Havuz

Instinct: single-forward-pass decision models (choice / true-false / score) returning a probability per candidate

I Ran Qwen 3.8 27B on Just 4GB VRAM! (FreeToken Tutorial) - Edge-native MoE serving engine for local LLMs on low VRAM GPUs

中文公司调研 Agent

Telemetria cifrada con TypeScript, agente Go, analisis Python y Ollama, Cloudflare Tunnel y lectura mTLS.

Agentic RAG consumer decision agent with ReAct, Text2SQL, multi-hop retrieval, memory and deterministic constraint checking.

Full-pipeline ComfyUI node pack tailored for Krea 2 Turbo: NAG negative guidance, Variance diversity, custom schedulers, samplers & aestheti

Private, offline AI assistant for Android and Windows. Run local LLMs (Gemma 4, Qwen, GGUF) on your phone or PC with llama.cpp, CUDA and Vul

Open-source AI dating agent — bring your own Tinder token and any LLM (OpenAI, Claude, Gemini, DeepSeek, Qwen, and more). Screens dates, mak

Use third-party and Chinese model APIs as bounded Codex subagents while keeping Codex as the main agent.

An agent that builds agents — closed-loop AI system that interviews, designs, tests, and ships standalone Python agents.

Run stock llama.cpp with the stock Unsloth Qwen3.6-35B-A3B 4-bit on a 16 GB Mac.

The honest, local-first usage meter for AI coding tools. With full privacy and with no server or reading your chats.

Qwen 3.8 27B on a $1,400 Strix Halo: frozen optimal config, 262k needle-exact retrieval, vision fixed, delegation-card certified

Typed decisions (choice / score / yes-no) from local Qwen models on Apple Silicon. Probabilities come straight from the logits, no text gene

One command spins up Qwen3.8-27B (abliterated, NVFP4) on a rented Blackwell GPU at ~150 tok/s and wires it into your local coding agent via

Fine-tune Qwen2.5-Coder-7B for free (QLoRA on Kaggle T4) and run a local coding agent that runs your code instead of guessing — 8 GB Mac, ll

Open-source CLI bridge for shell-capable AI agents to work in existing Chrome tabs. Local Node server, no third-party runtime dependencies,

迹录 Tracord —— 基于本地小模型多智能体的智能旅行规划助手 / AI travel planner with local SLM multi-agents — light up visited cities on a map, log expenses with one

Docker compose configs for serving Qwen3.8-27B on a DGX Spark (GB10): vLLM+MTP current, SGLang+DSPARK + FP8 rollback stacks

Offline LLM app for iPhone: run open models from Hugging Face on-device with llama.cpp, MLX, Core ML and Apple Foundation Models. No account

Run Qwen3.8-Flash-Next on ONE RTX 3090 (24 GB) + 64 GB RAM: 128K context, up to 2,100 tok/s prefill, 43–51 tok/s decode. vLLM runtime with h

Tiny engine, immense models — run large MoE LLMs (gpt-oss, Mixtral, Qwen3-MoE) on ordinary machines by streaming experts from disk. OpenAI-c

《大学生AI全场景使用指南》· 13 个校园场景的千问实操手册：逐步截图、指令原文与可复用提示词模板

Advanced local AI prompt generator for image & video models — reference images, dataset creation, prompt refinement, model-specific outputs

Cross-model transfer of persistent recurrent inference state without prefix replay

A 4B decision model built for Korean.

Qwen3.6 vLLM Toolkit — Launcher + Templates, Optimized for 2×24 GB. vLLM launcher with an optimized hybrid chat template drawing from the be

Governed neural weight editing and coding model adaptation. Native Rust patches, exact rollback, LoRA, Claude Code routing, and reproducible

High-performance CUDA inference engine for Qwen models, written from scratch in C++/CUDA

Raw benchmark logs for Swift-Qwen3.8-27B vs the Qwen3.8-27B BF16 base: every response, score, and config across 9 benchmarks

High-performance, high-precision, locally-fast-deployed Jev-class medical decision service on national medical exams — four readings, one ro

模型路由器：在 Codex Desktop 和 Claude Desktop 使用 Claude、第三方 API 与自定义模型。Local LLM proxy for Codex & Claude Desktop: OpenAI Responses / Chat Completi

Distil an expensive LLM API call on a narrow task into a small local model: capture traffic, curate, LoRA fine-tune with MLX on Apple Silico

Serve DeepSeek and Qwen, run an on-demand model library, and fine-tune with Unsloth QLoRA on 2x NVIDIA DGX Spark. TP=2 vLLM lanes behind one