[WIP] Machine learning library implementing neural networks from scratch for a school project.
You can build the library with the following make commands :
| Command | Effect | Output folder |
|---|---|---|
make / make all |
Build library, tests, benchmarks and examples | lib && bin |
make lib |
Build library | lib |
make test |
Build tests | bin |
make perf |
Build benchmarks | bin |
make examples |
Build examples | bin |
The library's .o and .so files are built in the lib directory.
Other executables are built in the bin directory with an adequate suffix.
Note : Executables in the bin directory must be run from the root of the
project to be able to find libtensorslow.so, so you have to use the
following command :
./bin/example_test
In order to clean the build directories, you can use the following phonies :
| Command | Effect | Target folder |
|---|---|---|
make clean |
Clean everything | lib && bin |
make clean_lib |
Clean everything in lib |
lib |
make clean_o |
Clean .o files in lib |
lib |
make clean_so |
Clean .so files in lib |
lib |
make clean_bin |
Clean everything in bin |
bin |
make clean_test |
Clean tests in bin |
bin |
make clean_perf |
Clean benchmarks in bin |
bin |
make clean_examples |
Clean examples in bin |
bin |
- googletest is required to build the tests.