Meta Spirit LM Integrates Speech and Text in New Multimodal GenAI Model
Presented in a recent paper, Spirit LM enables the creation of pipelines that mixes spoken and written text to integrate speech and text in the same multimodal model. According to Meta, their novel approach, based on interleaving text and speech tokens, makes it possible to circumvent the inherent limitations of prior solutions that use distinct pipelines for speech and text.
Meta MobileLLM Advances LLM Design for On-Device Use Cases
With MobileLLM, Meta researchers aim to show that, for smaller models, quality is not a direct product of how many billions parameters they have; rather, it is the result of carefully designing their architecture. To prove their point, they coupled deep and thin architectures with embedding sharing and grouped-query attention mechanisms to improve accuracy over prior state-of-the-art models.
Google Debuts OpenAI-compatible API for Gemini
In an effort to make it easier for developers who adopted OpenAI for their LLM-based solutions to switch to Gemini, Google has launched a new endpoint for its Gemini API that allows them to easily switch from one service to the other. The new endpoint is still in beta and provides only partial coverage of OpenAI capabilities.
How Uber Sped Up SQL-based Data Analytics with Presto and Express Queries
Uber uses Presto, an open-source distributed SQL query engine, to provide analytics across several data sources, including Apache Hive, Apache Pinot, MySQL, and Apache Kafka. To improve its performance, Uber engineers explored the advantages of dealing with quick queries, a.k.a. express queries, in a specific way and found they could improve both Presto utilization and response times.
LLaVA-CoT Shows How to Achieve Structured, Autonomous Reasoning in Vision Language Models
Chinese researchers fine-tuned Llama-3.2-11B to improve its ability to solve multimodal reasoning problems by going beyond the direct-response or chain-of-thought (coT) approaches to reason step by step in a structured way. Named LLava-CoT, the new model outperforms its base model and proves better than larger models, including Gemini-1.5-pro, GPT-4o-mini, and Llama-3.2-90B-Vision-Instruct.