Spaces:

---

Duplicated from  demo-leaderboard-backend/leaderboard

Salesforce

/

GIFT-Eval

like 159

Running

I am currently exploring multivariate, multi-step minutely time series forecasting, specifically:

Predicting tens to hundreds of variables simultaneously

Minute-level granularity (e.g., 15-minute intervals)

Forecast horizon of ~96 steps

Support for covariates as input

Models that can be fine-tuned on my own datasets

I have been looking into models such as TimesFM, Moirai, Chronos, and TTM, but I am feeling somewhat lost on which approach would be most suitable for my scenario.

Could you kindly provide guidance or recommendations on which models are most appropriate for these requirements, especially regarding multivariate support, covariates, and fine-tuning capabilities? Any pointers would be greatly appreciated.

Thank you very much for your time and assistance.