The Technology Innovation Institute (TII), a leading global scientific research center, has launched the second iteration of its renowned large language model (LLM) series, Falcon 2. The series includes two groundbreaking versions: Falcon 2 11B, an efficient LLM trained on 5.5 trillion tokens with 11 billion parameters, and Falcon 2 11B VLM, a vision-to-language model (VLM) enabling seamless conversion of visual inputs into textual outputs.
Falcon 2 11B VLM stands out as TII’s first multimodal model, and the only top-tier market model with image-to-text conversion capability, marking a significant advancement in AI innovation.
Surpassing Competitors
Tested against several prominent AI models, Falcon 2 11B surpasses the performance of Meta’s newly launched Llama 3 with 8 billion parameters (8B) and matches Google’s Gemma 7B. These results were independently verified by Hugging Face, a US-based platform hosting an objective evaluation tool and global leaderboard for open LLMs.
Open-Source Accessibility
Both Falcon 2 11B and 11B VLM are open-source, allowing developers worldwide unrestricted access. TII plans to expand the Falcon 2 next-generation models, introducing various sizes and advanced machine learning capabilities like ‘Mixture of Experts’ (MoE) to further enhance performance.
Multilingual and Multimodal Capabilities
Equipped with multilingual capabilities, Falcon 2 11B models excel in English, French, Spanish, German, Portuguese, and more. The vision-to-language model Falcon 2 11B VLM can identify and interpret images, providing applications across healthcare, finance, e-commerce, education, and legal sectors. These models run efficiently on just one GPU, making them highly scalable and easy to integrate into lighter infrastructures.
Leadership Comments
Faisal Al Bannai, Secretary General of ATRC, emphasized the commitment to the open-source movement and the Falcon Foundation. Dr. Hakim Hacid, Executive Director of the AI Cross-Centre Unit at TII, highlighted the benefits of smaller, efficient models and the potential of vision-to-language capabilities for AI accessibility.
Future Innovations
The versatility of Falcon 2 11B is leading TII to explore more GenAI innovations, including the ‘Mixture of Experts’ approach. This involves combining smaller networks with distinct specializations to deliver sophisticated and customized responses, improving accuracy and accelerating decision-making.
