Google’s Gemma suite of artificial intelligence models has now been downloaded more than 150 million times. This announcement, shared by a Google DeepMind engineer on social media, also highlighted that developers have fashioned over 70,000 unique variants of Gemma through the Hugging Face AI development platform.
Debuting in February 2024, Gemma was introduced to compete against other widely known families of open AI models. The most up to date Gemma versions can process both text and images, and are capable of supporting communication in over one hundred languages.
Gemma’s Growth and Competition
Specific adaptations of Gemma have been tailored to fields such as pharmaceutical research, leveraging fine tuning for specialized applications. However, despite the impressive download count, Gemma remains behind Meta’s Llama model family, which achieved over 1.2 billion downloads as of late April.
Both the Gemma and Llama models have sparked debate within the development community surrounding their licensing terms. Some developers argue that nonstandard licensing associated with these models creates legal uncertainties for commercial use.
Gemma’s expansion signals significant interest in flexible AI tools, yet its progress is accompanied by clear challenges. The ongoing competition and scrutiny over licensing will likely influence its trajectory in the rapidly evolving landscape of artificial intelligence.
