Closed nachtmann-iis closed 1 year ago
The issue still persists. I'm currently trying to docker everything in order to narrow down the exact issue for this behavior here is my current configuration.
version: '3.9'
services:
keyword_spotting_compile:
image: ghcr.io/armmbed/mbed-os-env:master-2021.12.25T04.23.51
command:
- /bin/bash
- -c
- |
cd /app/reference_submissions/keyword_spotting
ln -s /usr/bin/make /usr/bin/gmake
apt update
apt install unzip curl -y
./setup_example.sh
mbed compile -m NUCLEO_L4R5ZI -t GCC_ARM -v
volumes:
- .\tiny-0.7\benchmark:/app
Overview
I tried to reproduce the keyword recognition benchmark using the same hardware (Arduino Uno rev3, STM32-LPM01-XN_PowerShield, STM32-Nucleo-l4r5zi). I followed this guide up until the compilation of the exported model. The compilation failed with an exception related to the usage of
tr_info
inws_bootstrap_ffn.c
.Environment
Windows 10 21H2 WSL 2 Ubuntu 22.04 python 3.8.x
Conda env
Issue
after running setup_example.sh I continued with compiling the model into the binaries i'd like to upload by entering the following command into wsl
The compilation succeeded up around 40% and I got the following error: