# Core-cuda-bionic

Core-cuda-bionic is a CUDA (Compute Unified Device Architecture) browser workspace image that lets users run data-heavy and AI software in a web browser with graphics processing unit acceleration, without needing a. It is open source and self-hosted, installable on Docker Hub.

Canonical page: https://ourlocalhost.com/app/6461/core-cuda-bionic

## Description

Core-cuda-bionic is a CUDA (Compute Unified Device Architecture) browser workspace image that lets users run data-heavy and AI software in a web browser with graphics processing unit acceleration, without needing a local high-end computer.

## Platform availability

Available on 0 of 10 platforms.

## Links

- Docker Hub: https://hub.docker.com/r/kasmweb/core-cuda-bionic

## Common questions

### What OOM and token problems should I troubleshoot in self-hosted AI?

The troubleshooting guide points to out-of-memory model loads, slow tokens, missing models, and service failures.

### Can a GPU/OpenCL core fail to detect a device?

Yes, a forum thread reports Core 0x22 unable to detect an OpenCL device, with community replies in that thread.

### What Stable Diffusion GitHub setup issue is a user describing?

A user says they are trying Stable Diffusion from GitHub, want to use a high-end PC, and have many issues and questions.

---
Source: Our Local Host (https://ourlocalhost.com/app/6461/core-cuda-bionic), the self-hosted and open-source app directory.
