IPEX-LLM

info

IPEX-LLM is a PyTorch library for running LLM on Intel CPU and GPU (e.g., local PC with iGPU, discrete GPU such as Arc A-Series, Flex and Max) with very low latency.

IPEX-LLM supports accelerated Ollama backend to be hosted on Intel GPU. Refer to this guide from IPEX-LLM official documentation about how to install and run Ollama serve accelerated by IPEX-LLM on Intel GPU. You can then configure Continue to use the IPEX-LLM accelerated "ollama" provider as follows:

config.json

{
  "models": [
    {
      "title": "IPEX-LLM",
      "provider": "ollama",
      "model": "AUTODETECT"
    }
  ]
}

If you would like to reach the Ollama service from another machine, make sure you set or export the environment variable OLLAMA_HOST=0.0.0.0 before executing the command ollama serve. Then, in the Continue configuration, set 'apiBase' to correspond with the IP address / port of the remote machine. That is, Continue can be configured to be:

config.json

{
  "models": [
    {
      "title": "IPEX-LLM",
      "provider": "ollama",
      "model": "AUTODETECT",
      "apiBase": "http://your-ollama-service-ip:11434"
    }
  ]
}

tip

If you would like to preload the model before your first conversation with that model in Continue, you could refer to here for more information.