Model Parallelism is a parallel computing paradigm in the context of deep learning that involves distributing and parallelizing the components of a neural network across multiple computing devices or processing units. In contrast to data parallelism, where multiple copies of the entire model are trained on different subsets of the data, model parallelism focuses on splitting and distributing the model itself. This approach is particularly useful when a neural network is too large to fit into the memory of a single device.


SplitBrain: Hybrid Data and Model Parallel Deep Learning

The recent success of deep learning applications has coincided with those widely available powerful computational resources for training sophisticated machine learning models with huge datasets. Nonetheless, training large models such as convolutional neural networks using model parallelism (as opposed to data parallelism) is challenging because the complex nature of communication between model shards makes it difficult to partition the computation efficiently across multiple machines with an acceptable trade off. This paper presents SplitBrain, a high performance distributed deep learning framework supporting hybrid data and model parallelism. Specifically, SplitBrain provides layer specific partitioning that co locates compute intensive convolutional layers while sharding memory demanding layers. A novel scalable group communication is proposed to further improve the training throughput with reduced communication overhead. The results show that SplitBrain can achieve nearly linear speedup while saving up to 67% of memory consumption for data and model parallel VGG over CIFAR 10.