Pinned Loading

1. fast-rl-grpo fast-rl-grpo Public

   Optimizing RL training with GRPO. This repo implements the RL pipeline logic and optimizing it with different technique including custom kernels, reducing overhead mainly in grpo step. Features det…

   Python 1

2. Rainference Rainference Public

   Self-hosted LLM inference stack with OpenAI-compatible API. Deploy LLaMA models on your own bare-metal Kubernetes cluster with built-in billing, analytics, and management dashboard

   Python 2 1

3. Diffusion.cu Diffusion.cu Public

   Diffusion.cu A high-performance diffusion model built from scratch with core components optimized in CUDA for speed. It features custom Conv2d, GroupNorm, and attention using Flash Attention for ef…

   Python 1

4. llm-kernel-patch llm-kernel-patch Public

   patch transformer library llm part with specific llm cuda kernel

   Cuda 1

5. VectorSum-kernel VectorSum-kernel Public

   Vector sum GPU kernel optimization in CUDA

   Python 1

6. llm-d-inference-scheduler llm-d-inference-scheduler Public

   Forked from llm-d/llm-d-inference-scheduler

   Inference scheduler for llm-d

   Go