skills.rest
K-Dense-AI
K-Dense-AIOFFICIAL
Explore all skills by this author →
Website𝕏@k_dense_aiGitHub

timesfm-forecasting

Forecast univariate time series with TimesFM point and quantile outputs.

33.0k|3.2k|Updated Oct 19, 2025
Search Tags:#calibration#uncertainty#forecasting#preflight#timeseries#timesfm#univariate
One-click install
npx skills add https://github.com/K-Dense-AI/scientific-agent-skills --skill timesfm-forecasting-k-dense-ai
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: timesfm-forecasting
Source: https://github.com/K-Dense-AI/scientific-agent-skills/tree/main/scientific-skills/timesfm-forecasting
Command: npx skills add https://github.com/K-Dense-AI/scientific-agent-skills --skill timesfm-forecasting-k-dense-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires timesfm, torch, pandas, numpy, and includes scripts (resource) and references (resource) components.

What problem does it solve?

Forecast univariate time series without training using the TimesFM foundation model.

Core Features & Use Cases

  • Zero-shot forecasting: forecast from a single pass over a univariate series.
  • Uncertainty estimates: produces point forecasts plus quantile-based intervals.
  • Preflight safety: includes a system checker to verify RAM/GPU/disk space before inference, preventing crashes.
  • Flexible inputs: accepts CSVs, DataFrames, or raw arrays for batch forecasting.

Quick Start

Inference with a single univariate time series loaded into a 1-D numpy array: create the TimesFM predictor, compile with a ForecastConfig, and call forecast() to obtain point and quantiles.

Frequently Asked Questions about timesfm-forecasting

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I forecast univariate time series without training a model?▼

You can forecast univariate time series without training by using the TimesFM foundation model to generate predictions directly from your input data in a single pass.

Can I generate calibrated prediction intervals for time-series forecasting?▼

Yes, you can generate calibrated prediction intervals by using the TimesFM model, which produces point forecasts alongside quantile-based uncertainty estimates for your time series.

What input formats work with TimesFM for batch forecasting?▼

TimesFM batch forecasting works with CSV files, pandas DataFrames, and raw numpy arrays, allowing you to process multiple univariate time series inputs flexibly.

Do I need a preflight system check before running TimesFM inference?▼

Yes, a mandatory preflight system check verifies your available RAM, GPU, and disk space before running TimesFM inference to prevent crashes during model execution.

How do I configure TimesFM to output point forecasts and quantiles?▼

You configure TimesFM outputs by creating a ForecastConfig object, compiling the predictor with it, and calling forecast() to obtain both point and quantile results.

What are the limitations of zero-shot time-series forecasting?▼

Zero-shot forecasting with TimesFM is limited to univariate series and requires specific versions (2.0 or 2.5) along with sufficient RAM, GPU, and disk space to run successfully.

Related Skills

Explore More Data & Analytics Skills→
crazymsncrazymsn
22

timesfm-forecasting

Forecast univariate time-series with TimesFM from CSV, DataFrame, or arrays.

Community
Advanced
SciMate-AISciMate-AI
1

timesfm-forecasting

Forecast univariate time series with TimesFM and quantile intervals.

Official
Advanced
jasrajtulsijasrajtulsi

timesfm-forecasting

Forecast univariate time series with TimesFM zero-shot inference.

Community
Advanced
Zeyuyang-0420Zeyuyang-0420

timesfm-forecasting

Forecast univariate time series with Google's TimesFM zero-shot model.

Community
Advanced