An anonymous reader cites a TechCrunch report. Recognizing the demand for alternatives, several years ago AI startup Hugging Face partnered with workflow automation platform ServiceNow to create StarCoder, an open source code generator with a less restrictive license than others. Did. The original was released online early last year, and a sequel, StarCoder 2, has been in development ever since. StarCoder 2 is not a single code generation model, but a family. This version released today has three variations, the first two of which can run on modern consumer GPUs. A 3 billion parameter (3B) model trained by ServiceNow. A 7 billion parameter (7B) model trained by Hugging Face. The newest supporter of the StarCoder project, he is a 15 billion parameter (15B) model trained by Nvidia. (Note that “parameters” are the parts of the model learned from the training data that essentially define the model's skill for the problem, and in this case, the code it generates.)
Like most other code generators, StarCoder 2 can suggest ways to complete unfinished lines of code or retrieve summarized snippets of code in response to questions in natural language. Trained on four times the data of the original StarCoder (67.5 terabytes vs. 6.4 terabytes), StarCoder 2 offers what Hugging Face, ServiceNow, and Nvidia characterize as “significantly” improved performance at lower operating costs. Masu. StarCoder 2 can use GPUs like the Nvidia A100 to fine-tune first-party or third-party data in “hours” to create apps like chatbots and personal coding assistants. StarCoder 2 was also trained on a larger and more diverse dataset than the original StarCoder (approximately 619 programming languages), so it can, at least hypothetically, make more accurate and context-aware predictions.
[I]■ StarCoder 2 is truly superior to other code generators. Is it free or paid? Depending on the benchmarks, it appears to be more efficient than one of his versions of Code Llama, Code Llama 33B. According to Hugging Face, StarCoder 2 15B matches Code Llama 33B twice as fast on a subset of code completion tasks. It's not clear which task. Hug face is not specified. As an open source collection of models, StarCoder 2 also has the advantage of being able to be deployed locally to “learn” the developer's source code or codebase, allowing developers who are wary of exposing their code to cloud-hosted AI. This is an attractive prospect for individuals and businesses. . Hugging Face, ServiceNow, and Nvidia also claim that StarCoder 2 is more ethical and has fewer legal issues than its competitors. […] In contrast to code generators trained using copyrighted code (particularly GitHub Copilot), StarCoder 2 is based on a license from Software Heritage, a nonprofit organization that provides code archiving services. Trained on data only. Prior to StarCoder 2 training, BigCode, the cross-organizational team behind much of the StarCoder 2 roadmap, gave code owners the opportunity to opt out of the training set if they wished. Like the original StarCoder, StarCoder 2's training data is available for developers to fork, clone, and audit as needed. The StarCoder 2 license may still be an obstacle for some people. “StarCoder 2 is licensed under BigCode Open RAIL-M 1.0, which is intended to encourage responsible use by imposing 'light-touch' restrictions on both model licensees and downstream users,” TechCrunch said. writes Kyle Wiggers. “While RAIL-M is less restrictive than many other licenses, it is not truly 'open' in the sense that it does not allow developers to use StarCoder 2 in every possible application. (For example, apps that provide medical advice are strictly off-limits.) Some commentators argue that RAIL-M's requirements may be too vague to comply with in any case, and that RAIL-M does not comply with EU AI law. It states that there is a possibility of violating AI-related regulations such as . ”