Accelerate your deep learning development lifecycle

Determined AI streamlines your deep learning workflows, tracks your work, and manages your GPU clusters.

Determined AI Platform

Determined AI takes a pragmatic, results-driven approach to deep learning, with a goal of dramatically improving the productivity of deep learning developers. Our integrated AutoML platform simplifies the entire deep learning workflow from data management to model training and deployment. We manage your heterogeneous hardware and optimize your GPU resource utilization.

Determined AI

Integration

Product Stack

AutoML at scale

Deliver quality models in less time

  • State-of-the-art hyperparameter and neural architecture search
  • Push-button transfer learning and retraining
  • Automatic data-parallel distributed training
  • Data access accelerator
  • TensorFlow, Keras, or PyTorch support

Seamless infrastructure

Optimize resource utilization and ensure experiment reliability

  • Fault-tolerant training
  • Efficient scheduling and execution of experiments
  • Share GPU resources on premises, in the cloud, or both.
  • Run on bare metal, and in Kubernetes

Experiment tracking

Enable reproducibility and collaboration

  • Integrated metrics capture and model version management
  • Reproducible, containerized training
  • Label and share experiments
  • Multi-tenancy in a shared infrastructure
  • Real-time experiment metrics visualization

Model deployment optimization

Ensure model deployability and reduce time-to-market

  • Explore models with deployment constraints
  • Simulate inference speed on deployment hardware
  • Automated architecture search for the edge, cloud, and mobile
  • Containerized deployment in a single click

Recent posts

AUG 19, 2019

Specialized AI chips hold both promise and peril for developers

AUG 13, 2019

[Product feature series] One-click access to TensorBoard for model development and experimentation

JUN 04, 2019

The cloud giants have an AI problem

Stay updated