#BF16
224 related articles
Product Reviews·1 min
Qwen3.6 Quantized Local Deployment Benchmark: Which Is Best — NVFP4, APEX, Q4, or Q6?
Benchmarking 7-8 Qwen3.6 quantized models across 8 dimensions including tool calling, CLI ops, and bug fixing. Comparing NVFP4, APEX, Q4, Q6 with rankings and recommendations.
Read more →
Tutorials·2 min
vLLM and SGLang Local Deployment Tutorial: A Practical Guide to 3-8x Performance Gains
Complete guide to deploying vLLM and SGLang locally. Compare performance vs LM Studio, deploy in 3 steps with Docker + AI assistant. Covers SGLang vs vLLM selection, 5090 VRAM optimization, and Cherry Studio integration.
Read more →
Tutorials·3 min
Taobao Sign Signature Reverse Engineering: AI-Assisted MD5 Encryption Cracking Full Workflow
Complete walkthrough of reverse engineering Taobao's sign signature MD5 encryption, covering packet capture, breakpoint debugging, algorithm verification, and Python code generation with AI assistance.
Read more →
Tutorials·2 min
Reverse Engineering Tencent Slider CAPTCHA: Full Walkthrough of the CO Value Encryption Algorithm
Full walkthrough of reverse engineering Tencent Slider CAPTCHA: CO value encryption tracing, tdcgs analysis, AI-assisted browser env simulation, gap detection, POW, and Python automation.
Read more →
Tutorials·3 min
Cursor/Windsurf Rules Configuration Guide: Generate Consistent UI Components with AI
Learn how to write Rules files in Cursor and Windsurf to generate consistent UI components with AI. Includes Apple Liquid Glass style case study, writing tips, and best practices for Angular, React, and Vue.
Read more →
Deep Dives·4 min
Will Frequent Git Commits Fill Up Your Hard Drive? Delta Compression & Packfile Mechanism Explained
Will frequent Git commits fill your hard drive? Deep dive into Git's Delta compression and Packfile mechanism, revealing why pure code commits cost almost nothing and identifying the real culprit — large binary files.
Read more →
Tech Frontiers·4 min
Step 3.5 Flash Deep Dive: How a 196B Parameter Open-Source Model Beat Gemini
Deep dive into StepFun's Step 3.5 Flash: 196B parameter MoE model activating only 11B, 350 tokens/sec coding speed, 256K context window, local deployment ready. How it beats Gemini 3 Flash.
Read more →
Tutorials·2 min
NCCL Multi-GPU Communication Optimization in Practice: A Guide to Scaling HPC High-Performance Computing
Deep dive into NVIDIA NCCL multi-GPU communication library principles and optimization strategies, covering AllReduce, NVLink, and GPUDirect RDMA to help HPC and AI developers master scaling from single-node to massive clusters.
Read more →
Tutorials·4 min
NVIDIA Model Optimizer Post-Training Quantization (PTQ) Practical Guide
A deep dive into NVIDIA Model Optimizer's PTQ workflow, covering INT8/INT4 quantization principles, calibration methods, RTX GPU optimization, and best practices for deploying quantized LLMs on consumer GPUs.
Read more →
Product Reviews·3 min
8K Stars on GitHub: Deep Dive into the Most Comprehensive LLM Resource Repository
Deep analysis of the awesome-LLM-resources project (8200+ GitHub Stars), covering multimodal AI, Agents, MCP protocol, model training, o1 reasoning, SLMs, and more for LLM practitioners.
Read more →
Tutorials·2 min
Simon Willison's Full Process of Fixing a Datasette Map Bug with GPT-5.5 + Codex
Datasette creator Simon Willison debugged an OpenStreetMap display issue, uncovering CAPTCHA and Referrer-Policy conflicts, then built a plugin fix using GPT-5.5 and Codex.
Read more →
Product Reviews·4 min
Hugging Face Transformers: Deep Dive into the 160K-Star Open Source AI Model Framework
Deep dive into Hugging Face Transformers: architecture, multimodal support, ecosystem, and trends. Learn how this 160K-Star project became essential for AI developers.
Read more →
Deep Dives·3 min
Google TPU v8t and v8i Deep Dive: The New Era of Separated Training and Inference AI Chips
Google Cloud Next unveils TPU v8t (training) and TPU v8i (inference) chips. Deep analysis of their architecture, strategic significance, and impact on AI chip competition.
Read more →
Product Reviews·3 min
GPT-OSS 120B Local Deployment Review: Comprehensive Comparison with O4 Mini on Code Generation and Reasoning
Hands-on review of OpenAI's first open-weight reasoning model GPT-OSS 120B, locally deployed via Ollama, benchmarked against O4 Mini on code generation, reasoning, and UI design.
Read more →
Product Reviews·3 min
Multi-Agent Orchestration Framework: A Super AI Development Team of 98 Expert Agents
Deep dive into a trending GitHub multi-agent framework with 98 expert agents, swarm collaboration, HNSW vector memory, autonomous learning, and Claude Code plugin support for full-lifecycle development.
Read more →
Deep Dives·3 min
98 Expert Agents Working in Concert: A Deep Dive into an Open-Source Multi-Agent Orchestration Framework
Deep dive into a trending open-source multi-agent framework with 98 expert agents, swarm orchestration, HNSW vector memory, autonomous learning, and Agent Federation for distributed collaboration.
Read more →
Deep Dives·3 min
Microsoft Tutel: A Deep Dive into the MoE Model Acceleration Library with FP4/FP8 Low-Precision Support
A deep dive into Microsoft's open-source Tutel MoE optimization library, supporting FP8, NVFP4, and MXFP4 multi-precision computation for DeepSeek, Kimi-K2, Qwen3, and other leading MoE models.
Read more →
Deep Dives·4 min
Hugging Face Transformers: Deep Dive into the 160K-Star Open Source AI Framework
Deep dive into Hugging Face Transformers: the 160K-Star open-source framework covering Pipeline API, Auto Classes, multi-modal models, and the full HF ecosystem for AI inference and training.
Read more →
Product Reviews·2 min
Unsloth: 80% VRAM Reduction — The Best Open-Source Tool for Local LLM Fine-Tuning
Unsloth is an open-source LLM training tool with 63K+ GitHub stars, supporting Gemma 4, Qwen 3, DeepSeek. Reduces VRAM by 50–80%, enabling RTX 4090 to fine-tune 7B models with a no-code Web UI.
Read more →
Tutorials·3 min
Unsloth Tutorial: Fine-Tune LLMs Locally with Multi-fold Speed Boost and 60% VRAM Savings
Unsloth is an open-source LLM fine-tuning tool with 63K GitHub stars, supporting Gemma 4, Qwen3, and DeepSeek. It achieves multi-fold training speedup and 60% VRAM reduction through kernel optimization, enabling fine-tuning on consumer GPUs.
Read more →