Micro Batch Streaming: Allowing the Training of DNN Models to Use a large Batch Size in Memory Constrained Environments

Recent deep learning models are difficult to train using a large batch size, because commodity machines may not have enough memory to accommodate both the model and a large data batch size. The batch size is one of the hyper-parameters used in the training model, and it is dependent on and is limited by the target machine memory capacity because the batch size can only fit into the remaining memory after the model is uploaded. Moreover, the data item size is also an important factor because if each data item size is larger then the batch size that can fit into the remaining memory becomes smaller. This paper proposes a framework called Micro-Batch Streaming (MBS) to address this problem. This method helps deep learning models to train by providing a batch streaming method that splits a batch into a size that can fit in the remaining memory and streams them sequentially. A loss normalization algorithm based on the gradient accumulation is used to maintain the performance. The purpose of our method is to allow deep learning models to train using larger batch sizes that exceed the memory capacity of a system without increasing the memory size or using multiple devices (GPUs).