google / gemmlowp

Low-precision matrix multiplication
Apache License 2.0
1.77k stars 451 forks source link

Issues compiling for bare metal application #196

Closed dav1d-wright closed 4 years ago

dav1d-wright commented 4 years ago

Hi,

I am trying to compile TFLite for a bare metal application, and have run into issues with gemmlowp while doing so. For my target platform I do not have unistd.h, can anyone help me find a workaround?

bjacob commented 4 years ago

Does TFLite For Microcontrollers meet your needs? https://www.tensorflow.org/lite/microcontrollers/get_started Either way you should ask on a TFLite support channel - even if you overcame any issue with gemmlowp (just getting rid of a unistd.h include would be fairly easy), you would run into more issues with TFLite, which is much larger and has more OS dependencies. TFLite For Microcontrollers is meant to address this class of use cases.

dav1d-wright commented 4 years ago

Hi @bjacob and thanks for your reply. You are right, I should have added more information regarding what I'm up to. I am trying to use tflite for microcontrollers on a new target (TM4C), and I haven't managed to figure out the dependencies that are truly necessary. For example lite/kernels/internal/common.h depends on gemmlowp. It's not in the micro folder, but in all of the examples I found some parts of the lite folder are still included, especially parts of the kernels. But yes, I might have to ask in the tensorflow project. Thanks anyway!