What is the difference between the model.h and model.cc files used in tinyml in tensorflow lite for microcontrollers? After optimizing my model and converting it to be used on a 32bit ARM Cortex microcontroller I got only the model.cc file. However in all the tutorials that I have seen both the .h and .cc files are used. Where does the .h file come from?
Asked
Active
Viewed 205 times
0
-
Does this answer your question? [Why have header files and .cpp files?](https://stackoverflow.com/questions/333889/why-have-header-files-and-cpp-files) – Dr. Snoopy May 21 '21 at 15:51
2 Answers
0
As mentioned in the comment by Dr.Snoopy, it separates the interface and implementation - just a standard practice. We define both the files (they are not automatically generated)
Meghna Natraj
- 673
- 7
- 15
0
Yes you are both right. My question was in relation to converting a tensorflow model into a quantized tflite model for arduino because I noticed that most people would generate the .cc without a .h file but in usage the .cc file was no where to be found and only the .h file was used. Turns out in this instance it does not matter since what is required is the quantized binary model. So the .cc file will just be renamed to a .h file later and deployed on the arduino.
Atia
- 100
- 1
- 3
- 15