kwuking/TimeMixer
TimeMixer is a deep learning time series forecasting model using decomposable multiscale mixing, published at ICLR 2024.
Not currently ranked — collecting fresh signals.
star history
TimeMixer provides a PyTorch implementation of a time series forecasting architecture that decomposes temporal patterns across multiple scales and mixes them for accurate predictions. The model leverages decompose-then-forecast mechanisms to capture both short-term and long-term temporal dependencies. This repository includes training scripts, evaluation benchmarks, and the complete model architecture.
Frequently asked
- What is kwuking/TimeMixer?
- TimeMixer is a deep learning time series forecasting model using decomposable multiscale mixing, published at ICLR 2024.
- Is TimeMixer open source?
- Yes — kwuking/TimeMixer is open source, released under the Apache-2.0 license.
- What language is TimeMixer written in?
- kwuking/TimeMixer is primarily written in Python.
- How popular is TimeMixer?
- kwuking/TimeMixer has 2k stars on GitHub.
- Where can I find TimeMixer?
- kwuking/TimeMixer is on GitHub at https://github.com/kwuking/TimeMixer.