Here are detailed the steps necessary to use the tools included in the Docker image to execute and compile EDDL code.
The Docker image can be cloned into any device with the command:
docker pull solverml/eddl_riscv:cross_compilation
Once cloned, we can run the image as a container by executing the command:
docker run -t -d solverml/eddl_riscv:cross_compilation /bin/bash
To access the Docker container from a terminal run the following command:
docker exec -i -t container_id /bin/bash where 'container_id' needs to be replaced by the ID assign to the Docker container we have just created. To check this ID simply run the command docker ps. This will keep open a terminal allowing to execute commands as a root user inside the docker container.
In this docker container is possible to cross-compile programs since the cross-compiling tools are already installed.
To show how to do this, included with the Docker container there is a code example using the MNIST dataset in the folder /home/eddl_examples/cifar10_eddl_cross_train. Along with the code there is also a CMakeLists.txt file we can use to cross-compile this code. Instead of using CMake to compile the code using the default compiler of the system we need to specify the C++ compiler in the CMake command. For example, to cross-compile the code available in /home/eddl_examples/cifar10_eddl_cross_train move to the folder and execute this commands on the terminal:
mkdir buildcd buildcmake .. -DCMAKE_CXX_COMPILER="/home/isar-riscv/sdk/sdk-debian-sid-ports-riscv64/usr/bin/riscv64-linux-gnu-g++"make
This will generate and executable file compiled for a RISC-V system. Transferring the file to a RISC-V emulation like the ones available in the other dockers on this repository should be enough to execute it without any more modifications.