IntenseWebs/openvino
3
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

[TF FE] Align opset usage in utils (#16656)

Browse Source 
Signed-off-by: Kazantsev, Roman <roman.kazantsev@intel.com>
This commit is contained in:
Roman Kazantsev
2023-03-30 21:58:46 +04:00
committed by GitHub
parent e94f7b25c0
commit fc95d8e544
7 changed files with 25 additions and 27 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
src/frontends/tensorflow/src/op_table.cpp
View File

@@ -6,6 +6,7 @@
#include "common_op_table.hpp"
#include "openvino/opsets/opset10.hpp"
#include "openvino/opsets/opset8.hpp"
#include "openvino/opsets/opset9.hpp"
using namespace std;

1
src/frontends/tensorflow/src/translate_session.cpp
View File

@@ -6,6 +6,7 @@
#include "input_model.hpp"
#include "openvino/opsets/opset10.hpp"
#include "openvino/opsets/opset8.hpp"
#include "tf_framework_node.hpp"
#include "utils.hpp"

9
src/frontends/tensorflow_common/include/utils.hpp
View File

@@ -7,7 +7,6 @@
#include "openvino/core/validation_util.hpp"
#include "openvino/frontend/node_context.hpp"
#include "openvino/opsets/opset10.hpp"
#include "openvino/opsets/opset8.hpp"
#include "openvino/pass/graph_rewrite.hpp"
#ifndef TENSORFLOW_OP_VALIDATION
@@ -99,11 +98,11 @@ Output<Node> compute_subgraph_scalar_rank(const Output<Node>& output,
element::Type output_type,
bool as_scalar = false);
std::shared_ptr<ov::opset8::Transpose> make_transpose(const ov::Output<ov::Node>& arg,
const ov::AxisVector& input_order);
std::shared_ptr<ov::opset10::Transpose> make_transpose(const ov::Output<ov::Node>& arg,
const ov::AxisVector& input_order);
std::shared_ptr<ov::opset8::Reshape> make_reshape(const ov::Output<ov::Node>& arg,
const std::vector<int64_t>& new_shape);
std::shared_ptr<ov::opset10::Reshape> make_reshape(const ov::Output<ov::Node>& arg,
const std::vector<int64_t>& new_shape);
template <typename T>
void convert_nhwc_to_hw(const std::vector<T>& src, std::vector<size_t>& dst) {

2
src/frontends/tensorflow_common/src/op/fused_batch_norm.cpp
View File

@@ -8,7 +8,7 @@
using namespace std;
using namespace ov;
using namespace ov::opset8;
using namespace ov::opset10;
namespace ov {
namespace frontend {

37
src/frontends/tensorflow_common/src/utils.cpp
View File

@@ -8,7 +8,6 @@
#include "common_op_table.hpp"
#include "openvino/opsets/opset10.hpp"
#include "openvino/opsets/opset8.hpp"
using namespace ov;
using namespace ov::op;
@@ -35,7 +34,7 @@ void set_out_name(const string& out_name, const Output<Node>& output) {
output.get_tensor().add_names({out_name});
}
PadType convert_tf_padding(const ov::frontend::NodeContext& node, const string& tf_padding) {
PadType convert_tf_padding(const frontend::NodeContext& node, const string& tf_padding) {
set<string> supported_ops = {"Conv2D",
"Conv2DBackpropInput",
"Conv3D",
@@ -82,7 +81,7 @@ PadType convert_tf_padding(const ov::frontend::NodeContext& node, const string&
return PadType::EXPLICIT;
}
void fill_explicit_pads_vectors(const ov::frontend::NodeContext& node,
void fill_explicit_pads_vectors(const frontend::NodeContext& node,
bool is_nhwc,
size_t spatial_dims_num,
const vector<int64_t>& tf_explicit_paddings,
@@ -130,7 +129,7 @@ void fill_explicit_pads_vectors(const ov::frontend::NodeContext& node,
}
}
OutputVector translate_convolution_op(const ov::frontend::NodeContext& node, size_t spatial_dims_num) {
OutputVector translate_convolution_op(const frontend::NodeContext& node, size_t spatial_dims_num) {
TENSORFLOW_OP_VALIDATION(node,
spatial_dims_num == 2 || spatial_dims_num == 3,
"Conv2D or Conv3D are supported only.");
@@ -240,9 +239,7 @@ OutputVector translate_convolution_op(const ov::frontend::NodeContext& node, siz
return {conv};
}
void default_op_checks(const ov::frontend::NodeContext& node,
size_t min_input_size,
const vector<string>& supported_ops) {
void default_op_checks(const frontend::NodeContext& node, size_t min_input_size, const vector<string>& supported_ops) {
auto op_type = node.get_op_type();
TENSORFLOW_OP_VALIDATION(node,
find(supported_ops.begin(), supported_ops.end(), op_type) != supported_ops.end(),
@@ -291,17 +288,17 @@ PadMode convert_padding_mode(const NodeContext& node, const string& padding_mode
}
Output<Node> compute_subgraph_scalar_rank(const Output<Node>& output, element::Type output_type, bool as_scalar) {
auto shape_of = make_shared<opset10::ShapeOf>(output, output_type);
auto rank_of = make_shared<opset10::ShapeOf>(shape_of, output_type);
auto shape_of = make_shared<ShapeOf>(output, output_type);
auto rank_of = make_shared<ShapeOf>(shape_of, output_type);
if (as_scalar) {
auto const_zero = make_shared<opset10::Constant>(element::i32, Shape{}, 0);
return make_shared<opset10::Squeeze>(rank_of, const_zero);
auto const_zero = make_shared<Constant>(element::i32, Shape{}, 0);
return make_shared<Squeeze>(rank_of, const_zero);
}
return rank_of;
}
void convert_nhwc_to_nchw(bool need_convert, ov::Output<ov::Node>& node, ov::Rank input_rank) {
void convert_nhwc_to_nchw(bool need_convert, Output<Node>& node, Rank input_rank) {
if (need_convert) {
if (input_rank.is_dynamic()) {
// TODO: use ShapeOf sub-graph to generate permutation vector
@@ -319,7 +316,7 @@ void convert_nhwc_to_nchw(bool need_convert, ov::Output<ov::Node>& node, ov::Ran
}
}
void convert_nchw_to_nhwc(bool need_convert, ov::Output<ov::Node>& node, ov::Rank input_rank) {
void convert_nchw_to_nhwc(bool need_convert, Output<Node>& node, Rank input_rank) {
if (need_convert) {
if (input_rank.is_dynamic()) {
// TODO: use ShapeOf sub-graph to generate permutation vector
@@ -337,17 +334,15 @@ void convert_nchw_to_nhwc(bool need_convert, ov::Output<ov::Node>& node, ov::Ran
}
}
std::shared_ptr<ov::opset8::Transpose> make_transpose(const ov::Output<ov::Node>& arg,
const ov::AxisVector& input_order) {
auto order = std::make_shared<ov::opset8::Constant>(element::i64, Shape{input_order.size()}, input_order);
auto transpose = std::make_shared<ov::opset8::Transpose>(arg, order);
shared_ptr<Transpose> make_transpose(const Output<Node>& arg, const AxisVector& input_order) {
auto order = make_shared<Constant>(element::i64, Shape{input_order.size()}, input_order);
auto transpose = make_shared<Transpose>(arg, order);
return transpose;
}
std::shared_ptr<ov::opset8::Reshape> make_reshape(const ov::Output<ov::Node>& arg,
const std::vector<int64_t>& new_shape) {
auto new_shape_node = std::make_shared<ov::opset8::Constant>(element::i64, Shape{new_shape.size()}, new_shape);
auto reshape = std::make_shared<ov::opset8::Reshape>(arg, new_shape_node, true);
shared_ptr<Reshape> make_reshape(const Output<Node>& arg, const vector<int64_t>& new_shape) {
auto new_shape_node = make_shared<Constant>(element::i64, Shape{new_shape.size()}, new_shape);
auto reshape = make_shared<Reshape>(arg, new_shape_node, true);
return reshape;
}

1
src/frontends/tensorflow_lite/src/op/op_translation_utils.hpp
View File

@@ -13,6 +13,7 @@
#include "openvino/core/node_vector.hpp"
#include "openvino/frontend/tensorflow_lite/node_context.hpp"
#include "openvino/opsets/opset10.hpp"
#include "openvino/opsets/opset8.hpp"
#include "utils.hpp"
namespace ov {

1
src/frontends/tensorflow_lite/src/op_table.cpp
View File

@@ -6,6 +6,7 @@
#include "decoder_map.hpp"
#include "openvino/opsets/opset10.hpp"
#include "openvino/opsets/opset8.hpp"
#include "utils.hpp"
using namespace std;