cuda-graphs

Open repos and models their makers tagged cuda-graphs, most starred first. The same for everyone.

AI on my own machineRepo

StartLux-Prism

Fast, exact inference for StartLux-Decision models on NVIDIA GPUs: the released GGUF files on int8 and 4-bit tensor cores, behind the /v1/systemone API, with up to 5.3× lower latency than vLLM in FP8 and bit-identical answers.

  • 560
  • Python
  • 4 wk ago
StartLuxLabsTake it
AI on my own machineRepo

sequence-ai

SequenceLink: open source serving for robot policies on remote GPUs, verified with openpi, LeRobot, GR00T and Octo. An inference layer (CUDA graphs with the same outputs, replicas) and a network layer (a robot's connection to its GPU), on your own GPU or on the hosted SequenceLink platform. pip install sequence-ai

  • 58
  • Python
  • Apache-2.0
  • 1 d ago
AI on my own machineRepo

Annotated-LLM-Runtime

From-scratch, heavily-annotated CUDA inference runtime for Qwen2.5-Coder-7B on H100 (sm_90). Custom INT4 packer, fused GEMV, paged KV, split-KV attention, CUDA graph decode — every hot path commented for the why. Educational, not a llama.cpp replacement.

  • 21
  • Cuda
  • 3 mo ago
Photos, music and video at homeRepo

tacuda

GPU-accelerated technical analysis: 237 TA-Lib 0.8.3 functions, typed outputs, portfolio batches, CUDA Graphs and incremental streaming. Python, Node.js and .NET.

  • 5
  • C++
  • Apache-2.0
  • 3 d ago
Ta-CudaTake it
GitHubRepo

cutile-forge

Tile-Fusion Graph compiler + hybrid PCIe-streaming LLM engine on NVIDIA CuTile (sm_89 ↔ sm_120). 8 CUDA kernels → 3 Python tiles; 117 tok/s 7B, 70B in one 32 GB GPU at 80.8% of PCIe 5.0 ceiling; torch.compile backend.

  • 0
  • Python
  • MIT
  • 2 d ago
samudraguptoTake it
Search GitHub and Hugging Face for cuda-graphs

Get , open it once and sign in. Then come back here and press the button again.

cuda-graphs: open source repos and models on Take My Repo