Ollama is a versatile platform designed for running and fine-tuning machine learning models, including advanced language models like Llama3. The Ubuntu 24.04 with Ollama VPS template on Hostinger comes pre-installed with Ollama, the Llama3 model, and Open WebUI, providing an efficient way to manage and run these models. This guide will walk you through accessing and setting up Ollama.
If you don't have a VPS yet, check the available options here: LLM VPS hosting 🚀
Accessing Ollama
Open your web browser and navigate to:
https://[your-vps-ip]:8080
Replacing [your-vps-ip] with the IP address of your VPS.
The first you access, you'll be prompted to create an account:
These will be your login credentials for subsequent access.
Getting to Know Open WebUI
Once logged in, explore the Open WebUI dashboard, where you can:
Monitor active models
Upload new datasets and models
Track model training and inference tasks
The pre-installed Llama3 model can be fine-tuned and managed through the interface. You can also add more models in the settings by clicking on the gear icon and selecting Models:
Running Inference with Llama3
You can use the Llama3 pre-trained model for inference directly through the Open WebUI interface, input custom prompts or datasets to see how the model responds.
Adjust the hyperparameters and experiment with fine-tuning on custom data to improve performance:
It's important to note that the response speed of models highly depends on the specific language model used and the number of CPU cores available. Larger models with more parameters often require more computational resources. If you're interested in expanding your machine-learning project, you can consider upgrading your VPS and get more CPU core counts for optimal performance.
Additional Resources
For more detailed information and advanced configurations, refer to the official Ollama documentation