More Productive Data Science

Spend less time wrangling machines and accelerate development with the workload and application management solution designed with deep learning in mind.

Simplified Operations

Whether managing servers yourself or working with a DevOps or IT team — now you have a simpler, more powerful way to manage deep learning resources.

GPU as a First-Class Citizen

No more unfair trade offs — our platform lets you gain both the flexibility, versioning, and leanness of containers as well as the immense computation power of GPUs.

Elasticity and Efficiency

AI infrastructure doesn’t have to be a money pit. Maximize resource utilization with smart scheduling and auto-allocation across multiple concurrent workloads and users.

Ready to start innovating?

No Software Setup

All the tools you need for development are pre-installed so you can get straight to modeling.
Deep Learning Frameworks

We include the most popular open source frameworks such as Tensorflow (plus Tensorboard and Tensorflow Serving), Caffe, MXNet, Chainer, Torch, Theano, and others.

Data Science Libraries

Sometimes the task at hand doesn’t require a neural net in order to get the job done. That’s why NumPy, SciPy, Scikit Learn, Pandas, and much more come standard with Bitfusion.

GPU Libraries and Drivers

A common tripping point for AI developers is managing GPU driver complexity. We ensure CUDA, CUDNN, and the underlying GPU drivers are out of sight, out of mind.

Example Code and Data Sets

When you are experimenting with a new neural net architecture or learning deep learning in the first place, it helps to have examples and sample data for getting your feet wet.

Interactive Workspaces

For the fastest way to get started developing with deep learning, create an interactive
workspace that connects seamlessly with your code and data.
Notebooks

Fully-integrated Jupyter, for a sharable GUI development environment that supports 40 different languages, including Python, Julia, R, and Scala.

Terminal Access

Jupyter includes a web-based Terminal session to make it easy for you to navigate subdirectories and pull in additional packages or code via command line.

Code and Data Saving

Spin up and down workspaces as needed — your code and data is instantly available when you start and automatically preserved when you shut down.

Job Runner Preview

Kick off workloads anytime, anywhere that run in the background. Maximize the utilization of your cluster with multiple, parallel model runs and by sharing resources across your team or company.
Local CLI

Download our Bitfusion command line interface and initiate jobs from your own local environment or development box.

Workspace CLI

Workspaces come with the Bitfusion CLI pre-bundled so you can initiate jobs from your interactive development environment.

Parallel Training

Train multiple models simultaneously for 10X faster development and rapid hyperparameter exploration.

Status & Notifications

Track the quantity and progress of your workloads via GUI or CLI, and get notified automatically when your jobs are complete.

Model Deployment Preview

Ready to deploy your model for inference? We’ve made it way simpler to expose and manage API endpoints.
Tensorflow Serving

Currently supports Tensorflow Serving, which provides a foundation for versioning, deploying, and managing Tensorflow models.

Endpoint Management

Automates the deployment of an API server and URL endpoint based on your model export and serving files.

Smart Resource Management

Managing a large cluster of compute resources across multiple users and multiple workloads just got a lot simpler. Our management software takes care of initial placement, elastic resource attachment, event logging, and ongoing monitoring and health.
Auto-Allocation

Resources are automatically assigned and attached based on the workload requirements with intelligent scheduling, queuing, and job management.

Bitfusion Core

At the heart of the platform is our groundbreaking Bitfusion Core GPU virtualization, which provides unmatched resource flexibility and elasticity.

Advanced Cluster
Management Enterprise

For startups or enterprises that deploy our software platform into their cloud or data center, Bitfusion makes shared, heterogeneous cluster infrastructure simpler and more efficient to manage.
Custom Containers

Upload and manage your own container images to use as base workspace and job configurations, tailored for your unique requirements.

Workspace Snapshots

Create a new base workspace or job container image out of an active workspace. Leverage snapshots for just yourself, or share them with others.

User Management

Administrate multiple users and groups with different permission tiers to fit your team’s business processes and policies.

Node Management

Manage and monitor your cluster infrastructure, track utilization and system events, and modify configurations and policy as needed.

 

Bitfusion Core

Our transparent middleware layer combines multiple systems into a single elastic compute cluster that supports sharing, scaling and management of heterogeneous compute resources.
GPU Superpowers

Co-processor compute virtualization enables powerful new capabilities including on demand elastic GPUs, seamless multi-node scaling, and automatic high availability.

Runs in Userspace

Bitfusion Core runs in userspace, ensuring it can run securely, on almost any OS, and without requiring any changes to existing hypervisors or cloud infrastructure. 

No Code Changes

As completely transparent middleware, Bitfusion Core requires zero application changes. That’s why including Core in the AI Platform makes for a perfect combination.

See how Apposphere leverages Bitfusion to provide a more powerful GPU cloud services offering.

 

See the Software in Action

Schedule a Demo