to start with,please forgive me that the mistake of grammar and words ,thank you very much!
I found the tutorial on the ST website, where the test file data set was 10 rows, 270 columns, and then it was divided into 10 branches. then divided into (10,90,3,1), and finally succeeded. I used my own data set, 80 rows, 2 columns, and it didn't work. Do you know why?
this is the print information:
Starting AI validation on target with custom dataset : D:\pycharm\my projects\ch4_线性回归代�?\data.csv...
Neural Network Tools for STM32 v1.2.0 (AI tools v5.0.0)
-- Importing model
-- Importing model - done (elapsed time 2.455s)
-- Building X86 C-model
..............
-- Building X86 C-model - done (elapsed time 13.749s)
-- Setting inputs (and outputs) data
Using user-provided input file(s), shapes=(80, 2)
-- Running STM32 C-model
ON-DEVICE STM32 execution ("hargithub", None, 115200)..
<Stm32com id=0x1f336cb4d30 - CONNECTED(COM4/115200) devid=0x413/UNKNOW msg=2.1>
0x413/UNKNOW @168MHz/168MHz (FPU is present) lat=5 ART: PRFTen ICen DCen
found network(s): ['hargithub']
description : 'hargithub' 1->[7]->1 macc=874970 rom=775.52KiB ram=24.00KiB
tools versions : rt=(4, 1, 0) tool=(5, 0, 0)/(1, 3, 0) api=(1, 1, 0) "Wed Apr 29 14:07:33 2020"
TOOL ERROR: cannot reshape array of size 160 into shape (80,90,3,1)
Creating report file C:\Users\lenovo\.stm32cubemx\stm32ai_output\hargithub_validate_report.txt
TOOL ERROR: cannot reshape array of size 160 into shape (80,90,3,1)
Using TensorFlow backend.
Fatal Python error: PyImport_GetModuleDict: no module dictionary!
Current thread 0x00000f34 (most recent call first):
File "network_training\generate\validation\stm32com\
stm32nn.py", line 324 in _reset
File "network_training\generate\validation\stm32com\
stm32nn.py", line 369 in __del__
TOOL ERROR: cannot reshape array of size 160 into shape (80,90,3,1)
and then if i want to do the recognition of number by neural network in stm32,how to import the data set? please give me some article to refer to,thank you very much!
