I was trying to run KWS example using main.c and kws_wights.h on my pc. I used the test dataset as my input data instead of usage of the mic. I convert wav file to 16-bit c array and iterate through array to extract audio buffer each iteration of size 512. The preprocessing was done before calling the model using mfcc.c. However, the output of the model was not correct. I compared the output of the mfcc features using python files to the c files and they were not the same. Moreover, the size of mfcc features in the python file is 12 62 while for the c file files is 1263. My question is why the size and output are not the same? you can file the maim.c file used attachedmain.c.zip
I was trying to run KWS example using main.c and kws_wights.h on my pc. I used the test dataset as my input data instead of usage of the mic. I convert wav file to 16-bit c array and iterate through array to extract audio buffer each iteration of size 512. The preprocessing was done before calling the model using mfcc.c. However, the output of the model was not correct. I compared the output of the mfcc features using python files to the c files and they were not the same. Moreover, the size of mfcc features in the python file is 12 62 while for the c file files is 1263. My question is why the size and output are not the same? you can file the maim.c file used attachedmain.c.zip
Thanks in advance