Moirai-1.0-R-Large
Model directorySalesforce/moirai-1.0-R-large311M params | Upstream checkpoint
Moirai-1.0-R-Large is the highest-capacity dense checkpoint in Salesforce's original Moirai family, the top tier above Moirai-1.0-R-Small and Moirai-1.0-R-Base. At 311M parameters it is the first-generation Moirai design at its largest published scale, aimed at the strongest broad zero-shot performance the dense 1.0-R line offers.
It preserves the same masked-encoder transformer as the smaller checkpoints, with multi-patch projections, any-variate attention, and a mixture-distribution output head. The any-variate attention handles arbitrary numbers of target variables and dynamic covariates in one model, and the mixture output head yields probabilistic forecasts with calibrated quantiles. Like the rest of the family it is pretrained on the LOTSA corpus, the large open time-series archive behind the Moirai work, which underpins its zero-shot generalization.
Where to run this model
Compute availability, hardware requirements, and pricing depend on the execution provider.
Get model weights, deployment instructions, and current provider listings.
- sktime TServe
self-hosted runtime · Family / provider variant · Source checked 2026-10-06
Run a model-serving container on your own hardware or compute account. Select the catalog checkpoint and matching runtime image.
Hugging Face provider mappings checked 2026-10-06. No Inference Provider mapping was listed for this checkpoint.
Model Classification
Family
Moirai-1.0-R
Type
time series foundation model
Pretrained time-series model for zero-shot or few-shot forecasting workloads.
Resources
Training Data
LOTSA, the Large-scale Open Time Series Archive, with roughly 27B observations across nine domains including energy, transport, finance, healthcare, sales, climate, web, and social data.
Recommended For
- • Multivariate forecasting across heterogeneous domains
- • Workloads that benefit from probabilistic outputs and arbitrary variate counts
Strengths
- • Strong multivariate coverage across the Moirai family
- • Well-suited to covariates and correlated series
Limitations
- • Model cards for some newer Moirai variants are still sparse on exact checkpoint details
- • Heavier family choices can be more expensive than tiny single-purpose baselines
Capabilities
Tags
Specifications
- Parameters
- 311M
- Architecture
- masked encoder transformer with multi-patch projections, any-variate attention, and mixture-distribution output
- Context length
- n/a
- Max context
- n/a
- Minimum history
- n/a
- Recommended history
- n/a
- Input step
- n/a
- Required target series
- 1
- Temperature
- Ignored
- Top P
- Ignored
- Max output
- n/a
- License
- n/a