Sign up to save tools and stay up to date with the latest in AI
bg
bg
1

GitHub - wearesulie/sulie: Access to Sulie foundation models for time-series forecasting 📈

Nov 12, 2024 - github.com
The article introduces the Sulie SDK, a platform for advanced time series forecasting powered by Mimosa, a transformer-based foundation model optimized for time series data. The Sulie SDK offers features like zero-shot forecasting, automatic fine-tuning, and seamless integration with the Sulie platform. It also handles MLOps tasks, allowing users to focus on utilizing the forecasts. To use the Sulie SDK, an API key is required, which can be generated from the Sulie Dashboard.

The article also provides a guide on how to get started with the Sulie SDK, including installation and initialization using the API key. It details the features of the SDK, such as forecasting with Mimosa, fine-tuning for customized forecasting, managing datasets, and forecasting with custom models. The article concludes by providing additional resources like API documentation, forecasting guide, fine-tuning guide, and support for assistance and feedback on the SDK.

Key takeaways:

  • The Sulie SDK provides seamless integration with the Sulie platform for advanced time series forecasting powered by Mimosa, a transformer-based foundation model optimized for time series data.
  • Users can generate accurate time series forecasts using Mimosa’s zero-shot inference capabilities, ideal for fast, reliable predictions without training the model.
  • With automatic fine-tuning, users can optimize Mimosa for unique datasets and business cases, using Weighted Quantile Loss (WQL) for evaluation to ensure high accuracy.
  • Sulie’s Dataset API allows users to manage and version their datasets, making them accessible for forecasting and fine-tuning across teams.
View Full Article

Comments (0)

Be the first to comment!