With all the excitement around large language models, one question has been repeatedly asked throughout – “How do I make my own?”. There are plenty of techniques to pass additional context and reference material into a generic, pre-trained model, but if you want a model that is super fast and already understands your context, you need to fine-tune a model.
We recently saw announcements from Microsoft that the Azure Open AI service now supports both training a fine-tuned model, but also hosting that model at the click of a button!
In this video, Simon and Gavi take a look at the concept of Fine Tuning LLMs, then review a super quick demo of how we pass some data to the gpt3.5 turbo model, which is what powers the public ChatGPT service!
For more details around the recent announcements, check out: https://techcommunity.microsoft.com/t5/azure-ai-services-blog/fine-tuning-now-available-with-azure-openai-service/ba-p/3954693
And as always, if you need an accelerator behind your large language model aspirations, get in touch with Advancing Analytics!
