GPU & LLM Infrastructure

8 articles

Implementation Record: Balancing Accuracy and Determinism with Qwen2.5-14B + GPUStack

Explains why BASTION chose a local LLM (Qwen2.5-14B on GPUStack) over cloud LLM APIs. The LLM only handles tasks like log summaries and reports, while severity and blocking decisions are left to deterministic logic. Also covers enforced JSON output and low temperature.

Steps to Connect Bestnet Cloud and GPU-VPS via 10G Private Network and Implement a GPUStack Cluster

How-toBESTNET-CLOUDGPU-VPSGPUStack10G Private NetworkNFS Shared CacheTech Blog / Infrastructure Build GuideBes

Migration Story: GPUStack from v0.7.1 to v2.1.1

Tech BlogGPUStackDocker MigrationNFS Shared CacheGPU Cluster Operations NotesMigrating GPUStack from v0.7.1 to

Installing NVIDIA Driver and CUDA 12.9 for Tesla V100 on Ubuntu 24.04 VM in Proxmox VE

Proxmox VE / Ubuntu 24.04 / NVIDIA CUDAInstalling NVIDIA Driver and CUDA 12.9 on Tesla V100 in Ubuntu 24.04 VM

Resolving the issue where pipeline_parallel=2 crashes on V100

Tech BlogGPUStackvLLMTesla V100Multi-Node GPU Cluster Operations Memo Solving the pipeline_parallel=2 c

GPUStack v2.1.1: Resolving All Multi-Node Inference Startup Issues

Tech BlogGPUStackMulti-Node vLLMTesla V100TroubleshootingGPU Cluster Operations MemoResolved All Issues with G

The approach and pitfalls when upgrading Dify from v1.9.2 to v1.13.3

Infrastructure / Self-hosted AI Platform How to Upgrade Dify from v1.9.2 to v1.13.3 and Pitfalls to Avoid Upgr

Installing NVIDIA Driver and CUDA on Rocky Linux 9.3

Installation ProcedureSTEP1. Add Official NVIDIA RepositoryRegister the official NVIDIA CUDA repository for Ro