Hedra launches custom model training and deployment platform

Hedra launched Model Factory on November 10, offering teams GPU capacity, researcher-assisted model training and deployment using their own data and simulations. The company says customers can rent GPU sandboxes without long-term commitments or bring existing models to its inference engine. Hedra researchers work with customers and autonomous agents to build training environments, evaluate models and analyze production performance. Deployment options include customer premises and partner data centers. After deployment, researchers retrain models and release only versions that pass the customer's evaluations.