Stability AI unveiled the StableCode artificial intelligence model for code generation

StableCode supports languages such as Python, JavaScript, and C and comes in three different options.

Stability AI is probably best known for its Stable Diffusion text-to-image generation model, but the startup has other AI models at its disposal. Now Stability AI has launched a large new language model called StableCode that can be used to generate code.

Stability AI Releases the First Public Version of StableCode; A new large language model (LLM) designed to help users generate code. StableCode comes in three different levels: a basic model for general use cases, an instruction model, and a long-text model that can support up to 16,000 tokens.

What languages does StableCode AI support?

The StableCode model leverages the data set of the BigCode open-source programming language project with additional filters and fine-tuning from Stability AI. Initially, this model supports Python, Go, Java, JavaScript, C, and C++ programming languages.

Christine Lafort, head of research at Stability AI, told VentureBeat:

“What we’re trying to do with this model is similar to what we’ve done with Stable Diffusion to help everyone in the world become an artist. “We want to do the same thing with the StableCode model so that anyone who has a good idea and might have a problem implementing it can write a program that just fixes that problem.”

Nathan Cooper, Senior Research Scientist at Stability AI, also talked about using BigCode data to train the StableCode model in an interview and said:

“We love BigCode; They do amazing work on data governance, model governance, and model training. “We took that dataset and applied additional filters for quality as well as building a long background version and then trained our model on that.”

Cooper also explained that with StableCode’s long context model, which supports up to 16,000 tokens, more specialized and complex code can be generated.