Repeating the input prompt improves performance for popular LLMs (Gemini, GPT, Claude, and Deepseek) without increasing the number of generated tokens or latency, when not using reasoning.
This article explores different chunking strategies for Retrieval-Augmented Generation (RAG) systems, comparing nine approaches using the agenticmemory library to improve retrieval accuracy and reduce hallucinations.
A detailed guide for running the new gpt-oss models locally with the best performance using `llama.cpp`. The guide covers a wide range of hardware configurations and provides CLI argument explanations and benchmarks for Apple Silicon devices.
LocalScore is an open benchmark to evaluate local AI task performance across various hardware configurations, measuring Prompt Processing speed, Token Generation speed, Time-to-First-Token (TTFT), and a combined LocalScore.
The article explores the evolution of large language model (LLM) serving, highlighting significant advancements from pre-2020 frameworks to the introduction of vLLM in 2023. It discusses the challenges of efficient memory management in LLM serving and how vLLM's PagedAttention technique revolutionizes the field by reducing memory wastage and enabling better utilization of GPU resources.
A tool to estimate the memory requirements and performance of Hugging Face models based on quantization levels.
Investigation into the effect of DDR5 speed on local LLM inference speed.
The article discusses the importance of fine-tuning machine learning models for optimal inference performance and explores popular tools like vLLM, TensorRT, ONNX Runtime, TorchServe, and DeepSpeed.
This repository contains scripts for benchmarking the performance of large language models (LLMs) served using vLLM.
A startup called Backprop has demonstrated that a single Nvidia RTX 3090 GPU, released in 2020, can handle serving a modest large language model (LLM) like Llama 3.1 8B to over 100 concurrent users with acceptable throughput. This suggests that expensive enterprise GPUs may not be necessary for scaling LLMs to a few thousand users.
