Kruk2 / jasna
Desktop app and CLI that restores mosaics in videos and images using GPU models like BasicVSR++ and LTX.
Kruk2 / jasna
eagle-hq / EAGLE-4
Official Inference Code of EAGLE-4
kanaworksai-hue / audio-studio-mcp
Local AI BGM and sound effects for Claude Code and any MCP client. Stable Audio 3 on Apple Silicon (MLX) and Windows NVIDIA GPUs (PyTorch). No cloud, no API key.
academa-labs / manimgx
The animation engine for agents: blazingly fast math and 3D videos with Manim's Python API, on a Rust GPU core.
Panchovix / nvidia-windows-vbios-vix-loader
Load a modified VBIOS for NVIDIA GPUs on Windows through the driver's registry override, without flashing
tile-ai / tilelang
A Python-embedded domain-specific language and TVM-based compiler for writing high-performance GPU, CPU, and NPU kernels.
bojieli / ai-infra-book
《深入理解 AI Infra：量化分析与系统设计》（李博杰 著）开源书稿：从硬件约束和模型架构出发，量化推导 LLM 推理与训练系统设计。含全书正文、PDF、配套计算工具与实验
MakazhanAlpamys / Soup
Fine-tune LLMs from one YAML. Layer streaming trains an 8B model on a 4 GB laptop GPU.
NVIDIA / TensorRT-LLM
TensorRT LLM provides users with an easy-to-use Python API to define Large Language Models (LLMs) and supports state-of-the-art optimizations to perform inference efficiently on NVIDIA GPUs. TensorRT LLM also contains components to create Python and C++ runtimes that orchestrate the inference execution in a performant way.
ovg-project / kvcached
Virtualized Elastic KV Cache for Dynamic GPU Sharing and Beyond
llm-d / llm-d
Achieve state of the art inference performance with modern accelerators on Kubernetes
syv-ai / HyperQwen
Serve large Qwen models fast on the GPUs you actually own. Qwen3.8-27B on a single 24 GB card with vLLM: 127 tok/s single-user (381 when the answer quotes the prompt), ~1,035 tok/s at 64 concurrent, 150k-262k context. vLLM patches, requant pipeline, benchmarks.
Lore-Hex / token-exchange-neocloud-in-a-box
Open-source confidential GPU inference cloud in a box: Intel TDX + NVIDIA confidential-computing nodes for DeepSeek-V4.1-Flash and GLM-5.3, verifiable and listed on TrustedRouter
Panchovix / nvidia-linux-vbios-vix-loader
Load a modified VBIOS per NVIDIA GPU at boot on Linux, in memory, through the open kernel modules
kovidgoyal / kitty
If you live in the terminal, kitty is made for you! Cross-platform, fast, feature-rich, GPU based.
baidu-baige / LoongForge
A high-performance framework for training LLMs, VLMs, diffusion, and embodied models on NVIDIA GPUs and Kunlun XPUs. Supports Pi0.5, GR00T-N1.7, FastWAM, Cosmos3, DreamZero, DeepSeek-V4, GLM-5.3, Kimi-K3 and more.
j1s4nn / jev-adaptive-granularity-WM
Official code, locked protocols, and audit artifacts for "Jev-Routed Temporal Granularity with Recovery and Hysteresis for Causal Video Generation": per-segment fine/coarse/hold routing for Causal-rCM (Wan2.1-1.3B) via the Jev oracle, with two-GPU (T4/A100) feasibility audits.
surya-koritala / condor
Qwen3.8-Flash-Next (180B MoE) on one 12-16 GB NVIDIA GPU, Linux or Windows. Own engine, own compression.
newton-physics / newton
An open-source, GPU-accelerated physics simulation engine built upon NVIDIA Warp, specifically targeting roboticists and simulation researchers.
turboderp-org / exllamav3
An optimized quantization and inference library for running LLMs locally on modern consumer-class GPUs
wafer-ai / gpu-perf-engineering-resources
A curated resource list for learning AI performance engineering, from GPU fundamentals to production inference.
noonghunna / club-3090
Community recipes for serving LLMs on RTX 3090/4090/5090 CUDA gpus. Multi-engine (vLLM, sglang, llama.cpp and other custom engines) and model-agnostic. Currently shipping Qwen3.8-27B Qwen3.6 35B Gemma 4 26B Gemma 4 31B configs for 1×/2×/4x/8x cards.
tiantianGPU / reg-factory
天天卡网
ChenlizheMe / Infernux
Building a Neural Network-Native Engine (3N). Python authoring, C++ runtime, Vulkan/WebGPU. Windows, Linux, Android and Web.
marcinz606 / NegPy
Tool for processing film negatives.
Gxj230958 / TritonAscendBench
A migration benchmark that measures whether production GPU Triton kernels — extracted from vLLM and SGLang inference paths — can be ported to the Huawei Ascend 910B (Atlas A2) via `triton-ascend` without losing correctness or performance.
deepspeedai / DeepSpeed
DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.
jax-ml / jax
Composable transformations of Python+NumPy programs: differentiate, vectorize, JIT to GPU/TPU, and more
gpustack / gpustack
A GPU cluster manager for high-performance AI model serving (vLLM, SGLang) and on-demand SSH-accessible GPU instances.
remsky / Kokoro-FastAPI
Dockerized OpenAI-compatible wrapper for Kokoro-82M text-to-speech w/multiplatform CPU, AMD, NVIDIA GPU PyTorch; multi-speaker, clone-tuning, caption timestamps, SSML, optional readalong web UI
NVIDIA / TransformerEngine
A library for accelerating Transformer models on NVIDIA GPUs, including using 8-bit and 4-bit floating point (FP8 and FP4) precision on Hopper, Ada and Blackwell GPUs, to provide better performance with lower memory utilization in both training and inference.
EMI-Group / evox
Distributed GPU-Accelerated Framework for Evolutionary Computation. Comprehensive Library of Evolutionary Algorithms & Benchmark Problems.
selkies-project / selkies
Low-Latency Accelerated Web Remote Desktop Streaming Platform for Self-Hosting, Containers, Kubernetes, or Cloud/HPC
SemiAnalysisAI / InferenceX
Open Source AI Accelerator Research Platform Standard / 开源推理研究平台
tenstorrent / tt-metal
:metal: TT-NN operator library, and TT-Metalium low level kernel programming model.
local-inference-lab / rtx6kpro
RTX 6000 Pro Wiki — Running Large LLMs (Qwen3.5-397B, Kimi-K2.5, GLM-5) on PCIe GPUs without NVLink
canberk7 / ema-lightning
Tiny, fast and accurate Turkish text-to-speech. 8.6M parameters, 0.92% WER on Freya-TR-Eval, first audio in ~4 ms and 1,300× real time on one GPU. Streams audio, batches many callers on one GPU, and runs offline on a GPU or CPU.
jpezzulli / sglang-rtxpro6000
Optimized SGLang runtime for Qwen3.8-27B FP8 with DFlash2 and Qwen3.8 Flash-Next NVFP4 with FR-Spec on one NVIDIA RTX PRO 6000 Blackwell 96 GB GPU (SM120): 524K context, HiCache and NIXL.
Jit-Roy / WeeLLM
WeeLLM runs large diffusion models with as little as 4 GB of VRAM, without any quantization. It dynamically determines how many layers can fit within the available VRAM and streams the text encoder and transformer layers to the GPU layer by layer, enabling inference on hardware with limited VRAM. It supports both safetensors and GGUF models.
chris-ritsen / samsung-tv-root
Samsung QN90B and QN90F root
RodriMora / dgx-spark-grafana-dashboard
Grafana dashboard for NVIDIA DGX Spark clusters running vLLM: throughput, contention, and GB10 GPU hardware telemetry (NVIDIA-themed, 55 panels)
Han-DongHeun / nvmon
A fancy NVIDIA GPU monitor for the terminal. One Python file, no dependencies.
stas00 / ml-engineering
Machine Learning Engineering Open Book
skypilot-org / skypilot
The AI Compute Platform for frontier teams. SkyPilot turns fragmented AI compute into one AI supercomputer, so frontier AI teams build custom intelligence faster.
XuehaiPan / nvitop
An interactive NVIDIA-GPU process viewer and beyond, the one-stop solution for GPU process management.
NVIDIA / warp
A Python framework for GPU-accelerated simulation, robotics, and machine learning.
pytorch / executorch
On-device AI across mobile, embedded and edge for PyTorch
sktime / pytorch-forecasting
Time series forecasting with PyTorch
NVIDIA / physicsnemo
Open-source deep-learning framework for building, training, and fine-tuning deep learning models using state-of-the-art Physics-ML methods
pytorch / TensorRT
PyTorch/TorchScript/FX compiler for NVIDIA GPUs using TensorRT
thu-pacman / chitu
High-performance inference framework for large language models, focusing on efficiency, flexibility, and availability.
tensorflow / lingvo
Lingvo
meta-pytorch / torchrec
Pytorch domain library for recommendation systems
amap-cvlab / ABot-World
Infinite Interactive World Rollout on a Single Desktop GPU
NVIDIA / cutile-python
cuTile is a programming model for writing parallel kernels for NVIDIA GPUs
geomstats / geomstats
Computations and statistics on manifolds with geometric structures.
chubbyguan / chubbyskills
把中文全渠道内容（抖音 / B站 / 小红书 / 公众号 / X / 播客）采集进个人知识库的 14 个 AI Skill：图文存图、视频转文字稿、字幕优先免 GPU、RSS/YouTube 订阅调度与每日情报简报，附带知识库 MCP server。｜ Ingest Chinese content into your personal knowledge base — image/video routing, subtitle-first transcription, feed subscriptions with daily digests, and a KB MCP server.
NVIDIA / cudnn-frontend
cuDNN Frontend is NVIDIA's modern, open-source entry point to the cuDNN library and a growing collection of high-performance open-source kernels.
Motphys / UniLab
UniLab: A Heterogeneous Architecture for Robot RL Beyond GPU-Dominant Paradigms
NVlabs / SOMA-X
SOMA: Unifying Parametric Human Body Models