Dataset · autogluon
Time series datasets for Chronos
autogluon/chronos_datasets
Time series collections for training and evaluating Chronos forecasting models.
What's inside
This dataset brings together the time series used to train and evaluate Chronos models. Subsets requiring a custom builder script have been moved to a separate repository, chronos_datasets_extra. The recommended way to work with the data is via the fev library; it includes tabular, text, and time series modalities, with a size of 10M–100M.
What you can do with it
- Training and fine-tuning time series forecasting models
- Benchmarking Chronos and comparing it against other forecasters
- Validating pipelines using fev
- Tasks
- Time series forecasting
- Size
- 10M–100M rows
- Downloads this month
- 43K
- Likes
- 80
Models trained on it
- autogluon/chronos-2Time series forecasting 41.5M
- amazon/chronos-2Time series forecasting 9.2M
- autogluon/chronos-2-smallTime series forecasting 2.3M
- NX-AI/TiRexTime series forecasting 109.5K
- thuml/sundial-base-128mTime series forecasting 52.9K
- Datadog/Toto-Open-Base-1.0Time series forecasting 43.1K
- NX-AI/TiRex-2Time series forecasting 3.6K
- NX-AI/TiRex-1.1-giftevalTime series forecasting 1.7K
How to load it
from datasets import load_dataset
ds = load_dataset("autogluon/chronos_datasets")