Skip to content
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.

Sign up for GitHub

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

[IE Samples] Improved processing outputs for model with more than one output #10737

Merged
merged 3 commits into from
Mar 3, 2022
Merged

[IE Samples] Improved processing outputs for model with more than one output #10737

Changes from 2 commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
31c9edb
Improved processing outputs for model with more than one output
mgrdv Mar 2, 2022
7ea5f81
fixed condition
mgrdv Mar 2, 2022
018e99b
added checking count of output/reference files
mgrdv Mar 3, 2022
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
19 changes: 9 additions & 10 deletions samples/cpp/speech_sample/main.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -345,8 +345,10 @@ int main(int argc, char* argv[]) {
// -----------------------------------------------------------------------------------------------------
// --------------------------- Step 5. Do inference --------------------------------------------------------
std::vector<std::vector<uint8_t>> ptrUtterances;
std::vector<std::vector<uint8_t>> vectorPtrScores((outputs.size() == 0) ? 1 : outputs.size());
std::vector<uint16_t> numScoresPerOutput((outputs.size() == 0) ? 1 : outputs.size());
std::vector<std::vector<uint8_t>> vectorPtrScores((outputs.size() == 0) ? executableNet.outputs().size()
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks like it is not necessary to define the size of those vectors here it will reduce code size

: outputs.size());
std::vector<uint16_t> numScoresPerOutput((outputs.size() == 0) ? executableNet.outputs().size()
: outputs.size());
std::vector<std::vector<uint8_t>> vectorPtrReferenceScores(reference_name_files.size());
std::vector<ScoreErrorT> vectorFrameError(reference_name_files.size()),
vectorTotalError(reference_name_files.size());
Expand Down Expand Up @@ -474,8 +476,9 @@ int main(int argc, char* argv[]) {
inferRequest.inferRequest.wait();
if (inferRequest.frameIndex >= 0)
for (size_t next_output = 0; next_output < count_file; next_output++) {
std::string outputName = (outputs.size() == 0) ? executableNet.output(0).get_any_name()
: output_names[next_output];
std::string outputName = (outputs.size() == 0)
? executableNet.output(next_output).get_any_name()
: output_names[next_output];
auto dims = executableNet.output(outputName).get_shape();
numScoresPerOutput[next_output] = std::accumulate(std::begin(dims),
std::end(dims),
Expand All @@ -493,10 +496,6 @@ int main(int argc, char* argv[]) {

ov::Tensor outputBlob =
inferRequest.inferRequest.get_tensor(executableNet.output(outputName));
if (!outputs.empty()) {
outputBlob =
inferRequest.inferRequest.get_tensor(executableNet.output(outputName));
}
// locked memory holder should be alive all time while access to its buffer happens
auto byteSize = numScoresPerOutput[next_output] * sizeof(float);
std::memcpy(outputFrame, outputBlob.data<float>(), byteSize);
Expand Down Expand Up @@ -654,8 +653,8 @@ int main(int argc, char* argv[]) {
}
if (!FLAGS_r.empty()) {
// print statistical score error
std::string outputName =
(outputs.size() == 0) ? executableNet.output(0).get_any_name() : output_names[next_output];
std::string outputName = (outputs.size() == 0) ? executableNet.output(next_output).get_any_name()
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

we are setting this variable 2 times in the code...I would recommend moving this replacement when we are defining outputs vector, but let's do it next time

: output_names[next_output];
std::cout << "Output name: " << outputName << std::endl;
std::cout << "Number scores per frame: " << numScoresPerOutput[next_output] / batchSize << std::endl
<< std::endl;
Expand Down