Fast ES-RNN: A GPU Implementation of the ES-RNN Algorithm
This makes state-of-the-art time series forecasting faster and more accessible, though it is an incremental improvement focused on implementation efficiency.
The paper tackled the slow training of the ES-RNN forecasting algorithm by vectorizing and porting it to a GPU, achieving up to 322x speedup while maintaining similar accuracy as the original.
Due to their prevalence, time series forecasting is crucial in multiple domains. We seek to make state-of-the-art forecasting fast, accessible, and generalizable. ES-RNN is a hybrid between classical state space forecasting models and modern RNNs that achieved a 9.4% sMAPE improvement in the M4 competition. Crucially, ES-RNN implementation requires per-time series parameters. By vectorizing the original implementation and porting the algorithm to a GPU, we achieve up to 322x training speedup depending on batch size with similar results as those reported in the original submission. Our code can be found at: https://github.com/damitkwr/ESRNN-GPU