timesfm-forecasting

Forecast univariate time series with Google's TimesFM model.

Updated Aug 23, 2026
One-click install
npx skills add https://github.com/Rowtion/Bioclaw --skill timesfm-forecasting
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: timesfm-forecasting
Source: https://github.com/Rowtion/Bioclaw/tree/main/scientific-skills/timesfm-forecasting
Command: npx skills add https://github.com/Rowtion/Bioclaw --skill timesfm-forecasting

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires torch, numpy, pandas, scikit-learn, matplotlib, huggingface_hub, safetensors, pypdf2, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill automates time series forecasting for any univariate data, eliminating the need to train custom models or tune classical statistical parameters.

Core Features & Use Cases

  • Zero-Shot Forecasting: Predicts future values for sales, sensor readings, stock prices, or any other time series using Google's TimesFM foundation model.
  • Probabilistic Forecasts: Provides calibrated prediction intervals (quantiles) for uncertainty estimation.
  • System Requirements Check: Includes a mandatory preflight script to ensure your machine has sufficient RAM/GPU before loading the large model.
  • Use Case: Forecast next year's monthly website traffic or predict upcoming energy demand based on historical patterns, all without model training.

Quick Start

Run the python script scripts/check_system.py to verify your machine's compatibility before using the forecasting model.

Frequently Asked Questions about timesfm-forecasting

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

FAQPage Schema
How do I perform zero-shot time series forecasting without training a custom model?

Zero-shot time series forecasting uses Google's TimesFM foundation model to predict future values directly from historical data, eliminating custom model training. You provide univariate inputs via CSV, DataFrame, or array.

Can I get prediction intervals for time series forecasts using Python?

Probabilistic forecasting provides calibrated prediction intervals, or quantiles, for uncertainty estimation alongside point forecasts. This allows you to quantify the confidence range of future time series predictions in Python.

What are the system requirements for loading TimesFM for univariate forecasting?

Loading the TimesFM model requires sufficient RAM, GPU, and disk space. A preflight system checker script verifies these hardware requirements before model loading to prevent memory allocation errors during time series forecasting.

What input formats does AI time series forecasting support?

AI time series forecasting accepts CSV files, Pandas DataFrames, and NumPy arrays as input formats. This flexibility allows direct integration with existing Python data pipelines for univariate prediction tasks.

What is the best way to forecast future sales or energy demand without statistical parameter tuning?

Zero-shot forecasting with a foundation model bypasses classical statistical parameter tuning by pre-learning patterns from large datasets. It predicts upcoming sales or energy demand directly from historical sequences.