🚀 New Open-Source Package for Time Series Forecasting

Hi everyone,

We’ve been working on a new forecasting package that we believe pushes the boundaries of current state-of-the-art tools. After extensive benchmarks, it consistently outperformed many leaders — not just in accuracy (lower MAPE and sMAPE across diverse datasets), but also in training efficiency.

What makes it different?

⚡ Scalability: Handles large datasets with ease.

🎯 Accuracy: Outperforms popular baselines on real-world benchmarks.

🛠️ Ease of Use: Simple API designed with practitioners in mind.

🔄 Production-ready: Fast training and inference make it suitable for real-time applications.

We’re really excited about the results and would love for the community to try it out, share feedback, or challenge it with your own datasets.

👉python package

👉R package

Looking forward to your thoughts and discussions!