OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Models

Entrope

EntroPE is an open-source Python reference implementation for long-term time-series forecasting. It places dynamic patch boundaries using a frozen, pretrained causal GPT entropy model, then forecasts with a patch encoder, global transformer and fusion decoder.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Entrope — OpenViST project profile in AI Models.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use EntroPE when working with its documented long-term forecasting benchmarks and you need a reference implementation with separate evaluation, forecasting-training and optional entropy-model-training stages.

Who it is for

  • Researchers reproducing long-term forecasting benchmark results.
  • Practitioners experimenting with entropy-guided segmentation in transformer forecasting.
  • Developers evaluating provided checkpoints or training forecasting models on the supported datasets.

Documented uses

  • Evaluate supplied forecasting checkpoints across the repository's dataset and horizon configurations.
  • Train forecasting models from scratch on supported benchmark CSVs.
  • Optionally train the entropy models used to determine dynamic patch boundaries.
Before you choose

Requirements & limits

The implementation is in a public GitHub repository. Benchmark data and forecasting checkpoints are hosted on Hugging Face; the frozen entropy model checkpoints are included in the repository.

  • The method combines entropy-guided patch placement with a patch encoder, global transformer and fusion decoder.
  • Evaluation requires both forecasting checkpoints and the separately loaded entropy model checkpoints.
  • The documented metrics are computed in standardized (z-scored) space.
  • The documented environment is Python 3.10, PyTorch 2.x and CUDA 12.x, with one GPU described as sufficient.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Set up the EntroPE forecasting environment

Create the documented Python environment and install EntroPE's dependencies.

Before you start

  • Conda
  • Python 3.10

Steps

  1. Create a Conda environment named entrope with Python 3.10: conda create -n entrope python=3.10 -y.

  2. Activate the environment: conda activate entrope.

  3. From the repository directory, install the dependencies: the command below.

    Documented command
    pip install -r requirements.txt

Expected outcome

The EntroPE Python environment is created and its requirements are installed.

Guide limitations

  • The repository reports testing with Python 3.10 and PyTorch 2.x (CUDA 12.x); these are reported test conditions, not a guarantee for other configurations.
  • This recipe covers setup only. The forecasting data and checkpoints are hosted on Hugging Face; frozen GPT entropy-model files ship in the repository and are required for evaluation.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
18 October 2026
Original website title
Sachithx/EntroPE · Official homepage title · retained in its source language
How we review sourcesSuggest a correction