4 ms·Data Parallelism vs. Model Parallelism in Distributed Deep Learning Training2 points by keyboardman 7y ago