# TIME Benchmark

Canonical source: https://huggingface.co/datasets/Real-TSF/TIME

Official leaderboard: https://huggingface.co/spaces/Real-TSF/TIME-leaderboard

Paper: https://arxiv.org/abs/2602.12147

## Dataset Type

Contamination-resistant task-centric zero-shot forecasting benchmark.

## Temporal Structure

TIME is described as a benchmark of fresh forecasting datasets and task configurations aligned with real-world operational requirements. The dataset artifact archives task-level data and window-level prediction results for benchmark visualization.

## Actions Or Interventions

None. TIME is a passive forecasting benchmark.

## Reported Scale

The Hugging Face dataset card describes TIME as 50 fresh datasets and 98 forecasting tasks. The artifact is about 220 MB.

## Suitability Note

TIME is relevant to Toto 2.0 because Datadog uses it for contamination-resistant zero-shot claims. It should be treated as a benchmark-hygiene source rather than an observability-specific or high-dimensional-channel benchmark.

## Access And License Notes

The Hugging Face dataset page lists Apache-2.0.