Google releases Gemma 3, the latest 'open' AI model, boasting single-GPU efficiency and enhanced performance. It surpasses competitors in benchmarks, supports 35+ languages, and handles text, images, and short videos. With a focus on safety and accessibility, Gemma 3 aims to empower developers and researchers, furthering the democratization of powerful AI technology.
Researchers have introduced Tradutor, an open-source AI translator designed for European Portuguese. It addresses the underrepresentation of this language variety in machine translation. Using a new parallel corpus, PTradutor, and fine-tuning techniques on LLMs like Gemma and LLaMA-3, Tradutor achieves performance comparable to commercial models, promoting linguistic inclusivity in AI.
This research explores the effectiveness of scaling inference time in diffusion models. It introduces a framework that uses verifiers and algorithms to optimize sampling noise, leading to higher quality generated samples. The study examines various verifier-algorithm combinations and their compatibility with fine-tuned models, demonstrating that inference-time scaling can be very efficient, even on smaller models.
