Databricks releases Dolly 2.0, an open-source AI like ChatGPT for commercial use

Databricks has released Dolly 2.0, an open-source large language model designed to democratize access to generative AI. Unlike many proprietary systems locked behind paywalls or restricted by restrictive licensing, Dolly 2.0 is fully open-sourced for commercial use. This approach allows businesses to run the model on internal servers, keeping sensitive data private and enabling organizations to customize and monetize the technology without relying on third-party vendors. The development of Dolly 2.0 specifically addresses the legal barriers faced by earlier open models like Alpaca, which were trained on data sourced from closed APIs and thus forbidden from commercial application. To ensure full freedom of use, Databricks trained this iteration using a high-quality dataset crowdsourced from its own employees. This unique training method removes legal ambiguities, offering enterprises a compliant alternative that supports tuning with proprietary datasets and independent development. This release is highly relevant to the open-data movement because it demonstrates a viable path toward sustainable, commercially friendly open-source AI. By providing both the model weights and the underlying dataset under permissive licenses, Databricks encourages transparency and collaboration. It empowers developers and researchers to build upon existing work without fear of intellectual property infringement, fostering an ecosystem where innovation is driven by shared knowledge rather than closed silos.

Source: siliconangle.com
Published on 2023-04-13

Crece producción industrial de México 3.5 % en febrero »

Open Data

Related news