Unweave

Open source platform for machine learning dev environments.

Freemium

Description

Unweave is an open-source tool engineered to streamline machine learning development workflows. It provides a unified platform for accessing and managing GPU-powered development environments, allowing users to SSH into instances from multiple cloud providers such as Unweave's own, LambdaLabs, GCP, and AWS. The platform aims to remove the complexities of cloud infrastructure setup, offering sane defaults while retaining flexibility for customization.

Users can manage their projects and sessions through a cross-platform Command Line Interface (CLI) that integrates with existing local or browser-based IDEs, or via a web-based Dashboard for a visual overview. Unweave focuses on transparent cloud costs, converting complex provider pricing into an easily understandable format, and supports finding the best GPU price and availability across different services. Future developments include enhanced Git integration for seamless code, data, and artifact versioning.

Key Features

Cross-Cloud GPU Access: SSH into GPU machines from providers like Unweave's own, LambdaLabs, GCP, and AWS.
Transparent Cloud Costs: Converts complex cloud provider pricing into a simple, parseable format, with no markup on user-provided cloud accounts.
CLI & Dashboard Management: Control environments via a cross-platform terminal CLI or a visual web dashboard.
Zero Setup Remote GPUs: Automatically configures and launches cloud dev environments.
Optimized Defaults & Flexibility: Provides sane defaults that work out-of-the-box, with options for customization.
Multi-Provider Search & Selection: Find and choose the best GPU price and availability across different cloud services, with auto-selection on paid plans.
IDE Integration: Works with existing code editors, whether local or browser-based.
Open Source: Community-driven development with code available on GitHub.
Git-Based Workflow (Coming Soon): Features planned for automatic versioning of code, data, artifacts, and seamless cloning from Git repositories.

Use Cases

Developing and training machine learning models on remote GPUs.
Managing and standardizing ML development environments across teams.
Accessing diverse GPU types from multiple cloud providers seamlessly.
Optimizing cloud spending for ML workloads through transparent pricing and provider choice.
Simplifying the setup of cloud infrastructure for ML projects.

Frequently Asked Questions

Which cloud providers does Unweave support?

Unweave allows you to connect to GPU machines on its own hosted platform (on paid plans) and supports bringing your own accounts for LambdaLabs, GCP, and AWS.

Does Unweave add a markup to cloud provider costs?

Unweave states it does not charge a markup on the pricing of cloud providers when you bring your own account. For its hosted Unweave cloud provider option, GPU costs are part of its service pricing.

Will the free plan always remain free?

Yes, Unweave states that its free-forever plan will always be free.

What are the limitations of the free plan?

The free plan allows up to 3 projects and up to 5 concurrent sessions. It requires users to bring their own cloud providers, with direct integration for LambdaLabs accounts mentioned.