DNNL_ARG_TO <doxid-group__dnnl__api__primitives__common_1gaf700c3396987b450413c8df5d78bafd9>`, conv_weights_memory}});
}
// convert bias from f32 to bf16 as convolution descriptor is created with
// bias data type as bf16.
auto conv_bias_memory = conv_user_bias_memory;
if (conv_pd.bias_desc() != conv_user_bias_memory.get_desc()) {
conv_bias_memory = :ref:`memory <doxid-structdnnl_1_1memory>`(conv_pd.bias_desc(), eng);
net_fwd.push_back(:ref:`reorder <doxid-structdnnl_1_1reorder>`(conv_user_bias_memory, conv_bias_memory));
net_fwd_args.push_back({{:ref:`DNNL_ARG_FROM <doxid-group__dnnl__api__primitives__common_1ga953b34f004a8222b04e21851487c611a>`, conv_user_bias_memory},
{:ref:`DNNL_ARG_TO <doxid-group__dnnl__api__primitives__common_1gaf700c3396987b450413c8df5d78bafd9>`, conv_bias_memory}});
}
// create memory for conv dst
auto conv_dst_memory = :ref:`memory <doxid-structdnnl_1_1memory>`(conv_pd.dst_desc(), eng);
// finally create a convolution primitive
net_fwd.push_back(:ref:`convolution_forward <doxid-structdnnl_1_1convolution__forward>`(conv_pd));
net_fwd_args.push_back({{:ref:`DNNL_ARG_SRC <doxid-group__dnnl__api__primitives__common_1gac37ad67b48edeb9e742af0e50b70fe09>`, conv_src_memory},
{:ref:`DNNL_ARG_WEIGHTS <doxid-group__dnnl__api__primitives__common_1gaf279f28c59a807e71a70c719db56c5b3>`, conv_weights_memory},
{:ref:`DNNL_ARG_BIAS <doxid-group__dnnl__api__primitives__common_1gad0cbc09942aba93fbe3c0c2e09166f0d>`, conv_bias_memory},
{:ref:`DNNL_ARG_DST <doxid-group__dnnl__api__primitives__common_1ga3ca217e4a06d42a0ede3c018383c388f>`, conv_dst_memory}});
// AlexNet: relu
// {batch, 96, 55, 55} -> {batch, 96, 55, 55}
:ref:`memory::dims <doxid-structdnnl_1_1memory_1afdd20764d58c0b517d5a31276672aeb8>` relu_data_tz = {batch, 96, 55, 55};
const float negative_slope = 0.0f;
// create relu primitive desc
// keep memory format tag of source same as the format tag of convolution
// output in order to avoid reorder
auto relu_pd = :ref:`eltwise_forward::primitive_desc <doxid-structdnnl_1_1eltwise__forward_1_1primitive__desc>`(eng, :ref:`prop_kind::forward <doxid-group__dnnl__api__attributes_1ggac7db48f6583aa9903e54c2a39d65438fa965dbaac085fc891bfbbd4f9d145bbc8>`,
:ref:`algorithm::eltwise_relu <doxid-group__dnnl__api__attributes_1gga00377dd4982333e42e8ae1d09a309640aba09bebb742494255b90b43871c01c69>`, conv_pd.dst_desc(), conv_pd.dst_desc(),
negative_slope);
// create relu dst memory
auto relu_dst_memory = :ref:`memory <doxid-structdnnl_1_1memory>`(relu_pd.dst_desc(), eng);
// finally create a relu primitive
net_fwd.push_back(:ref:`eltwise_forward <doxid-structdnnl_1_1eltwise__forward>`(relu_pd));
net_fwd_args.push_back(
{{:ref:`DNNL_ARG_SRC <doxid-group__dnnl__api__primitives__common_1gac37ad67b48edeb9e742af0e50b70fe09>`, conv_dst_memory}, {:ref:`DNNL_ARG_DST <doxid-group__dnnl__api__primitives__common_1ga3ca217e4a06d42a0ede3c018383c388f>`, relu_dst_memory}});
// AlexNet: lrn
// {batch, 96, 55, 55} -> {batch, 96, 55, 55}
// local size: 5
// alpha: 0.0001
// beta: 0.75
// k: 1.0
:ref:`memory::dims <doxid-structdnnl_1_1memory_1afdd20764d58c0b517d5a31276672aeb8>` lrn_data_tz = {batch, 96, 55, 55};
const uint32_t local_size = 5;
const float alpha = 0.0001f;
const float beta = 0.75f;
const float k = 1.0f;
// create a lrn primitive descriptor
auto lrn_pd = :ref:`lrn_forward::primitive_desc <doxid-structdnnl_1_1lrn__forward_1_1primitive__desc>`(eng, :ref:`prop_kind::forward <doxid-group__dnnl__api__attributes_1ggac7db48f6583aa9903e54c2a39d65438fa965dbaac085fc891bfbbd4f9d145bbc8>`,
:ref:`algorithm::lrn_across_channels <doxid-group__dnnl__api__attributes_1gga00377dd4982333e42e8ae1d09a309640ab9e2d858b551792385a4b5b86672b24b>`, relu_pd.dst_desc(),
relu_pd.dst_desc(), local_size, alpha, beta, k);
// create lrn dst memory
auto lrn_dst_memory = :ref:`memory <doxid-structdnnl_1_1memory>`(lrn_pd.dst_desc(), eng);
// create workspace only in training and only for forward primitive
// query lrn_pd for workspace, this memory will be shared with forward lrn
auto lrn_workspace_memory = :ref:`memory <doxid-structdnnl_1_1memory>`(lrn_pd.workspace_desc(), eng);
// finally create a lrn primitive
net_fwd.push_back(:ref:`lrn_forward <doxid-structdnnl_1_1lrn__forward>`(lrn_pd));
net_fwd_args.push_back(
{{:ref:`DNNL_ARG_SRC <doxid-group__dnnl__api__primitives__common_1gac37ad67b48edeb9e742af0e50b70fe09>`, relu_dst_memory}, {:ref:`DNNL_ARG_DST <doxid-group__dnnl__api__primitives__common_1ga3ca217e4a06d42a0ede3c018383c388f>`, lrn_dst_memory},
{:ref:`DNNL_ARG_WORKSPACE <doxid-group__dnnl__api__primitives__common_1ga550c80e1b9ba4f541202a7ac98be117f>`, lrn_workspace_memory}});
// AlexNet: pool
// {batch, 96, 55, 55} -> {batch, 96, 27, 27}
// kernel: {3, 3}
// strides: {2, 2}
:ref:`memory::dims <doxid-structdnnl_1_1memory_1afdd20764d58c0b517d5a31276672aeb8>` pool_dst_tz = {batch, 96, 27, 27};
:ref:`memory::dims <doxid-structdnnl_1_1memory_1afdd20764d58c0b517d5a31276672aeb8>` pool_kernel = {3, 3};
:ref:`memory::dims <doxid-structdnnl_1_1memory_1afdd20764d58c0b517d5a31276672aeb8>` pool_strides = {2, 2};
:ref:`memory::dims <doxid-structdnnl_1_1memory_1afdd20764d58c0b517d5a31276672aeb8>` pool_dilation = {0, 0};
:ref:`memory::dims <doxid-structdnnl_1_1memory_1afdd20764d58c0b517d5a31276672aeb8>` pool_padding = {0, 0};
// create memory for pool dst data in user format
auto pool_user_dst_memory
= :ref:`memory <doxid-structdnnl_1_1memory>`({{pool_dst_tz}, dt::f32, tag::nchw}, eng);
// create pool dst memory descriptor in format any for bfloat16 data type
auto pool_dst_md = :ref:`memory::desc <doxid-structdnnl_1_1memory_1_1desc>`({pool_dst_tz}, :ref:`dt::bf16 <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725afe2904d9fb3b0f4a81c92b03dec11424>`, :ref:`tag::any <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725a100b8cad7cf2a56f6df78f171f97a1ec>`);
// create a pooling primitive descriptor
auto pool_pd = :ref:`pooling_forward::primitive_desc <doxid-structdnnl_1_1pooling__forward_1_1primitive__desc>`(eng, :ref:`prop_kind::forward <doxid-group__dnnl__api__attributes_1ggac7db48f6583aa9903e54c2a39d65438fa965dbaac085fc891bfbbd4f9d145bbc8>`,
:ref:`algorithm::pooling_max <doxid-group__dnnl__api__attributes_1gga00377dd4982333e42e8ae1d09a309640a8c73d4bb88a0497586a74256bb338e88>`, lrn_dst_memory.get_desc(), pool_dst_md,
pool_strides, pool_kernel, pool_dilation, pool_padding,
pool_padding);
// create pooling workspace memory if training
auto pool_workspace_memory = :ref:`memory <doxid-structdnnl_1_1memory>`(pool_pd.workspace_desc(), eng);
// create a pooling primitive
net_fwd.push_back(:ref:`pooling_forward <doxid-structdnnl_1_1pooling__forward>`(pool_pd));
// leave DST unknown for now (see the next reorder)
net_fwd_args.push_back({{:ref:`DNNL_ARG_SRC <doxid-group__dnnl__api__primitives__common_1gac37ad67b48edeb9e742af0e50b70fe09>`, lrn_dst_memory},
// delay putting DST until reorder (if needed)
{:ref:`DNNL_ARG_WORKSPACE <doxid-group__dnnl__api__primitives__common_1ga550c80e1b9ba4f541202a7ac98be117f>`, pool_workspace_memory}});
// create reorder primitive between pool dst and user dst format
// if needed
auto pool_dst_memory = pool_user_dst_memory;
if (pool_pd.dst_desc() != pool_user_dst_memory.get_desc()) {
pool_dst_memory = :ref:`memory <doxid-structdnnl_1_1memory>`(pool_pd.dst_desc(), eng);
net_fwd_args.back().insert({:ref:`DNNL_ARG_DST <doxid-group__dnnl__api__primitives__common_1ga3ca217e4a06d42a0ede3c018383c388f>`, pool_dst_memory});
net_fwd.push_back(:ref:`reorder <doxid-structdnnl_1_1reorder>`(pool_dst_memory, pool_user_dst_memory));
net_fwd_args.push_back({{:ref:`DNNL_ARG_FROM <doxid-group__dnnl__api__primitives__common_1ga953b34f004a8222b04e21851487c611a>`, pool_dst_memory},
{:ref:`DNNL_ARG_TO <doxid-group__dnnl__api__primitives__common_1gaf700c3396987b450413c8df5d78bafd9>`, pool_user_dst_memory}});
} else {
net_fwd_args.back().insert({:ref:`DNNL_ARG_DST <doxid-group__dnnl__api__primitives__common_1ga3ca217e4a06d42a0ede3c018383c388f>`, pool_dst_memory});
}
//-----------------------------------------------------------------------
//----------------- Backward Stream -------------------------------------
// ... user diff_data in float data type ...
std::vector<float> net_diff_dst(batch * 96 * 27 * 27);
for (size_t i = 0; i < net_diff_dst.size(); ++i)
net_diff_dst[i] = sinf((float)i);
// create memory for user diff dst data stored in float data type
auto pool_user_diff_dst_memory
= :ref:`memory <doxid-structdnnl_1_1memory>`({{pool_dst_tz}, dt::f32, tag::nchw}, eng);
write_to_dnnl_memory(net_diff_dst.data(), pool_user_diff_dst_memory);
// Backward pooling
// create memory descriptors for pooling
auto pool_diff_src_md = :ref:`memory::desc <doxid-structdnnl_1_1memory_1_1desc>`({lrn_data_tz}, :ref:`dt::bf16 <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725afe2904d9fb3b0f4a81c92b03dec11424>`, :ref:`tag::any <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725a100b8cad7cf2a56f6df78f171f97a1ec>`);
auto pool_diff_dst_md = :ref:`memory::desc <doxid-structdnnl_1_1memory_1_1desc>`({pool_dst_tz}, :ref:`dt::bf16 <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725afe2904d9fb3b0f4a81c92b03dec11424>`, :ref:`tag::any <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725a100b8cad7cf2a56f6df78f171f97a1ec>`);
// backward primitive descriptor needs to hint forward descriptor
auto pool_bwd_pd = :ref:`pooling_backward::primitive_desc <doxid-structdnnl_1_1pooling__backward_1_1primitive__desc>`(eng,
:ref:`algorithm::pooling_max <doxid-group__dnnl__api__attributes_1gga00377dd4982333e42e8ae1d09a309640a8c73d4bb88a0497586a74256bb338e88>`, pool_diff_src_md, pool_diff_dst_md,
pool_strides, pool_kernel, pool_dilation, pool_padding,
pool_padding, pool_pd);
// create reorder primitive between user diff dst and pool diff dst
// if required
auto pool_diff_dst_memory = pool_user_diff_dst_memory;
if (pool_dst_memory.get_desc() != pool_user_diff_dst_memory.get_desc()) {
pool_diff_dst_memory = :ref:`memory <doxid-structdnnl_1_1memory>`(pool_dst_memory.get_desc(), eng);
net_bwd.push_back(
:ref:`reorder <doxid-structdnnl_1_1reorder>`(pool_user_diff_dst_memory, pool_diff_dst_memory));
net_bwd_args.push_back({{:ref:`DNNL_ARG_FROM <doxid-group__dnnl__api__primitives__common_1ga953b34f004a8222b04e21851487c611a>`, pool_user_diff_dst_memory},
{:ref:`DNNL_ARG_TO <doxid-group__dnnl__api__primitives__common_1gaf700c3396987b450413c8df5d78bafd9>`, pool_diff_dst_memory}});
}
// create memory for pool diff src
auto pool_diff_src_memory = :ref:`memory <doxid-structdnnl_1_1memory>`(pool_bwd_pd.diff_src_desc(), eng);
// finally create backward pooling primitive
net_bwd.push_back(:ref:`pooling_backward <doxid-structdnnl_1_1pooling__backward>`(pool_bwd_pd));
net_bwd_args.push_back({{:ref:`DNNL_ARG_DIFF_DST <doxid-group__dnnl__api__primitives__common_1gac9302f4cbd2668bf9a98ba99d752b971>`, pool_diff_dst_memory},
{:ref:`DNNL_ARG_DIFF_SRC <doxid-group__dnnl__api__primitives__common_1ga18ee0e360399cfe9d3b58a13dfcb9333>`, pool_diff_src_memory},
{:ref:`DNNL_ARG_WORKSPACE <doxid-group__dnnl__api__primitives__common_1ga550c80e1b9ba4f541202a7ac98be117f>`, pool_workspace_memory}});
// Backward lrn
auto lrn_diff_dst_md = :ref:`memory::desc <doxid-structdnnl_1_1memory_1_1desc>`({lrn_data_tz}, :ref:`dt::bf16 <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725afe2904d9fb3b0f4a81c92b03dec11424>`, :ref:`tag::any <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725a100b8cad7cf2a56f6df78f171f97a1ec>`);
const auto &lrn_diff_src_md = lrn_diff_dst_md;
// create backward lrn primitive descriptor
auto lrn_bwd_pd = :ref:`lrn_backward::primitive_desc <doxid-structdnnl_1_1lrn__backward_1_1primitive__desc>`(eng,
:ref:`algorithm::lrn_across_channels <doxid-group__dnnl__api__attributes_1gga00377dd4982333e42e8ae1d09a309640ab9e2d858b551792385a4b5b86672b24b>`, lrn_diff_src_md, lrn_diff_dst_md,
lrn_pd.src_desc(), local_size, alpha, beta, k, lrn_pd);
// create reorder primitive between pool diff src and lrn diff dst
// if required
auto lrn_diff_dst_memory = pool_diff_src_memory;
if (lrn_diff_dst_memory.get_desc() != lrn_bwd_pd.diff_dst_desc()) {
lrn_diff_dst_memory = :ref:`memory <doxid-structdnnl_1_1memory>`(lrn_bwd_pd.diff_dst_desc(), eng);
net_bwd.push_back(:ref:`reorder <doxid-structdnnl_1_1reorder>`(pool_diff_src_memory, lrn_diff_dst_memory));
net_bwd_args.push_back({{:ref:`DNNL_ARG_FROM <doxid-group__dnnl__api__primitives__common_1ga953b34f004a8222b04e21851487c611a>`, pool_diff_src_memory},
{:ref:`DNNL_ARG_TO <doxid-group__dnnl__api__primitives__common_1gaf700c3396987b450413c8df5d78bafd9>`, lrn_diff_dst_memory}});
}
// create memory for lrn diff src
auto lrn_diff_src_memory = :ref:`memory <doxid-structdnnl_1_1memory>`(lrn_bwd_pd.diff_src_desc(), eng);
// finally create a lrn backward primitive
// backward lrn needs src: relu dst in this topology
net_bwd.push_back(:ref:`lrn_backward <doxid-structdnnl_1_1lrn__backward>`(lrn_bwd_pd));
net_bwd_args.push_back({{:ref:`DNNL_ARG_SRC <doxid-group__dnnl__api__primitives__common_1gac37ad67b48edeb9e742af0e50b70fe09>`, relu_dst_memory},
{:ref:`DNNL_ARG_DIFF_DST <doxid-group__dnnl__api__primitives__common_1gac9302f4cbd2668bf9a98ba99d752b971>`, lrn_diff_dst_memory},
{:ref:`DNNL_ARG_DIFF_SRC <doxid-group__dnnl__api__primitives__common_1ga18ee0e360399cfe9d3b58a13dfcb9333>`, lrn_diff_src_memory},
{:ref:`DNNL_ARG_WORKSPACE <doxid-group__dnnl__api__primitives__common_1ga550c80e1b9ba4f541202a7ac98be117f>`, lrn_workspace_memory}});
// Backward relu
auto relu_diff_src_md = :ref:`memory::desc <doxid-structdnnl_1_1memory_1_1desc>`({relu_data_tz}, :ref:`dt::bf16 <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725afe2904d9fb3b0f4a81c92b03dec11424>`, :ref:`tag::any <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725a100b8cad7cf2a56f6df78f171f97a1ec>`);
auto relu_diff_dst_md = :ref:`memory::desc <doxid-structdnnl_1_1memory_1_1desc>`({relu_data_tz}, :ref:`dt::bf16 <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725afe2904d9fb3b0f4a81c92b03dec11424>`, :ref:`tag::any <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725a100b8cad7cf2a56f6df78f171f97a1ec>`);
auto relu_src_md = conv_pd.dst_desc();
// create backward relu primitive_descriptor
auto relu_bwd_pd = :ref:`eltwise_backward::primitive_desc <doxid-structdnnl_1_1eltwise__backward_1_1primitive__desc>`(eng,
:ref:`algorithm::eltwise_relu <doxid-group__dnnl__api__attributes_1gga00377dd4982333e42e8ae1d09a309640aba09bebb742494255b90b43871c01c69>`, relu_diff_src_md, relu_diff_dst_md,
relu_src_md, negative_slope, relu_pd);
// create reorder primitive between lrn diff src and relu diff dst
// if required
auto relu_diff_dst_memory = lrn_diff_src_memory;
if (relu_diff_dst_memory.get_desc() != relu_bwd_pd.diff_dst_desc()) {
relu_diff_dst_memory = :ref:`memory <doxid-structdnnl_1_1memory>`(relu_bwd_pd.diff_dst_desc(), eng);
net_bwd.push_back(:ref:`reorder <doxid-structdnnl_1_1reorder>`(lrn_diff_src_memory, relu_diff_dst_memory));
net_bwd_args.push_back({{:ref:`DNNL_ARG_FROM <doxid-group__dnnl__api__primitives__common_1ga953b34f004a8222b04e21851487c611a>`, lrn_diff_src_memory},
{:ref:`DNNL_ARG_TO <doxid-group__dnnl__api__primitives__common_1gaf700c3396987b450413c8df5d78bafd9>`, relu_diff_dst_memory}});
}
// create memory for relu diff src
auto relu_diff_src_memory = :ref:`memory <doxid-structdnnl_1_1memory>`(relu_bwd_pd.diff_src_desc(), eng);
// finally create a backward relu primitive
net_bwd.push_back(:ref:`eltwise_backward <doxid-structdnnl_1_1eltwise__backward>`(relu_bwd_pd));
net_bwd_args.push_back({{:ref:`DNNL_ARG_SRC <doxid-group__dnnl__api__primitives__common_1gac37ad67b48edeb9e742af0e50b70fe09>`, conv_dst_memory},
{:ref:`DNNL_ARG_DIFF_DST <doxid-group__dnnl__api__primitives__common_1gac9302f4cbd2668bf9a98ba99d752b971>`, relu_diff_dst_memory},
{:ref:`DNNL_ARG_DIFF_SRC <doxid-group__dnnl__api__primitives__common_1ga18ee0e360399cfe9d3b58a13dfcb9333>`, relu_diff_src_memory}});
// Backward convolution with respect to weights
// create user format diff weights and diff bias memory for float data type
auto conv_user_diff_weights_memory
= :ref:`memory <doxid-structdnnl_1_1memory>`({{conv_weights_tz}, dt::f32, tag::nchw}, eng);
auto conv_diff_bias_memory = :ref:`memory <doxid-structdnnl_1_1memory>`({{conv_bias_tz}, dt::f32, tag::x}, eng);
// create memory descriptors for bfloat16 convolution data
auto conv_bwd_src_md = :ref:`memory::desc <doxid-structdnnl_1_1memory_1_1desc>`({conv_src_tz}, :ref:`dt::bf16 <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725afe2904d9fb3b0f4a81c92b03dec11424>`, :ref:`tag::any <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725a100b8cad7cf2a56f6df78f171f97a1ec>`);
auto conv_diff_weights_md
= :ref:`memory::desc <doxid-structdnnl_1_1memory_1_1desc>`({conv_weights_tz}, :ref:`dt::bf16 <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725afe2904d9fb3b0f4a81c92b03dec11424>`, :ref:`tag::any <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725a100b8cad7cf2a56f6df78f171f97a1ec>`);
auto conv_diff_dst_md = :ref:`memory::desc <doxid-structdnnl_1_1memory_1_1desc>`({conv_dst_tz}, :ref:`dt::bf16 <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725afe2904d9fb3b0f4a81c92b03dec11424>`, :ref:`tag::any <doxid-group__dnnl__api__fpmath__mode_1gga0ad94cbef13dce222933422bfdcfa725a100b8cad7cf2a56f6df78f171f97a1ec>`);
// use diff bias provided by the user
auto conv_diff_bias_md = conv_diff_bias_memory.:ref:`get_desc <doxid-structdnnl_1_1memory_1ad8a1ad28ed7acf9c34c69e4b882c6e92>`();
// create backward convolution primitive descriptor
auto conv_bwd_weights_pd = :ref:`convolution_backward_weights::primitive_desc <doxid-structdnnl_1_1convolution__backward__weights_1_1primitive__desc>`(eng,
:ref:`algorithm::convolution_direct <doxid-group__dnnl__api__attributes_1gga00377dd4982333e42e8ae1d09a309640a5028ad8f818a45333a8a0eefad35c5c0>`, conv_bwd_src_md,
conv_diff_weights_md, conv_diff_bias_md, conv_diff_dst_md,
conv_strides, conv_padding, conv_padding, conv_pd);
// for best performance convolution backward might chose
// different memory format for src and diff_dst
// than the memory formats preferred by forward convolution
// for src and dst respectively
// create reorder primitives for src from forward convolution to the
// format chosen by backward convolution
auto conv_bwd_src_memory = conv_src_memory;
if (conv_bwd_weights_pd.src_desc() != conv_src_memory.get_desc()) {
conv_bwd_src_memory = :ref:`memory <doxid-structdnnl_1_1memory>`(conv_bwd_weights_pd.src_desc(), eng);
net_bwd.push_back(:ref:`reorder <doxid-structdnnl_1_1reorder>`(conv_src_memory, conv_bwd_src_memory));
net_bwd_args.push_back({{:ref:`DNNL_ARG_FROM <doxid-group__dnnl__api__primitives__common_1ga953b34f004a8222b04e21851487c611a>`, conv_src_memory},
{:ref:`DNNL_ARG_TO <doxid-group__dnnl__api__primitives__common_1gaf700c3396987b450413c8df5d78bafd9>`, conv_bwd_src_memory}});
}
// create reorder primitives for diff_dst between diff_src from relu_bwd
// and format preferred by conv_diff_weights
auto conv_diff_dst_memory = relu_diff_src_memory;
if (conv_bwd_weights_pd.diff_dst_desc()
!= relu_diff_src_memory.get_desc()) {
conv_diff_dst_memory = :ref:`memory <doxid-structdnnl_1_1memory>`(conv_bwd_weights_pd.diff_dst_desc(), eng);
net_bwd.push_back(:ref:`reorder <doxid-structdnnl_1_1reorder>`(relu_diff_src_memory, conv_diff_dst_memory));
net_bwd_args.push_back({{:ref:`DNNL_ARG_FROM <doxid-group__dnnl__api__primitives__common_1ga953b34f004a8222b04e21851487c611a>`, relu_diff_src_memory},
{:ref:`DNNL_ARG_TO <doxid-group__dnnl__api__primitives__common_1gaf700c3396987b450413c8df5d78bafd9>`, conv_diff_dst_memory}});
}
// create backward convolution primitive
net_bwd.push_back(:ref:`convolution_backward_weights <doxid-structdnnl_1_1convolution__backward__weights>`(conv_bwd_weights_pd));
net_bwd_args.push_back({{:ref:`DNNL_ARG_SRC <doxid-group__dnnl__api__primitives__common_1gac37ad67b48edeb9e742af0e50b70fe09>`, conv_bwd_src_memory},
{:ref:`DNNL_ARG_DIFF_DST <doxid-group__dnnl__api__primitives__common_1gac9302f4cbd2668bf9a98ba99d752b971>`, conv_diff_dst_memory},
// delay putting DIFF_WEIGHTS until reorder (if needed)
{:ref:`DNNL_ARG_DIFF_BIAS <doxid-group__dnnl__api__primitives__common_1ga1cd79979dda6df65ec45eef32a839901>`, conv_diff_bias_memory}});
// create reorder primitives between conv diff weights and user diff weights
// if needed
auto conv_diff_weights_memory = conv_user_diff_weights_memory;
if (conv_bwd_weights_pd.diff_weights_desc()
!= conv_user_diff_weights_memory.get_desc()) {
conv_diff_weights_memory
= :ref:`memory <doxid-structdnnl_1_1memory>`(conv_bwd_weights_pd.diff_weights_desc(), eng);
net_bwd_args.back().insert(
{:ref:`DNNL_ARG_DIFF_WEIGHTS <doxid-group__dnnl__api__primitives__common_1ga3324092ef421f77aebee83b0117cac60>`, conv_diff_weights_memory});
net_bwd.push_back(:ref:`reorder <doxid-structdnnl_1_1reorder>`(
conv_diff_weights_memory, conv_user_diff_weights_memory));
net_bwd_args.push_back({{:ref:`DNNL_ARG_FROM <doxid-group__dnnl__api__primitives__common_1ga953b34f004a8222b04e21851487c611a>`, conv_diff_weights_memory},
{:ref:`DNNL_ARG_TO <doxid-group__dnnl__api__primitives__common_1gaf700c3396987b450413c8df5d78bafd9>`, conv_user_diff_weights_memory}});
} else {
net_bwd_args.back().insert(
{:ref:`DNNL_ARG_DIFF_WEIGHTS <doxid-group__dnnl__api__primitives__common_1ga3324092ef421f77aebee83b0117cac60>`, conv_diff_weights_memory});
}
// didn't we forget anything?
assert(net_fwd.size() == net_fwd_args.size() && "something is missing");
assert(net_bwd.size() == net_bwd_args.size() && "something is missing");
int n_iter = 1; // number of iterations for training
// execute
while (n_iter) {
// forward
for (size_t i = 0; i < net_fwd.size(); ++i)
net_fwd.at(i).execute(s, net_fwd_args.at(i));
// update net_diff_dst
// auto net_output = pool_user_dst_memory.get_data_handle();
// ..user updates net_diff_dst using net_output...
// some user defined func update_diff_dst(net_diff_dst.data(),
// net_output)
for (size_t i = 0; i < net_bwd.size(); ++i)
net_bwd.at(i).execute(s, net_bwd_args.at(i));
// update weights and bias using diff weights and bias
//
// auto net_diff_weights
// = conv_user_diff_weights_memory.get_data_handle();
// auto net_diff_bias = conv_diff_bias_memory.get_data_handle();
//
// ...user updates weights and bias using diff weights and bias...
//
// some user defined func update_weights(conv_weights.data(),
// conv_bias.data(), net_diff_weights, net_diff_bias);
--n_iter;
}
s.wait();
}
int main(int argc, char **argv) {
return handle_example_errors(simple_net, parse_engine_kind(argc, argv));
}