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

Limix

LimiX is a family of pretrained foundation models and Python inference tools for structured and tabular data. Its documented tasks are classification, regression and missing-value imputation, using downloadable checkpoints through a command-line interface or the LimiXPredictor Python entry point.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Limix — 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 LimiX when a structured-data workflow calls for pretrained inference across classification, regression or feature imputation, and the dataset can be prepared in the repository's documented CSV layout.

Who it is for

  • Researchers and developers working with structured or tabular datasets
  • Python practitioners who need pretrained models for classification, regression or missing-value imputation

Documented uses

  • Run classification on tabular datasets
  • Estimate regression targets from structured data
  • Impute missing feature values
  • Prepare benchmark-formatted CSV datasets for model inference
Before you choose

Requirements & limits

The repository documents Python 3.12 or later. LimiX-2 usage is subject to a non-commercial license.

  • LimiX-2 is documented for all three supported tasks without task-specific parameter updates.
  • The repository also lists 2M- and 16M-parameter checkpoints, with different configuration files for those models.
  • LimiX-infer expects one dataset per subdirectory and task-specific CSV inputs.
  • The documented installation uses Python packaging; optional PyTorch and flash-attn installations must match local CUDA.

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

Prepare tabular CSVs for LimiX command-line inference

Arrange benchmark data in the directory structure LimiX-infer expects.

Before you start

  • A benchmark root directory
  • One subdirectory per dataset
  • CSV files with headers
  • For classification or regression, train and test CSVs with the target in the final column
  • For feature imputation, a masked-data CSV whose filename suffix matches the mask ratio

Steps

  1. Create a dataset subdirectory inside the benchmark root.

  2. Add _train.csv and _test.csv.

  3. For feature imputation, also add _0.05.csv when using the documented default mask ratio of 0.05.

Expected outcome

A benchmark directory structured for LimiX-infer to read; folders without the required CSVs are skipped.

Guide limitations

  • This guide covers input preparation only, not running inference.
  • The documented imputation filename suffix follows the --mask_ratio setting.
Documentation 1 ↗
Source-based practical guide

Install the LimiX Python package

Install the repository's Python package and expose its inference entry points.

Before you start

  • Python 3.12 or later
  • Git
  • PyTorch and flash-attn versions aligned with the local CUDA environment if installing those optional dependencies

Steps

  1. Clone the LimiX repository.

  2. Change into the repository directory.

  3. Run python -m pip install -e .

Expected outcome

The documented installation provides the LimiX-infer command and makes LimiXPredictor importable from the inference.predictor and limix modules.

Guide limitations

  • The repository says PyTorch and flash-attn must match local CUDA when installed.
  • This is an installation guide; it does not demonstrate a completed inference run.
Documentation 1 ↗
COMPARE OPTIONS

Compare options for Model inference and serving

These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.

ProjectDocumented purposePlatformAccess and requirements
LimixRun classification on tabular datasetsCloudThe repository documents Python 3.12 or later. LimiX-2 usage is subject to a non-commercial license.
Pulse Of MotionEstimate a video's physical frame rate from its visual dynamics.LinuxThe repository is public. Its README describes a released checkpoint that downloads from Hugging Face on first inference, with a manual download option. The strongest internal model is not fully released; the README directs interested users to contact the project about access. No license information is provided in the repository evidence.
Neural Image AssessmentTrain the model using AVA images and rating annotations.Cloud, Linux, macOS, Windows
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
16 October 2026
Original website title
limix-ldm-ai/LimiX · Official homepage title · retained in its source language
How we review sourcesSuggest a correction