nvidia

Audio2Face Avatar with Riva SDK functionality

Distill teacher chains-of-thought into a LoRA adapter via a strict boxed-answer format contract + two-phase Train→Nudge (silver-medal NVIDIA Nemotron reasoning recipe, as a tested library).

Experiments to test different speech recognition systems for SEPIA Framework

OpenAI-compatible gateway aggregating 8 free LLM providers with automatic failover, circuit breakers, and smart routing. 32+ models, zero cost.

Push-to-talk voice dictation for Linux. 100% local, multilingual (25+ languages), with speaker diarization. Qt frontend, Rust backend on NVIDIA Parakeet via ONNX Runtime. KDE Plasmoid integrated.

DFlash vLLM for DGX Spark — Plug & Play Block-Diffusion Speculative Decoding

Autonomous AI SOC for Wazuh SIEM — 11 security-expert agents triage, correlate, investigate and plan responses 24/7, every action gated by two-tier human approval. Runs on OpenClaw, Hermes or NVIDIA NemoClaw; any LLM, self-hosted vLLM, or fully air-gapped Ollama. Evidence packs, SOC KPIs, Slack approvals.

Bleeding edge vLLM Docker image for the NVIDIA DGX Spark (GB10 / sm_121a).

Legacy: self-hosted LLM/RAG stack installer for DGX Spark (arm64). The work continued as AGmind Systems Lab — measured local-LLM performance evidence (Strix Halo, DGX Spark, llama.cpp, vLLM): https://agmind.ai

Native Windows vLLM 0.27.1 wheels: Python 3.13, PyTorch 2.13 + CUDA 13.0, SM 7.5-12.0 for RTX 20/30/40/50, OpenAI-compatible serving, FlashAttention/Rust, 10 KV formats, Multi-TurboQuant, and experimental CPU/RAM/NVMe prompt-KV offload - no WSL or Docker.

Real-time performance metrics and statistics for your Nvidia GPU

A production-ready Docker setup for ComfyUI that unlocks the full potential of NVIDIA Blackwell GPUs (RTX 50 series) through 4-bit quantization with NVFP4.

On-device speech-to-text for iOS/macOS powered by NVIDIA NeMo Conformer CTC Small (13M params). Pure Swift + CoreML implementation with automatic audio padding, chunking for long audio, and real-time recognition.

Dual-engine (llama.cpp + vLLM) LLM benchmarking pipeline for GGUF & safetensors on NVIDIA GPUs — speed, quality, live dashboard, publishable cards.

基于 SenseVoice 的 Windows 本地语音转文字工具，支持 OpenAI 格式 API 润色，低延迟，高精度。

A "standard library" of Triton kernels.

Unified KV-cache compression for LLM inference: 12 Python-native methods, guarded add-on composition and routing, analytical capacity simulation, Godzilla KVarN/TriAttention with SM86/SM89 qualification, CUDA weight sharing, and multi-GPU planning.

Run and manage both vLLM and llama.cpp model servers from one terminal UI + CLI — shared Docker Compose profiles, live tok/s, and cross-backend port/GPU conflict checks. NVIDIA + Linux.

LLM fine-tuning with LoRA + NVFP4/MXFP8 on NVIDIA DGX Spark (Blackwell GB10)

A curated list of tools, frameworks, and resources for running AI locally — inference, RAG, orchestration, monitoring, and more.