Is Copilot the best AI companion out there? Help us find out by answering a couple of quick questions!
Fine tuning allows you to customize large language models for specific tasks and it operates by customizing the large language model itself. Microsoft’s Azure OpenAI Service & Azure Machine Learning offer Supervised Fine Tuning, which allows you to provide custom data (prompt/completion or conversational chat) to teach the base model new skills.
Microsoft today announced a significant update for Azure OpenAI Service users. First, fine-tuning is now available for GPT-3.5-Turbo, Babbage-002, and Davinci-002 in Azure Machine Learning model catalog. Microsoft also announced the availability of two new base inference models (Babbage-002 and Davinci-002) in Azure Machine Learning model catalog.
- GPT-3.5-Turbo is the most capable and cost-effective OpenAI model available on Azure and it is optimized primarily for chat apps.
- Babbage-002 and Davinci-002 are much smaller in size and they can understand and generate natural language and code. But they don’t have instruction-following capabilities.
It is important to note that all three models depend on data before Sep 2021. Find the pricing details below:
|Model||Hourly Hosting||Input tokens||Output Tokens|
|Babbage-002||$1.70||$0.0004 / 1k||$0.0004 / 1k|
|Davinci-002||$3.00||$0.0020 / 1k||$0.0020 / 1k|
|GPT-35-Turbo||$7.00||$0.0015 / 1k||$0.0020 / 1k|
You can learn more about Fine Tuning models here.