The article walks through upgrading vLLM to 0.17.0, configuring Docker containers for 4090 GPUs, comparing FP8 and 4‑bit quantization of Qwen3.5 35B and 27B models, and presents detailed performance numbers and script parameters that reveal trade‑offs in memory usage and throughput.
This tutorial explains the concepts, methods, and practical commands for fine‑tuning large language models using efficient techniques like LoRA and QLoRA, covering model selection, resource considerations, Docker deployment, dataset preparation, training configuration, evaluation metrics, model merging, and deployment with GGUF and Ollama.
This article explains how a hierarchical sparse‑attention framework redesigns KVCache storage across GPU, CPU, and remote memory, eliminates bandwidth and capacity bottlenecks, and enables efficient inference for 128K‑token and larger contexts with dramatically reduced GPU memory usage and higher throughput.
This article analyzes Deepseek's EP‑based inference architecture for V3/R1 models, comparing it with TP, detailing how EP reduces memory and compute overhead, boosts batch size, cuts GPU memory usage, and introduces reliability, scalability, and maintainability challenges for large‑scale deployments.
This article presents a comprehensive overview of state‑of‑the‑art distributed training methods for large language models, using GPT‑175B as a case study to analyze memory, communication, and compute overheads, and to recommend practical optimization strategies such as tensor, pipeline, and sequence parallelism, ZeRO‑1 optimizer, and selective activation checkpointing.
This article explains how to overcome GPU memory limits by using PyTorch 1.9's meta device to create an empty model, load large‑scale model weights layer‑by‑layer, move each part to a 16 GB GPU for inference, and release memory, enabling a 70B FP16 model to run on a single consumer‑grade GPU.
This article explains how researchers applied quantization techniques to high‑resolution physical simulations, enabling over 50% memory reduction without noticeable visual loss, by modeling error propagation, using constrained optimization, and introducing dithering, with results demonstrated on GPU‑based smoke, fluid, and elastic body simulations.
This article summarizes an NIPS 2017 paper that introduces GPU memory‑optimization techniques—swap‑out/in and a memory‑efficient attention layer—integrated into TensorFlow, enabling significantly larger batch sizes and deeper models without sacrificing accuracy.
