Popular repositories Loading

1. Flash-Attention-v15 Flash-Attention-v15 Public

   High-performance CUDA implementation of FlashAttention featuring online softmax, shared-memory tiling, warp-level reductions, and vectorized memory accesses.

   Jupyter Notebook 2

2. a1coderxt a1coderxt Public

   Config files for my GitHub profile.

   Jupyter Notebook

3. LLM-Inference-Engine LLM-Inference-Engine Public

   Continuous LLM inference engine with paged KV-cache and dynamic batching

   Python

4. GeMM-CUDA-KERNEL GeMM-CUDA-KERNEL Public

   From naive matrix multiplication to an optimized CUDA GEMM kernel with hierarchical tiling and performance analysis.

   Jupyter Notebook

5. vllm vllm Public

   Forked from vllm-project/vllm

   A high-throughput and memory-efficient inference and serving engine for LLMs

   Python