Run Ollama, Stable Diffusion and Automatic Speech Recognition with your Intel Arc GPU
A Docker Compose setup that lets Intel Arc GPU owners run Ollama, Stable Diffusion and Whisper speech recognition together behind Open WebUI, using a native SYCL or Vulkan llama.cpp backend for Ollama.
This project packages Docker containers for Intel Arc GPUs: Ollama built with a native ggml-sycl llama.cpp backend compiled against Intel's oneAPI, or an alternative Vulkan backend using the stock Ollama image that needs no custom build, behind Open WebUI as the chat interface. ComfyUI or SD.Next containers add Stable Diffusion image generation, and an optional Whisper container adds automatic speech recognition, all optimized for Intel Arc GPUs through Intel's Extension for PyTorch.
Open WebUI is configured with authentication disabled for ease of use, image generation enabled through Automatic1111 or SD.Next, and access limited to the Ollama API rather than OpenAI's. Runtime behavior such as context length, KV-cache type and GPU offload is tuned through an environment file, and Ollama exposes its usual port for other tools to connect to.