Skip to content

Instantly share code, notes, and snippets.

@dwaq

dwaq/Arduino Tensorflow Setup.md

Created February 5, 2020 13:22
Show Gist options
  • Save dwaq/7c8c30a888b640c58488b6f78b61f2c9 to your computer and use it in GitHub Desktop.
Save dwaq/7c8c30a888b640c58488b6f78b61f2c9 to your computer and use it in GitHub Desktop.
Download ZIP
Raw
Arduino Tensorflow Setup.md

Installation

In WSL:

Download tensorflow:

git clone --depth 1 https://github.com/tensorflow/tensorflow.git

cd tensorflow

Download Python dependencies:

pip3 install pycrypto pyserial --user

Compile code for hello_world (important, because it downloads and installs other tensorflow dependencies)

(might need to install unzip first: sudo apt install unzip)

make -f tensorflow/lite/micro/tools/make/Makefile TARGET=sparkfun_edge hello_world_bin

Add Redboard Artemis ATP as a valid target:

  • In Makefile /tensorflow/lite/micro/tools/make/targets/apollo3evb_makefile.inc, add sparkfun_redboard_artemis_atp\ after sparkfun_edge\ (line 4)
  • In example project directory (eg: /tensorflow/lite/micro/examples/hello_world/), cp sparkfun_edge directory to sparkfun_redboard_artemis_atp and modify for new board (would use BSP at /tensorflow/lite/micro/tools/make/downloads/AmbiqSuite-Rel2.2.0/boards_sfe/redboard_artemis/bsp/am_bsp.c)

Compile the actual code that we want to run:

make -f tensorflow/lite/micro/tools/make/Makefile TARGET=sparkfun_redboard_artemis_atp micro_speech_bin

Flash it to the board:

./tensorflow/lite/micro/tools/make/downloads/AmbiqSuite-Rel2.2.0/boards_sfe/common/tools_sfe/ambiq/ambiq_bin2board.py --bin tensorflow/lite/micro/tools/make/gen/sparkfun_redboard_artemis_atp_cortex-m4/bin/micro_speech.bin

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment