Practical overview
When it fits
Use it for a tabular regression or classification task when you want an automated baseline and Python artifacts that can be inspected and adapted outside the tool.
Who it is for
- Developers who need model and pipeline code to integrate into prediction workflows.
- Data scientists and researchers seeking automated baselines for tabular prediction tasks.
- Citizen data scientists and engineers who want to start model training without manually defining preprocessing and model configurations.
Documented uses
- Search model and hyperparameter configurations for a selected tabular target.
- Review and modify generated data-processing and model code.
- Use generated scripts to make predictions or retrain a model on new data.
Requirements & limits
The project is distributed as a Python package and identifies its repository license as MIT.
- The documented supported frameworks are TensorFlow through Keras and XGBoost; CatBoost and LightGBM are listed as planned, not current support.
- The README describes pandas, scikit-learn and the chosen framework as the minimal dependency set.
- Trial metrics and hyperparameters are saved to a results CSV, and results are saved after each trial so a search can be stopped without losing prior trial output.
- Generated code can be used without continuing to run Automl Gs, but the project frames generated models as baselines that can be improved.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Generate a model and pipeline from a CSV
Train a tabular model and create Python files for later use in a prediction workflow.
Before you start
- Python 3
- A CSV file in the current directory and a target column to predict
- The relevant framework installed: TensorFlow or XGBoost; pandas and scikit-learn are also listed as dependencies
Steps
Install the package with the command below.
pip3 install automl_gsInstall the framework you plan to use.
Run
automl_gs data.csv TargetColumn, replacing the example names with your CSV path and target field.Inspect the generated model and pipeline files, requirements, encoders, model artifacts and
automl_results.csvin the timestamped output folder.
Expected outcome
A trained model, native Python model and pipeline scripts, and a CSV of training results and hyperparameters.
Guide limitations
- The CSV path must be in the current directory.
- The documented framework support covers regression and classification with TensorFlow through Keras or XGBoost.
- The project describes generated models as baselines that may need further tuning.
Use the generated script to predict or retrain
Run prediction or retraining with the model script produced by Automl Gs.
Before you start
- A completed Automl Gs output folder containing
model.pyand the generated model files - A CSV file to use as prediction or retraining data
Steps
Change to the generated output folder.
For prediction, run the command below.
python3 model.py -d data.csv -m predictTo retrain on new data, run the command below.
python3 model.py -d data.csv -m train
Expected outcome
The generated script is invoked in prediction or training mode with the supplied CSV.
Guide limitations
- The README documents the commands but does not specify the output format or detailed data requirements for these modes.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 16 October 2026
- Original website title
- minimaxir/automl-gs · Official homepage title · retained in its source language