Moirai 2.0 is a decoder-only transformer model developed by Salesforce AI Research for universal time series forecasting. The small variant is pretrained on a mixture of public, synthetic, and internal operational data using improved techniques including quantile loss and multi-token prediction. It supports zero-shot and few-shot forecasting across many domains and is distributed with open weights on Hugging Face.
Moirai 2.0 R Small is a Data science & ML workbench product. Accurately forecasting future values in diverse time series data without task-specific model training. Moirai 2.0 R Small is an open-source project aimed at data scientists and forecasters. The project is open source (Apache-2.0). It runs on the web, the command line, and API.
Salesforce AI Research builds and maintains Moirai 2.0 R Small, and the product first shipped in 2024. The project is developed in the open on GitHub with 1.6k stars and 1 commits in the last 90 days. Among its 4 catalogued features are Time Series Forecasting, Quantile Loss, and Multi-Token Prediction.
Latest indexed changes and source events
Salesforce/moirai-2.0-R-small verified by the PulseGate indexer
Other apps tracked under the same category.
