summaryrefslogtreecommitdiff
path: root/source/use_case/kws/src/UseCaseHandler.cc
diff options
context:
space:
mode:
Diffstat (limited to 'source/use_case/kws/src/UseCaseHandler.cc')
-rw-r--r--source/use_case/kws/src/UseCaseHandler.cc104
1 files changed, 56 insertions, 48 deletions
diff --git a/source/use_case/kws/src/UseCaseHandler.cc b/source/use_case/kws/src/UseCaseHandler.cc
index c20c32b..ce99ed3 100644
--- a/source/use_case/kws/src/UseCaseHandler.cc
+++ b/source/use_case/kws/src/UseCaseHandler.cc
@@ -1,6 +1,6 @@
/*
- * SPDX-FileCopyrightText: Copyright 2021-2022 Arm Limited and/or its affiliates <open-source-office@arm.com>
- * SPDX-License-Identifier: Apache-2.0
+ * SPDX-FileCopyrightText: Copyright 2021-2022 Arm Limited and/or its affiliates
+ * <open-source-office@arm.com> SPDX-License-Identifier: Apache-2.0
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
@@ -16,16 +16,16 @@
*/
#include "UseCaseHandler.hpp"
+#include "AudioUtils.hpp"
+#include "ImageUtils.hpp"
#include "InputFiles.hpp"
#include "KwsClassifier.hpp"
+#include "KwsProcessing.hpp"
+#include "KwsResult.hpp"
#include "MicroNetKwsModel.hpp"
-#include "hal.h"
-#include "AudioUtils.hpp"
-#include "ImageUtils.hpp"
#include "UseCaseCommonUtils.hpp"
-#include "KwsResult.hpp"
+#include "hal.h"
#include "log_macros.h"
-#include "KwsProcessing.hpp"
#include <vector>
@@ -42,15 +42,15 @@ namespace app {
/* KWS inference handler. */
bool ClassifyAudioHandler(ApplicationContext& ctx, uint32_t clipIndex, bool runAll)
{
- auto& profiler = ctx.Get<Profiler&>("profiler");
- auto& model = ctx.Get<Model&>("model");
+ auto& profiler = ctx.Get<Profiler&>("profiler");
+ auto& model = ctx.Get<Model&>("model");
const auto mfccFrameLength = ctx.Get<int>("frameLength");
const auto mfccFrameStride = ctx.Get<int>("frameStride");
- const auto scoreThreshold = ctx.Get<float>("scoreThreshold");
+ const auto scoreThreshold = ctx.Get<float>("scoreThreshold");
/* If the request has a valid size, set the audio index. */
if (clipIndex < NUMBER_OF_FILES) {
- if (!SetAppCtxIfmIdx(ctx, clipIndex,"clipIndex")) {
+ if (!SetAppCtxIfmIdx(ctx, clipIndex, "clipIndex")) {
return false;
}
}
@@ -58,9 +58,10 @@ namespace app {
constexpr uint32_t dataPsnTxtInfStartX = 20;
constexpr uint32_t dataPsnTxtInfStartY = 40;
- constexpr int minTensorDims = static_cast<int>(
- (MicroNetKwsModel::ms_inputRowsIdx > MicroNetKwsModel::ms_inputColsIdx)?
- MicroNetKwsModel::ms_inputRowsIdx : MicroNetKwsModel::ms_inputColsIdx);
+ constexpr int minTensorDims =
+ static_cast<int>((MicroNetKwsModel::ms_inputRowsIdx > MicroNetKwsModel::ms_inputColsIdx)
+ ? MicroNetKwsModel::ms_inputRowsIdx
+ : MicroNetKwsModel::ms_inputColsIdx);
if (!model.IsInited()) {
printf_err("Model is not initialised! Terminating processing.\n");
@@ -68,7 +69,7 @@ namespace app {
}
/* Get Input and Output tensors for pre/post processing. */
- TfLiteTensor* inputTensor = model.GetInputTensor(0);
+ TfLiteTensor* inputTensor = model.GetInputTensor(0);
TfLiteTensor* outputTensor = model.GetOutputTensor(0);
if (!inputTensor->dims) {
printf_err("Invalid input tensor dims\n");
@@ -79,20 +80,22 @@ namespace app {
}
/* Get input shape for feature extraction. */
- TfLiteIntArray* inputShape = model.GetInputShape(0);
+ TfLiteIntArray* inputShape = model.GetInputShape(0);
const uint32_t numMfccFeatures = inputShape->data[MicroNetKwsModel::ms_inputColsIdx];
- const uint32_t numMfccFrames = inputShape->data[arm::app::MicroNetKwsModel::ms_inputRowsIdx];
+ const uint32_t numMfccFrames =
+ inputShape->data[arm::app::MicroNetKwsModel::ms_inputRowsIdx];
/* We expect to be sampling 1 second worth of data at a time.
* NOTE: This is only used for time stamp calculation. */
const float secondsPerSample = 1.0 / audio::MicroNetKwsMFCC::ms_defaultSamplingFreq;
/* Set up pre and post-processing. */
- KwsPreProcess preProcess = KwsPreProcess(inputTensor, numMfccFeatures, numMfccFrames,
- mfccFrameLength, mfccFrameStride);
+ KwsPreProcess preProcess = KwsPreProcess(
+ inputTensor, numMfccFeatures, numMfccFrames, mfccFrameLength, mfccFrameStride);
std::vector<ClassificationResult> singleInfResult;
- KwsPostProcess postProcess = KwsPostProcess(outputTensor, ctx.Get<KwsClassifier &>("classifier"),
+ KwsPostProcess postProcess = KwsPostProcess(outputTensor,
+ ctx.Get<KwsClassifier&>("classifier"),
ctx.Get<std::vector<std::string>&>("labels"),
singleInfResult);
@@ -103,26 +106,29 @@ namespace app {
auto currentIndex = ctx.Get<uint32_t>("clipIndex");
/* Creating a sliding window through the whole audio clip. */
- auto audioDataSlider = audio::SlidingWindow<const int16_t>(
- get_audio_array(currentIndex),
- get_audio_array_size(currentIndex),
- preProcess.m_audioDataWindowSize, preProcess.m_audioDataStride);
+ auto audioDataSlider =
+ audio::SlidingWindow<const int16_t>(GetAudioArray(currentIndex),
+ GetAudioArraySize(currentIndex),
+ preProcess.m_audioDataWindowSize,
+ preProcess.m_audioDataStride);
/* Declare a container to hold results from across the whole audio clip. */
std::vector<kws::KwsResult> finalResults;
/* Display message on the LCD - inference running. */
std::string str_inf{"Running inference... "};
- hal_lcd_display_text(str_inf.c_str(), str_inf.size(),
- dataPsnTxtInfStartX, dataPsnTxtInfStartY, 0);
- info("Running inference on audio clip %" PRIu32 " => %s\n", currentIndex,
- get_filename(currentIndex));
+ hal_lcd_display_text(
+ str_inf.c_str(), str_inf.size(), dataPsnTxtInfStartX, dataPsnTxtInfStartY, 0);
+ info("Running inference on audio clip %" PRIu32 " => %s\n",
+ currentIndex,
+ GetFilename(currentIndex));
/* Start sliding through audio clip. */
while (audioDataSlider.HasNext()) {
const int16_t* inferenceWindow = audioDataSlider.Next();
- info("Inference %zu/%zu\n", audioDataSlider.Index() + 1,
+ info("Inference %zu/%zu\n",
+ audioDataSlider.Index() + 1,
audioDataSlider.TotalStrides() + 1);
/* Run the pre-processing, inference and post-processing. */
@@ -142,19 +148,21 @@ namespace app {
}
/* Add results from this window to our final results vector. */
- finalResults.emplace_back(kws::KwsResult(singleInfResult,
- audioDataSlider.Index() * secondsPerSample * preProcess.m_audioDataStride,
- audioDataSlider.Index(), scoreThreshold));
+ finalResults.emplace_back(kws::KwsResult(
+ singleInfResult,
+ audioDataSlider.Index() * secondsPerSample * preProcess.m_audioDataStride,
+ audioDataSlider.Index(),
+ scoreThreshold));
#if VERIFY_TEST_OUTPUT
DumpTensor(outputTensor);
-#endif /* VERIFY_TEST_OUTPUT */
+#endif /* VERIFY_TEST_OUTPUT */
} /* while (audioDataSlider.HasNext()) */
/* Erase. */
str_inf = std::string(str_inf.size(), ' ');
- hal_lcd_display_text(str_inf.c_str(), str_inf.size(),
- dataPsnTxtInfStartX, dataPsnTxtInfStartY, false);
+ hal_lcd_display_text(
+ str_inf.c_str(), str_inf.size(), dataPsnTxtInfStartX, dataPsnTxtInfStartY, false);
ctx.Set<std::vector<kws::KwsResult>>("results", finalResults);
@@ -164,7 +172,7 @@ namespace app {
profiler.PrintProfilingResult();
- IncrementAppCtxIfmIdx(ctx,"clipIndex");
+ IncrementAppCtxIfmIdx(ctx, "clipIndex");
} while (runAll && ctx.Get<uint32_t>("clipIndex") != initialClipIdx);
@@ -175,7 +183,7 @@ namespace app {
{
constexpr uint32_t dataPsnTxtStartX1 = 20;
constexpr uint32_t dataPsnTxtStartY1 = 30;
- constexpr uint32_t dataPsnTxtYIncr = 16; /* Row index increment. */
+ constexpr uint32_t dataPsnTxtYIncr = 16; /* Row index increment. */
hal_lcd_set_text_color(COLOR_GREEN);
info("Final results:\n");
@@ -190,28 +198,28 @@ namespace app {
float score = 0.f;
if (!result.m_resultVec.empty()) {
topKeyword = result.m_resultVec[0].m_label;
- score = result.m_resultVec[0].m_normalisedVal;
+ score = result.m_resultVec[0].m_normalisedVal;
}
- std::string resultStr =
- std::string{"@"} + std::to_string(result.m_timeStamp) +
- std::string{"s: "} + topKeyword + std::string{" ("} +
- std::to_string(static_cast<int>(score * 100)) + std::string{"%)"};
+ std::string resultStr = std::string{"@"} + std::to_string(result.m_timeStamp) +
+ std::string{"s: "} + topKeyword + std::string{" ("} +
+ std::to_string(static_cast<int>(score * 100)) +
+ std::string{"%)"};
- hal_lcd_display_text(resultStr.c_str(), resultStr.size(),
- dataPsnTxtStartX1, rowIdx1, false);
+ hal_lcd_display_text(
+ resultStr.c_str(), resultStr.size(), dataPsnTxtStartX1, rowIdx1, false);
rowIdx1 += dataPsnTxtYIncr;
if (result.m_resultVec.empty()) {
- info("For timestamp: %f (inference #: %" PRIu32
- "); label: %s; threshold: %f\n",
- result.m_timeStamp, result.m_inferenceNumber,
+ info("For timestamp: %f (inference #: %" PRIu32 "); label: %s; threshold: %f\n",
+ result.m_timeStamp,
+ result.m_inferenceNumber,
topKeyword.c_str(),
result.m_threshold);
} else {
for (uint32_t j = 0; j < result.m_resultVec.size(); ++j) {
info("For timestamp: %f (inference #: %" PRIu32
- "); label: %s, score: %f; threshold: %f\n",
+ "); label: %s, score: %f; threshold: %f\n",
result.m_timeStamp,
result.m_inferenceNumber,
result.m_resultVec[j].m_label.c_str(),
generated by cgit v1.2.1 (git 2.23.0) at 2024-05-19 15:17:50 +0000