Ning Sun

🎉🚀📈 New release for NeuralForecast 📈🚀🎉 We’re thrilled to share a multitude of new models and features in this release! Many new multivariate models are now available: 🤩 iTransformer: the latest Transformer-based forecasting model ☄ MLPMultivariate: an MLP model that supports multivariate tasks 🥣 TSMixer and TSMixerx: MLP-based model by Google with feature mixing New recent and baseline univariate models are now available: ⏱ BiTCN: a model leveraging temporal convolutional networks 📈 NLinear and DLinear We've also added Time-LLM! ⏰ This will allow you to reprogram an LLM for forecasting. We aim to give you all the tools you need to tackle any forecasting problem, and continue to build a hub for foundation forecasting models. Finally, everything got even faster ⚡ with: ✨ Support for multi-node distributed training with Spark (see example notebook) 🐻 Support for polars Make sure to check our documentation for all the new models as well as brand new tutorials on: 📗 Cross-validation with neuralforecast 📘 MLFlow logging See the full set of updates in the release notes. ⭐️ Don't forget to star the repo and check our other libraries for machine learning and statistical methods. Happy Forecasting! 🚀📈💻 #timeseries #forecasting #opensource #python #deeplearning #machinelearning