American enterprise software company Databricks has released Dolly 2.0, the next version of its large language model (LLM) with similar features to ChatGPT. This is the first LLM with open source code and a set of freely accessible training instructions that will help companies to use AI technology for their own commercial projects without having to pay for APIs or share data with third parties.
Course
Cinema 4D course
You will have free access to C4D. We took care of everything
Sign up
In recent months, a number of language models similar to OpenAI’s GPT have been released, which by many definitions could be considered open. One such is Meta’s LLaMA, which in turn was inspired by Alpaca, Koala, Vicuna and Dolly 1.0
However, many of these “open” models were under the control of system developers – for example, the Alpaca team’s AI project at Stanford, which was trained on GPT-3.5 instructions and built on LLaMA 7B. OpenAI’s terms of use include a rule that researchers cannot use products from systems that compete with the company.
Databricks aims to solve this problem. Dolly 2.0 is a large language model with 12 billion parameters, based on the open source Eleuther family of artificial intelligence models and tuned exclusively to a small block of instructions (databricks-dolly-15k) created by the Databricks team. The license terms of this dataset allow you to use, modify and extend it for any purpose, including academic or commercial applications.
The Databricks blog points out that, like the original Dolly, version 2.0 isn’t state-of-the-art, but “shows a surprisingly efficient level of instruction execution given the size of the training block.” The report adds that the level of effort and expense required to create powerful artificial intelligence technologies is “significantly less than previously thought”.
The Dolly 2.0 model can be downloaded from the Databricks Hugging Face page, and the instructions are available from GitHub. The company also offers to attend its webinar on April 25, which will explain how organizations can use the LLM.