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

Rtdl Revisiting Models

Rtdl Revisiting Models is an Apache-2.0-licensed Python research implementation of deep-learning architectures for tabular data, including FT-Transformer, MLP and ResNet. It includes a Python package, experiment code, configurations and reported results.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Rtdl Revisiting Models — 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 it when you want to inspect or build on the paper’s tabular deep-learning experiments, particularly comparisons between FT-Transformer and MLP-like baselines.

Who it is for

  • Researchers comparing neural-network architectures for structured datasets.
  • Python practitioners prototyping tabular classification or regression experiments.
  • Readers examining the metrics and tuned configurations reported in the accompanying paper.

Documented uses

  • Compare FT-Transformer with MLP-like baselines on tabular benchmarks.
  • Explore reported metrics and tuned hyperparameters across datasets.
  • Use the included implementations as a starting point for tabular classification or regression research.
Before you choose

Requirements & limits

  • The main documented environment is PyTorch; a separate TensorFlow environment is described only for experimenting with TabNet.
  • The README reports FT-Transformer as slower than MLP-like models.
  • Reproducing experiments requires obtaining datasets, and the licenses of dataset components apply.
  • The repository README points to TabM as a newer tabular deep-learning model.

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 PyTorch research environment and prepare the dataset

Create the documented Conda environment and unpack the project dataset in the repository root.

Before you start

  • Conda installed.
  • A repository root path represented by an absolute path.
  • Acceptance of the dataset components’ applicable licenses.

Steps

  1. Set PROJECT_DIR to the absolute repository root, clone the project into that path, and change to the directory.

  2. Create and activate the revisiting-models environment with Python 3.8.8; install the listed PyTorch, torchvision, CUDA toolkit, NumPy, cuDNN, requirements, Node.js, and JupyterLab widget manager dependencies.

  3. Set the documented Conda environment variables for PYTHONPATH, PROJECT_DIR, LD_LIBRARY_PATH, CUDA_HOME, and CUDA_ROOT; deactivate and reactivate the environment.

  4. Download the dataset archive, move it to $PROJECT_DIR, change to that directory, and unpack the archive.

Expected outcome

The documented PyTorch environment is configured, and the dataset archive is unpacked in the repository root.

Guide limitations

  • The repository’s setup instructions specify particular package versions and CUDA-related dependencies.
  • The dataset’s component licenses apply; the project states it imposes no additional restrictions.
  • These steps cover setup and data preparation, not model training or inference.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
14 October 2026
Original website title
yandex-research/rtdl-revisiting-models · Official homepage title · retained in its source language
How we review sourcesSuggest a correction