#KV Cache optimization
398 related articles
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 →
Tutorials·3 min
GPT-OSS Open-Source Model Local Deployment Tutorial: Real-World Performance and VRAM Usage Analysis
Complete guide to deploying OpenAI's open-source GPT-OSS model locally with Ollama. Real-world testing of the 20B version on RTX 4090 covering Chinese comprehension, logical reasoning, and VRAM usage analysis under MoE architecture.
Read more →
Product Reviews·3 min
Claude Haiku 4.5 Real-World Test: Fast but Outclassed in Value by GPT-5 Mini
Real-world comparison of Claude Haiku 4.5 vs GPT-5 Mini and GLM 4.6 on speed, code quality, and price. Haiku 4.5 beats Sonnet 4 by one minute but costs 4x more than GPT-5 Mini with 9 points lower coding scores.
Read more →
Product Reviews·3 min
Step Audio 2.5 In-Depth Review: How Does China's Voice AI Compare to GPT Realtime 2?
Hands-on comparison of StepFun's Step Audio 2.5 vs OpenAI GPT Realtime 2 across reasoning, role-playing, Chinese understanding, and API pricing for developers.
Read more →
Expert Opinions·3 min
Sam Altman in Conversation with Garry Tan: How OpenAI and YC Are Co-Shaping the AI Startup Ecosystem
Sam Altman and YC President Garry Tan discuss the convergence of OpenAI's foundation models and YC's startup ecosystem, revealing key signals about the next phase of AI entrepreneurship.
Read more →
Tutorials·3 min
4-bit QLoRA Fine-tuning LLaMA 3 in Practice: A Guide to Training 8 Billion Parameter LLMs on Consumer GPUs
Learn how to fine-tune LLaMA 3 8B on consumer GPUs using QLoRA and Unsloth. Covers 4-bit quantization, LoRA adaptation, Alpaca data format, training pipeline setup, and model deployment—runs on RTX 3090.
Read more →
Product Reviews·3 min
awesome-LLM-resources: A Comprehensive Analysis of the 8,200-Star LLM Learning Resource Repository on GitHub
In-depth analysis of the 8,200-star GitHub project awesome-LLM-resources, covering multimodal generation, Agents, model training, MCP protocol, and more — a one-stop LLM learning guide.
Read more →
Tech Frontiers·2 min
Cursor Launches Claude Opus 4 Fast Mode: 2.5x Faster, 6x More Expensive — Is It Worth It?
Cursor launches Claude Opus 4 Fast Mode with 2.5x speed but 6x cost. We analyze use cases, cost tradeoffs, and practical tips to help you decide if it's worth it.
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·3 min
Cube Studio: A Comprehensive Analysis of Tencent's Open-Source Cloud-Native AI Platform
Deep dive into Tencent Music's open-source Cube Studio cloud-native AI platform, covering distributed training, LLM fine-tuning & inference, Pipeline orchestration, and domestic hardware adaptation.
Read more →
Tutorials·3 min
GitHub 8K Stars: Deep Dive into the Most Comprehensive LLM Resource Repository
Deep analysis of the GitHub project awesome-LLM-resources with 8,200+ Stars, covering multimodal AI, Agents, MCP protocol, model training, inference optimization, and coding assistants.
Read more →
Tech Frontiers·2 min
Kimi K2 Official Verification Tool K2-Vendor-Verifier: One-Click API Vendor Model Accuracy Detection
Moonshot AI open-sources K2-Vendor-Verifier to verify third-party Kimi K2 API vendor inference accuracy. Learn how this tool helps developers detect over-quantization, model substitution, and other API market risks.
Read more →
Tech Frontiers·2 min
OpenAI Symphony: One Codex Agent Per Task — A New Paradigm for Multi-Agent Parallel Programming
Deep dive into OpenAI's Symphony: assigning an independent Codex agent to each task for multi-agent parallel programming. Explore its architecture, developer impact, and industry trends.
Read more →
Tutorials·3 min
AI Project Evaluation Framework: A Practical Guide to 5 Levels and 30 Key Metrics
A battle-tested AI project evaluation framework covering 5 levels and 30 core metrics—model quality, UX, system efficiency, business value, and data loops—to scientifically assess LLM Agent performance.
Read more →
Deep Dives·3 min
Augment Code Architecture Revealed: Dedicated Subagents Replace KV Cache, Cutting Costs by 90%
Deep dive into how Augment Code uses Mercury 2 dedicated subagents to replace traditional KV cache, achieving 82% faster context compaction, 90% lower summarization costs, and 30% reduced LLM spending.
Read more →
Tech Frontiers·2 min
DeepSeek V3.2 Released: Proprietary Sparse Attention (DSA) + 50% API Price Cut Explained
DeepSeek releases V3.2-Exp with proprietary DeepSeek Sparse Attention (DSA) for faster long-context training and inference, plus API prices cut over 50%.
Read more →
Product Reviews·2 min
DeepSeek-Reasonix: A Terminal AI Coding Agent Optimized with Prefix Cache Stability
DeepSeek-Reasonix is an open-source terminal AI coding agent natively designed for DeepSeek models, achieving lower latency and API costs through prefix cache stability optimization.
Read more →
Product Reviews·3 min
Cube Studio: Tencent's Open-Source One-Stop AI Platform — A Deep Dive into Full MLOps Lifecycle
Deep dive into Tencent Music's open-source Cube Studio cloud-native AI platform, covering Notebook development, Pipeline orchestration, distributed training, LLM fine-tuning, inference deployment, and domestic hardware adaptation for full MLOps lifecycle.
Read more →
Product Reviews·1 min
Hugging Face Transformers: A Complete Guide to the AI Model Framework with 160K Stars
Explore Hugging Face Transformers, the 160K-star open-source framework for AI models covering text, vision, audio, and multimodal with unified APIs.
Read more →
Product Reviews·4 min
Cube Studio: Tencent's Open-Source All-in-One AI Platform Covering the Full Pipeline from Large Model Training to Inference
Deep dive into Tencent Music's open-source Cube Studio cloud-native AI platform, covering distributed training, LLM fine-tuning, vLLM inference, VGPU virtualization, and Huawei Ascend adaptation.
Read more →