本文內容來自先楫開發者 @Xusiwei1236，介紹了如何在HPM6750上運行邊緣AI框架，感興趣的小伙伴快點來看看

--------------- 以下為測評內容---------------

TFLM是什么？

你或許都聽說過TensorFlow——由谷歌開發并開源的一個機器學習庫，它支持模型訓練和模型推理。

今天介紹的TFLM，全稱是TensorFlow Lite for Microcontrollers，翻譯過來就是“針對微控制器的TensorFlow Lite”。那TensorFlow Lite又是什么呢？

TensorFlow Lite（通常簡稱TFLite）其實是TensorFlow團隊為了將模型部署到移動設備而開發的一套解決方案，通俗的說就是手機版的TensorFlow。下面是TensorFlow官網上關于TFLite的一段介紹：

“TensorFlow Lite 是一組工具，可幫助開發者在移動設備、嵌入式設備和 loT 設備上運行模型，以便實現設備端機器學習。”

而我們今天要介紹的TensorFlow Lite for Microcontrollers（TFLM）則是 TensorFlow Lite的微控制器版本。這里是官網上的一段介紹：

“ TensorFlow Lite for Microcontrollers （以下簡稱TFLM）是 TensorFlow Lite 的一個實驗性移植版本，它適用于微控制器和其他一些僅有數千字節內存的設備。它可以直接在“裸機”上運行，不需要操作系統支持、任何標準 C/C++ 庫和動態內存分配。核心運行時(core runtime)在 Cortex M3 上運行時僅需 16KB，加上足以用來運行語音關鍵字檢測模型的操作，也只需 22KB 的空間。”

這三者一脈相承，都出自谷歌，區別是TensorFlow同時支持訓練和推理，而后兩者只支持推理。TFLite主要用于支持手機、平板等移動設備，TFLM則可以支持單片機。從發展歷程上來說，后兩者都是TensorFlow項目的“支線項目”。或者說這三者是一個樹形的發展過程，具體來說，TFLite是從TensorFlow項目分裂出來的，TFLite-Micro是從TFLite分裂出來的，目前是三個并行發展的。在很長一段時間內，這三個項目的源碼都在一個代碼倉中維護，從源碼目錄的包含關系上來說，TensorFlow包含后兩者，TFLite包含tflite-micro。

HPMSDK中的TFLM

• TFLM中間件

HPM SDK中集成了TFLM中間件（類似庫，但是沒有單獨編譯為庫），位于hpm_sdk\middleware子目錄：

這個子目錄的代碼是由TFLM開源項目裁剪而來，刪除了很多不需要的文件。

TFLM示例

HPM SDK中也提供了TFLM示例，位于hpm_sdk\samples\tflm子目錄：

示例代碼是從官方的persion_detection示例修改而來，添加了攝像頭采集圖像和LCD顯示結果。
由于我手里沒有配套的攝像頭和顯示屏，所以本篇沒有以這個示例作為實驗。

在HPM6750上運行TFLM基準測試

接下來以person detection benchmark為例，講解如何在HPM6750上運行TFLM基準測試。

將person detection benchmark源代碼添加到HPM SDK環境

按照如下步驟，在HPM SDK環境中添加person detection benchmark源代碼文件：

在HPM SDK的samples子目錄創建tflm_person_detect_benchmark目錄，并在其中創建src目錄；

從上文描述的已經運行過person detection benchmark的tflite-micro目錄中拷貝如下文件到src目錄：

tensorflow\lite\micro\benchmarks\person_detection_benchmark.cc

tensorflow\lite\micro\benchmarks\micro_benchmark.h

tensorflow\lite\micro\examples\person_detection\model_settings.h

tensorflow\lite\micro\examples\person_detection\model_settings.cc

在src目錄創建testdata子目錄，并將tflite-micro目錄下如下目錄中的文件拷貝全部到testdata中：

tensorflow\lite\micro\tools\make\gen\linux_x86_64_default\genfiles\tensorflow\lite\micro\examples\person_detection\testdata

修改person_detection_benchmark.cc、model_settings.cc、no_person_image_data.cc、person_image_data.cc 文件中部分#include預處理指令的文件路徑（根據拷貝后的相對路徑修改）；

person_detection_benchmark.cc文件中，main函數的一開始添加一行board_init();、頂部添加一行#include "board.h”

添加CMakeLists.txt和app.yaml文件

在src平級創建CMakeLists.txt文件，內容如下：

cmake_minimum_required(VERSION 3.13)

set(CONFIG_TFLM 1)

find_package(hpm-sdk REQUIRED HINTS $ENV{HPM_SDK_BASE})

project(tflm_person_detect_benchmark)

set(CMAKE_CXX_STANDARD 11)

sdk_app_src(src/model_settings.cc)

sdk_app_src(src/person_detection_benchmark.cc)

sdk_app_src(src/testdata/no_person_image_data.cc)

sdk_app_src(src/testdata/person_image_data.cc)

sdk_app_inc(src)

sdk_ld_options("-lm")

sdk_ld_options("--std=c++11")

sdk_compile_definitions(__HPMICRO__)

sdk_compile_definitions(-DINIT_EXT_RAM_FOR_DATA=1)

# sdk_compile_options("-mabi=ilp32f")

# sdk_compile_options("-march=rv32imafc")

sdk_compile_options("-O2")

# sdk_compile_options("-O3")

set(SEGGER_LEVEL_O3 1)

generate_ses_project()

在src平級創建app.yaml文件，內容如下：

dependency:

  - tflm

• 編譯和運行TFLM基準測試

接下來就是大家熟悉的——編譯運行了。首先，使用generate_project生產項目：接著，將HPM6750開發板連接到PC，在Embedded Studio中打卡剛剛生產的項目：這個項目因為引入了TFLM的源碼，文件較多，所以右邊的源碼導航窗里面的Indexing要執行很久才能結束。

然后，就可以使用F7編譯、F5調試項目了：

編譯完成后，先打卡串口終端連接到設備串口，波特率115200。啟動調試后，直接繼續運行，就可以在串口終端中看到基準測試的輸出了：

==============================

 hpm6750evkmini clock summary

==============================

cpu0:            816000000Hz

cpu1:            816000000Hz

axi0:            200000000Hz

axi1:            200000000Hz

axi2:            200000000Hz

ahb:             200000000Hz

mchtmr0:         24000000Hz

mchtmr1:         1000000Hz

xpi0:            133333333Hz

xpi1:            400000000Hz

dram:            166666666Hz

display:         74250000Hz

cam0:            59400000Hz

cam1:            59400000Hz

jpeg:            200000000Hz

pdma:            200000000Hz

==============================

----------------------------------------------------------------------

$$\   $$\ $$$$$$$\  $$\      $$\ $$\

$$ |  $$ |$$  __$$\ $$$\    $$$ |\__|

$$ |  $$ |$$ |  $$ |$$$$\  $$$$ |$$\  $$$$$$$\  $$$$$$\   $$$$$$\

$$$$$$$$ |$$$$$$$  |$$\$$\$$ $$ |$$ |$$  _____|$$  __$$\ $$  __$$\

$$  __$$ |$$  ____/ $$ \$$$  $$ |$$ |$$ /      $$ |  \__|$$ /  $$ |

$$ |  $$ |$$ |      $$ |\$  /$$ |$$ |$$ |      $$ |      $$ |  $$ |

$$ |  $$ |$$ |      $$ | \_/ $$ |$$ |\$$$$$$$\ $$ |      \$$$$$$  |

\__|  \__|\__|      \__|     \__|\__| \_______|\__|       \______/

----------------------------------------------------------------------

InitializeBenchmarkRunner took 114969 ticks (4 ms).

WithPersonDataIterations(1) took 10694521 ticks (445 ms)

DEPTHWISE_CONV_2D took 275798 ticks (11 ms).

DEPTHWISE_CONV_2D took 280579 ticks (11 ms).

CONV_2D took 516051 ticks (21 ms).

DEPTHWISE_CONV_2D took 139000 ticks (5 ms).

CONV_2D took 459646 ticks (19 ms).

DEPTHWISE_CONV_2D took 274903 ticks (11 ms).

CONV_2D took 868518 ticks (36 ms).

DEPTHWISE_CONV_2D took 68180 ticks (2 ms).

CONV_2D took 434392 ticks (18 ms).

DEPTHWISE_CONV_2D took 132918 ticks (5 ms).

CONV_2D took 843014 ticks (35 ms).

DEPTHWISE_CONV_2D took 33228 ticks (1 ms).

CONV_2D took 423288 ticks (17 ms).

DEPTHWISE_CONV_2D took 62040 ticks (2 ms).

CONV_2D took 833033 ticks (34 ms).

DEPTHWISE_CONV_2D took 62198 ticks (2 ms).

CONV_2D took 834644 ticks (34 ms).

DEPTHWISE_CONV_2D took 62176 ticks (2 ms).

CONV_2D took 838212 ticks (34 ms).

DEPTHWISE_CONV_2D took 62206 ticks (2 ms).

CONV_2D took 832857 ticks (34 ms).

DEPTHWISE_CONV_2D took 62194 ticks (2 ms).

CONV_2D took 832882 ticks (34 ms).

DEPTHWISE_CONV_2D took 16050 ticks (0 ms).

CONV_2D took 438774 ticks (18 ms).

DEPTHWISE_CONV_2D took 27494 ticks (1 ms).

CONV_2D took 974362 ticks (40 ms).

AVERAGE_POOL_2D took 2323 ticks (0 ms).

CONV_2D took 1128 ticks (0 ms).

RESHAPE took 184 ticks (0 ms).

SOFTMAX took 2249 ticks (0 ms).

NoPersonDataIterations(1) took 10694160 ticks (445 ms)

DEPTHWISE_CONV_2D took 274922 ticks (11 ms).

DEPTHWISE_CONV_2D took 281095 ticks (11 ms).

CONV_2D took 515380 ticks (21 ms).

DEPTHWISE_CONV_2D took 139428 ticks (5 ms).

CONV_2D took 460039 ticks (19 ms).

DEPTHWISE_CONV_2D took 275255 ticks (11 ms).

CONV_2D took 868787 ticks (36 ms).

DEPTHWISE_CONV_2D took 68384 ticks (2 ms).

CONV_2D took 434537 ticks (18 ms).

DEPTHWISE_CONV_2D took 133071 ticks (5 ms).

CONV_2D took 843202 ticks (35 ms).

DEPTHWISE_CONV_2D took 33291 ticks (1 ms).

CONV_2D took 423388 ticks (17 ms).

DEPTHWISE_CONV_2D took 62190 ticks (2 ms).

CONV_2D took 832978 ticks (34 ms).

DEPTHWISE_CONV_2D took 62205 ticks (2 ms).

CONV_2D took 834636 ticks (34 ms).

DEPTHWISE_CONV_2D took 62213 ticks (2 ms).

CONV_2D took 838212 ticks (34 ms).

DEPTHWISE_CONV_2D took 62239 ticks (2 ms).

CONV_2D took 832850 ticks (34 ms).

DEPTHWISE_CONV_2D took 62217 ticks (2 ms).

CONV_2D took 832856 ticks (34 ms).

DEPTHWISE_CONV_2D took 16040 ticks (0 ms).

CONV_2D took 438779 ticks (18 ms).

DEPTHWISE_CONV_2D took 27481 ticks (1 ms).

CONV_2D took 974354 ticks (40 ms).

AVERAGE_POOL_2D took 1812 ticks (0 ms).

CONV_2D took 1077 ticks (0 ms).

RESHAPE took 341 ticks (0 ms).

SOFTMAX took 901 ticks (0 ms).

WithPersonDataIterations(10) took 106960312 ticks (4456 ms)

NoPersonDataIterations(10) took 106964554 ticks (4456 ms)

可以看到，在HPM6750EVKMINI開發板上，連續運行10次人像檢測模型，總體耗時4456毫秒，每次平均耗時445.6毫秒。

在樹莓派3B+上運行TFLM基準測試

在樹莓派上運行TFLM基準測試

樹莓派3B+上可以和PC上類似，直接運行PC端的測試命令，得到基準測試結果：

可以看到，在樹莓派3B+上的，對于有人臉的圖片，連續運行10次人臉檢測模型，總體耗時4186毫秒，每次平均耗時418.6毫秒；對于無人臉的圖片，連續運行10次人臉檢測模型，耗時4190毫秒，每次平均耗時419毫秒。

HPM6750和樹莓派3B+、AMD R7 4800H上的基準測試結果對比

這里將HPM6750EVKMINI開發板、樹莓派3B+和AMD R7 4800H上運行人臉檢測模型的平均耗時結果匯總如下：

可以看到，在TFLM人臉檢測模型計算場景下，HPM6750EVKMINI和樹莓派3B+成績相當。雖然HPM6750的816MHz CPU頻率比樹莓派3B+搭載的BCM2837 Cortex-A53 1.4GHz的主頻低，但是在單核心計算能力上沒有相差太多。

這里樹莓派3B+上的TFLM基準測試程序是運行在64位Debian Linux發行版上的，而HPM6750上的測試程序是直接運行在裸機上的。由于操作系統內核中任務調度器的存在，會對CPU的計算能力帶來一定損耗。所以，這里進行的并不是一個嚴格意義上的對比測試，測試結果僅供參考。

（本文參考鏈接：http://m.eeworld.com.cn/bbs_thread-1208270-1-1.html）