Spell Developer Plan | Free
This plan is for individual developers, including researchers, practitioners, consultants, and students building and using deep learning models.
Use GPU compute for model training with minimal configuration
Manage, track, and report on multiple experiments
Automate hyperparameter discovery and optimization
Spell provides on-demand GPU instances for most popular machine types (e.g., K80, V100) with no individual cloud account configuration needed
Utilizing the Spell public compute instances, Spell performs the entire setup process from finding an instance, building the AMI, preconfiguring environment with popular frameworks (e.g. Pytorch, TF), syncing local code with git, mounting necessary datasets to the instance, and executing the code, all under one single CLI command or console form
Web console provides an integrated interface to track, measure, and compare each experiment run. Users can utilize either built in metrics streaming or customize their own via the Python API
Access to numerous in-depth articles and tutorials on deep learning optimization, workflow simplification, and more
The Spell Developer plan offers totally free access to the most comprehensive platform for deep learning model development and training.