Skip to content Skip to sidebar Skip to footer

Small Language Model

BRAG Unveils: Small Language Models (SLMs) Optimized for RAG Tasks Available for Under $25 each.

The BRAG series is a set of high-performance Retrieval Augmented Generation (RAG) models developed by Maximalists AI Researcher. They are a small language model designed to be a low-cost alternative for AI-driven language processing, proving effective in artificial intelligence due to their affordability and cost-effectiveness. They were created to meet the need for more powerful…

Read More

Improving Text Embeddings in Compact Language Models: A Comparative Refinement Method using MiniCPM.

Researchers from Tsinghua University have developed an approach to improve the performance of smaller language models such as MiniCPM, Phi-2, and Gemma by enhancing their text embeddings. By applying contrastive fine-tuning using the NLI dataset, the researchers significantly improved the text embedding quality across various benchmarks. In particular, MiniCPM showed a significant 56.33% performance improvement,…

Read More

The Gemma 2-2B model has been launched, featuring an advanced text generation capability with 2.6 billion parameters, enhanced security measures, and the ability to deploy on the device itself.

Google's AI research team, DeepMind, has unveiled Gemma 2 2B, its new, sophisticated language model. This version, supporting 2.6 billion parameters, is optimized for on-device use and is a top choice for applications demanding high performance and efficiency. It holds enhancements for handling massive text generation tasks with more precision and higher levels of efficiency…

Read More

OuteAI Introduces Innovative Lite-Oute-1 Variants: Lite-Oute-1-300M and Lite-Oute-1-65M as Robust Yet Space-Saving AI Platforms.

OuteAI has released two new models of its Lite series, namely Lite-Oute-1-300M and Lite-Oute-1-65M, which are designed to maintain optimum efficiency and performance, making them suitable for deployment across various devices. The Lite-Oute-1-300M model is based on the Mistral architecture and features 300 million parameters, while the Lite-Oute-1-65M, based on the LLaMA architecture, hosts around…

Read More

Neural Magic has launched a fully quantized FP8 iteration of Meta’s Llama 3.1 405B Model, including FP8 Dynamic and Static Quantization.

Neural Magic, an AI solutions provider, has recently announced a breakthrough in AI model compression with the introduction of a fully quantized FP8 version of Meta's Llama 3.1 405B model. This achievement is significant in the field of AI as it allows this massive model to fit on any 8xH100 or 8xA100 system without the…

Read More

Arcee AI has unveiled Arcee-Nova, a new open-source language model. This revolutionary model, based on Qwen2-72B, nears the performance level of GPT-4.

Arcee AI, known for its innovation in open-source artificial intelligence, has launched Arcee-Nova, which is hailed as a pioneering accomplishment in the AI sector. Arcee-Nova has quickly gained recognition as the highest-performing model within the open-source arena, nearly on par with the performance of GPT-4, a benchmark AI model as of May 2023. Arcee-Nova is an…

Read More