Question/Issue:
I am trying to run an inference model on the AURIX TC375 - the model i use was succesfully deployed on the TC397 already which though uses more RAM. Since I accidentaly burned parts of the power supply I can no longer use the TC397 and am forced to use the TC375. I am using the AURIX IDE and the ext. GCC - I can compile the project and flash it onto the TC375 - but I get the following errors when debugging over UART:
ERR: MFCC failed (-1002)<\n>
ERR: Failed to run DSP process (-1002)<\n>
ERR: Failed to run classifier (-5)<\n>
Steps Taken:
Ive looked up some files but do not really know how to fix that or where or if the model trained can even be ported like that. I looked up where the messaghes coe from and what they mean. I also checked for tensorarenasize and so forth but randomnly changing that didnt work (making ti bigger or smaller) and also doesn’t seem like a generally good approach.
Reproducibility:
- [x ] Always
Environment:
- Platform: AURIX IDE - tolchain Tasking - using external GCC
- Build Environment Details: [e.g., Arduino IDE 1.8.19 ESP32 Core for Arduino 2.0.4]
- OS Version: [ Windows 10] (not relevant though)