Hugging Face's logo

Datasets:
THULab
/
fastest

Tasks:
Time Series Forecasting
Modalities:
Time-series
Languages:
English
Size:
n<1K
Tags:
forecasting
tsfile
format:tsfile
timeseries
License:
Dataset card Files
xet
 Community
fastest / README.md
smilegeng's picture
smilegeng
chore: add bare timeseries tag, remove ineffective modality field
6a6a972 verified
preview code
|
Raw History Blame Contribute Delete
1.44 kB
metadata 
language:
  - en
license: apache-2.0
task_categories:
  - time-series-forecasting
tags:
  - forecasting
  - tsfile
  - format:tsfile
  - timeseries
pretty_name: fastest
size_categories:
  - n<1K

fastest (TsFile)

Apache TsFile version of FaalSa/fastest.

Overview

A univariate time-series forecasting dataset of 25 series, provided in the standard GluonTS-style train / validation / test layout. The validation and test splits extend the train series by one and two prediction horizons respectively (validation = train + 120 points, test = train + 240 points).

train/      faalsa_fastest_train.tsfile        25 series × 4320 points = 108,000 rows
validation/ faalsa_fastest_validation.tsfile   25 series × 4440 points = 111,000 rows
test/       faalsa_fastest_test.tsfile         25 series × 4560 points = 114,000 rows

The original train / validation / test split is preserved — each split is a separate TsFile (not merged).

Schema (TsFile structure)

  • Time (INT64, milliseconds) — built from the per-series start timestamp and position.
  • item_id (TAG) — the series identity (25 devices).
  • target (FIELD, FLOAT) — the observed value.

Usage

Read the .tsfile files with the Apache TsFile Java or Python SDK.

Source & license