chrishkchris commented on issue #14: rearrange contents in dist-train.md URL: https://github.com/apache/singa-doc/pull/14#issuecomment-609033831 > Then would it be better to design the APIs in this way: > > 1. For training with MPI > > ```python > # in mnist_mpi.py > if __name__ == '__main__': > sgd = ... > sgd = DistOpt(sgd) > train_mnist(sgd, sparse, topK) > ``` > > 1. For training via multiprocessing > > ```python > # in mnist_multiprocessing.py > if __name__ == '__main__': > nccl_id = ... > sgd = ... > sgd = DistOpt(sgd, num_gpu, nccl_id) > train_mnist(sgd, sparse, topK) > ``` > > Even if you use socket, multiprocessing can only run on a single node, hence num_gpu = gpu_per_node. sorry, Distopt cannot passed into the multiprocess, this is because DistOpt has the communicator initialized, if we pass it into multiprocess the same object will be replicated. The code cannot run
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
