gradsflow/gradsflow-automl
An open-source AutoML Library based on PyTorch
What it solves
GradsFlow simplifies the process of building and training deep learning models. It removes the complexity of manual setup by providing tools for automated machine learning (AutoML) and streamlined model training, allowing users to run experiments on their local laptops or scale them to remote clusters.
How it works
The library provides a modular set of tools built on top of PyTorch. For beginners, the AutoTasks component enables zero-code model building for tasks like image recognition and text analysis. For more experienced users, the Model API offers a customizable way to train any PyTorch model, while the Tuner component automates hyperparameter optimization to improve model performance with minimal code changes.
Who it’s for
It is designed for both beginners who need an easy, automated path into machine learning and experts who want a flexible, extensible framework to manage complex training workflows and hyperparameter tuning.
Highlights
- Provides zero-code AutoML for common tasks like image recognition and sentiment analysis.
- Includes a customizable Model Training API compatible with PyTorch.
- Offers a simplified API for hyperparameter tuning.
- Supports scaling experiments from a local laptop to remote clusters.
- Integrates with PyTorch Lightning Flash for advanced task support.
Related
- Project
- Project
- Project
- Project