最有价值项目

A multimodal live AI assistant designed to enhance the browsing experience using Gemini.

Claude Code hook toolkit that gives vision-blind models a text description of pasted/tool-produced images.

Run NVIDIA's tri-modal Nemotron Omni (text+vision+audio) entirely on Apple Silicon — pure MLX runtime, parity-verified against NVIDIA's reference

Vision-Language Models on AMD GPUs — LLaVA, MiniGPT-4, Idefics on ROCm 🚀

MiniMax H3 全域提示词工程 Skill｜T2VA / I2VA / FL2VA / L2VA / Ref2VA｜50+ 创作玩法｜人物一致性、视频编辑、动作参考、音频控制、广告、短剧、MV、3D、数字人、Prompt 修复｜ComfyUI / RunningHub / Codex

[ACCV 2022] Dataset with 1000+ video advertisemets proposed by "Multi-modal Segment Assemblage Network for Ad Video Editing with Importance-Coherence Reward"

[ICCV 2025] Know "No" Better: A Data-Driven Approach for Enhancing Negation Awareness in CLIP

An invisible AI desktop assistant that evades screen capture and taskbar detection. Features instant visual analysis, live Whisper voice transcription, and robust support for premium LLMs (Anthropic, OpenAI, OpenRouter, Gemini, and local Ollama).

Multi-Agent Comic Generator powered by the @google-gemini API. Turn simple prompts into comics with AI-generated stories, consistent characters, auto speech bubbles, and 7 language support. Built with FastAPI and deployed on Google Cloud Run.

nvtop for vLLM — an interactive terminal dashboard for vLLM serving performance (concurrency, throughput, cache & KV memory, latency, spec-decode, GPU)

A slim, local-LLM-first AI coding assistant. TUI, Web UI, and desktop app. Runs air-gapped with Ollama, vLLM, or any OpenAI-compatible endpoint.

Terraform setup for deploying a private coding LLM on Vast.ai with vLLM, Qwen3 Coder, and OpenCode.

在nano-vllm基础上支持moe模型以及Speculative Decoding技术

GLM-5.2 QuantTrio (Unpruned ) on 4xDGX Spark (GB10): 327K-655K context or up to 5 concurrent agents, one receipe. TP4 + DCP + MTP

Extend LLM context windows beyond GPU memory limits with disk-backed KV cache.

Kubernetes-native control plane for scale-to-zero serving of long-tail LLMs

Tencent Hy3 295B MoE (NVFP4) on 2x NVIDIA DGX Spark — TP2 over 200GbE, 256K context, 26 tok/s end-to-end. Tuned, benchmarked, agent-ready.

Multi-model LLM serving for NVIDIA DGX Spark with vLLM, web UI, and tool calling

Repository for Multililngual Generation, RAG evaluations, and surrogate judge training for Arena RAG leaderboard (NAACL'25)

Tiny, stateless Go router that dispatches OpenAI-compatible requests to single-model vLLM and sglang backends with zero external dependencies