Using Ollama in LobeChat
Ollama is a powerful framework for running large language models (LLMs) locally, supporting various language models including Llama 2, Mistral, and more. Now, LobeChat supports integration with Ollama, meaning you can easily enhance your application by using the language models provided by Ollama in LobeChat.
This document will guide you on how to use Ollama in LobeChat:
Using Ollama on macOS
Local Installation of Ollama
Download Ollama for macOS and unzip/install it.
Configure Ollama for Cross-Origin Access
Due to Ollama's default configuration, which restricts access to local only, additional environment variable setting OLLAMA_ORIGINS is required for cross-origin access and port listening. Use launchctl to set the environment variable:
launchctl setenv OLLAMA_ORIGINS "*"After setting up, restart the Ollama application.
Conversing with Local Large Models in LobeChat
Now, you can start conversing with the local LLM in LobeChat.
Using Ollama on Windows
Local Installation of Ollama
Download Ollama for Windows and install it.
Configure Ollama for Cross-Origin Access
Since Ollama's default configuration allows local access only, additional environment variable setting OLLAMA_ORIGINS is needed for cross-origin access and port listening.
On Windows, Ollama inherits your user and system environment variables.
- First, exit the Ollama program by clicking on it in the Windows taskbar.
- Edit system environment variables from the Control Panel.
- Edit or create the Ollama environment variable
OLLAMA_ORIGINSfor your user account, setting the value to*. - Click
OK/Applyto save and restart the system. - Run
Ollamaagain.
Conversing with Local Large Models in LobeChat
Now, you can start conversing with the local LLM in LobeChat.
Using Ollama on Linux
Local Installation of Ollama
Install using the following command:
curl -fsSL https://ollama.com/install.sh | shAlternatively, you can refer to the Linux manual installation guide.
Configure Ollama for Cross-Origin Access
Due to Ollama's default configuration, which allows local access only, additional environment variable setting OLLAMA_ORIGINS is required for cross-origin access and port listening. If Ollama runs as a systemd service, use systemctl to set the environment variable:
- Edit the systemd service by calling
sudo systemctl edit ollama.service:
sudo systemctl edit ollama.service- Add
Environmentunder[Service]for each environment variable:
[Service]
Environment="OLLAMA_HOST=0.0.0.0"
Environment="OLLAMA_ORIGINS=*"- Save and exit.
- Reload
systemdand restart Ollama:
sudo systemctl daemon-reload
sudo systemctl restart ollamaConversing with Local Large Models in LobeChat
Now, you can start conversing with the local LLM in LobeChat.
Deploying Ollama using Docker
Pulling Ollama Image
If you prefer using Docker, Ollama provides an official Docker image that you can pull using the following command:
docker pull ollama/ollamaConfigure Ollama for Cross-Origin Access
Since Ollama's default configuration allows local access only, additional environment variable setting OLLAMA_ORIGINS is needed for cross-origin access and port listening.
If Ollama runs as a Docker container, you can add the environment variable to the docker run command.
docker run -d --gpus=all -v ollama:/root/.ollama -e OLLAMA_ORIGINS="*" -p 11434:11434 --name ollama ollama/ollamaConversing with Local Large Models in LobeChat
Now, you can start conversing with the local LLM in LobeChat.
Installing Ollama Models
Ollama supports various models, which you can view in the Ollama Library and choose the appropriate model based on your needs.
Installation in LobeChat
In LobeChat, we have enabled some common large language models by default, such as llama3, Gemma, Mistral, etc. When you select a model for conversation, we will prompt you to download that model.
Once downloaded, you can start conversing.
Pulling Models to Local with Ollama
Alternatively, you can install models by executing the following command in the terminal, using llama3 as an example:
ollama pull llama3Custom Configuration
You can find Ollama's configuration options in Settings -> Language Models, where you can configure Ollama's proxy, model names, etc.
Visit Integrating with Ollama to learn how to deploy LobeChat to meet integration needs with Ollama.