OpenAI, the company behind the powerful and popular ChatGPT-3 language model, has announced the release of fine-tuning capabilities for its GPT-3.5 Turbo model. The new capability will enable developers to customise the existing models as per their use case and run them at scale.
OpenAI claims that the fine-tuned version of GPT-3.5 Turbo can even beat GPT-4’s capabilities. In early tests, a fine-tuned version of the model was able to “match, or even outperform, base GPT-4-level capabilities on certain narrow tasks.”
Fine-tuning is a process of training a language model on a specific dataset to improve its performance on any specified task. OpenAI’s latest move will now let businesses train GPT 3.5 Turbo using their own data set. With this, the developers can now customise GPT-3.5 Turbo for their specific needs.
ALSO READ l Alert! Another software professional loses over Rs 10 lakhs after clicking on an Instagram ad
For example, a company could fine-tune the model to generate more creative text, or to answer questions more accurately. Fine-tuning can also be used to improve the model’s performance on specific tasks, such as translation or code generation.
“Since the release of GPT-3.5 Turbo, developers and businesses have asked for the ability to customize the model to create unique and differentiated experiences for their users. With this launch, developers can now run supervised fine-tuning to make this model perform better for their use cases,” company said in the blog post.
Fine-tuning, apart from increasing performance, also helps businesses shorten their prompts for same performance. Fine-tuning with GPT-3.5-Turbo can work with longer texts, up to 4,000 words which is twice as much as before. Early testers have found that they could make their prompts to the model much shorter, even by up to 90%, by teaching the model how to follow the instructions. This made the model faster to use and also saved money. The support for fine-tuning with function calling and gpt-3.5-turbo-16k will be coming later this fall, says the company.
Follow FE Tech Bytes on Twitter, Instagram, LinkedIn, Facebook.
