Skip to content

Support for signature names instead of tensor names #203

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
47 changes: 40 additions & 7 deletions include/cppflow/model.h
Original file line number Diff line number Diff line change
Expand Up @@ -28,6 +28,7 @@ namespace cppflow {
explicit model(const std::string& filename, const TYPE type=TYPE::SAVED_MODEL);

std::vector<std::string> get_operations() const;
std::string get_tensor_mapping(const std::string& operation) const;
std::vector<int64_t> get_operation_shape(const std::string& operation) const;

std::vector<tensor> operator()(std::vector<std::tuple<std::string, tensor>> inputs, std::vector<std::string> outputs);
Expand All @@ -45,6 +46,8 @@ namespace cppflow {
std::shared_ptr<TF_Status> status;
std::shared_ptr<TF_Graph> graph;
std::shared_ptr<TF_Session> session;

std::vector<std::string> tensor_mapping;
};
}

Expand Down Expand Up @@ -72,6 +75,23 @@ namespace cppflow {
this->session = {TF_LoadSessionFromSavedModel(session_options.get(), run_options.get(), filename.c_str(),
&tag, tag_len, this->graph.get(), meta_graph.get(), this->status.get()),
session_deleter};

std::string_view sv((char*)meta_graph.get()->data, meta_graph.get()->length);

std::string output_name = "StatefulPartitionedCall:0";
std::size_t tensor_mapping_end = sv.npos;
while ((tensor_mapping_end = sv.find(output_name)) != sv.npos)
{
// found at position tensor_mapping_end, now reverse search for beginning
auto tensor_mapping_start = sv.rfind("\x0A", tensor_mapping_end - 4);
if (tensor_mapping_start != sv.npos)
{
tensor_mapping.emplace_back(sv.substr(tensor_mapping_start + 2 /* bytes used for identifying start */,
tensor_mapping_end - tensor_mapping_start - 4 /* byte spacing between the tensor name and output mapping */ - 2));

output_name.back()++; // Increment output index;
}
}
}
else if (type == TYPE::FROZEN_GRAPH) {
this->session = {TF_NewSession(this->graph.get(), session_options.get(), this->status.get()), session_deleter};
Expand Down Expand Up @@ -106,11 +126,29 @@ namespace cppflow {
return result;
}

inline std::tuple<std::string, int> parse_name(const std::string& name) {
auto idx = name.find(':');
return (idx == std::string::npos ? std::make_tuple(name, 0) : std::make_tuple(name.substr(0, idx), std::stoi(name.substr(idx + 1))));
}

inline std::string model::get_tensor_mapping(const std::string& operation) const {
std::string output_name;
auto it = std::find(tensor_mapping.begin(), tensor_mapping.end(), operation);
if (it != tensor_mapping.end())
{
output_name = "StatefulPartitionedCall:0";
output_name.back() += std::distance(tensor_mapping.begin(), it);
}

return output_name;
}

inline std::vector<int64_t> model::get_operation_shape(const std::string& operation) const {
// Get operation by the name
TF_Output out_op;
out_op.oper = TF_GraphOperationByName(this->graph.get(), operation.c_str());
out_op.index = 0;
const auto [op_name, op_idx] = parse_name(operation);
out_op.oper = TF_GraphOperationByName(this->graph.get(), op_name.c_str());
out_op.index = op_idx;

std::vector<int64_t> shape;

Expand Down Expand Up @@ -143,11 +181,6 @@ namespace cppflow {
return shape;
}

inline std::tuple<std::string, int> parse_name(const std::string& name) {
auto idx = name.find(':');
return (idx == std::string::npos ? std::make_tuple(name, 0) : std::make_tuple(name.substr(0, idx), std::stoi(name.substr(idx + 1))));
}

inline std::vector<tensor> model::operator()(std::vector<std::tuple<std::string, tensor>> inputs, std::vector<std::string> outputs) {

std::vector<TF_Output> inp_ops(inputs.size());
Expand Down