Mozilla.ai, an organization backed by the Mozilla Basis, has launched any-llm v1.0, an open-source Python library that gives a single interface to speak with completely different massive language mannequin (LLM) suppliers.
any-llm 1.0 was launched November 4 and is obtainable on GitHub. With any-llm, builders can use any mannequin, cloud or native, with out rewriting a stack each time. This implies much less boilerplate code, fewer integration complications, and extra flexibility to choose what works finest for the developer, Nathan Brake, machine studying engineer at Mozilla.ai, wrote in a weblog publish. “We needed to make it simple for builders to make use of any massive language mannequin with out being locked right into a single supplier,” Brake wrote.
Mozilla.ai initially launched any-llm on July 24. The 1.0 launch has a steady, constant API floor, async-first APIs, and re-usable consumer connections for high-throughput and streaming use instances, Brake wrote. Clear deprecation and experimental notices are offered to keep away from surprises when API adjustments might happen.