vEdge Affordable GPU Compute for Machine Learning

What is vEdge?

vEdge provides access to affordable GPU compute power by utilizing virtual GPU (vGPU) technology. This approach allows multiple applications or users to share the same physical hardware, significantly lowering operational costs compared to dedicated GPU instances. The platform is specifically tailored for machine learning workflows, supporting both model training and deployment.

Users benefit from rapid instance start times, typically under 10 seconds, enabling quick scaling for demanding tasks like parallel hyperparameter tuning. vEdge supports popular ML frameworks and allows deployment flexibility through pre-configured images with common ML tools or by using custom Docker images. The service aims to offer substantial savings compared to traditional cloud providers, making GPU acceleration more accessible.

Features

Shared GPU Hardware (vGPUs): Lowers costs by allowing multiple users/apps to share the same GPU.
Cost Savings: Aims for significant savings compared to using whole GPUs or other providers (estimated 5x-10x).
Fast Instance Start Time: Instances launch in less than 10 seconds.
Scalable ML Training: Facilitates easy scaling for parallel workflows like hyperparameter tuning.
Flexible Deployment Options: Supports deployment via pre-built ML images or custom Docker containers.
Affordable Pricing Structure: Plans include low hourly rates (starting around $0.20/h) with no intended egress fees.

Use Cases

Training machine learning models cost-effectively.
Deploying GPU-accelerated applications and AI models.
Running parallel hyperparameter tuning experiments.
Scaling machine learning development workflows quickly.
Developing AI applications with reduced infrastructure costs.

FAQs

How is vEdge different from other compute providers like AWS?

vEdge shares GPU hardware across multiple users which drives down costs and makes scaling a lot faster and easier.
When will vEdge become publicly available?

vEdge plans to release in early 2025. Early access will be provided to users who sign up for the waitlist.
How much will vEdge cost?

Pricing is not finalized yet. However, vEdge plans to offer lower power instances starting around $0.20/h and does not plan to charge for egress.
How can I contact vEdge?

You can contact vEdge by email: contact@vedge.cloud