luliyucoordinate / cute-flash-attentionLinks
Implement Flash Attention using Cute.
☆85Updated 5 months ago
Alternatives and similar repositories for cute-flash-attention
Users that are interested in cute-flash-attention are comparing it to the libraries listed below
Sorting:
- ☆73Updated 2 weeks ago
- 使用 cutlass 仓库在 ada 架构上实现 fp8 的 flash attention☆68Updated 9 months ago
- ⚡️Write HGEMM from scratch using Tensor Cores with WMMA, MMA and CuTe API, Achieve Peak⚡️ Performance.☆79Updated 3 weeks ago
- A GPU-optimized system for efficient long-context LLMs decoding with low-bit KV cache.☆36Updated last month
- SpInfer: Leveraging Low-Level Sparsity for Efficient Large Language Model Inference on GPUs☆47Updated 2 months ago
- ☆59Updated last month
- Examples of CUDA implementations by Cutlass CuTe☆188Updated 4 months ago
- A lightweight design for computation-communication overlap.☆132Updated 3 weeks ago
- ☆96Updated 8 months ago
- Decoding Attention is specially optimized for MHA, MQA, GQA and MLA using CUDA core for the decoding stage of LLM inference.☆36Updated 2 months ago
- ☆121Updated 5 months ago
- ☆33Updated last year
- We invite you to visit and follow our new repository at https://github.com/microsoft/TileFusion. TiledCUDA is a highly efficient kernel …☆182Updated 4 months ago
- Standalone Flash Attention v2 kernel without libtorch dependency☆109Updated 8 months ago
- Performance of the C++ interface of flash attention and flash attention v2 in large language model (LLM) inference scenarios.☆37Updated 3 months ago
- Several optimization methods of half-precision general matrix vector multiplication (HGEMV) using CUDA core.☆62Updated 8 months ago
- ☆18Updated 2 months ago
- PyTorch bindings for CUTLASS grouped GEMM.☆93Updated last week
- 📚FFPA(Split-D): Extend FlashAttention with Split-D for large headdim, O(1) GPU SRAM complexity, 1.8x~3x↑🎉 faster than SDPA EA.☆184Updated 3 weeks ago
- Optimize softmax in triton in many cases☆20Updated 8 months ago
- TileFusion is an experimental C++ macro kernel template library that elevates the abstraction level in CUDA C for tile processing.☆88Updated last week
- Optimize GEMM with tensorcore step by step☆26Updated last year
- ☆109Updated 3 weeks ago
- ☆73Updated 4 months ago
- A standalone GEMM kernel for fp16 activation and quantized weight, extracted from FasterTransformer☆92Updated last week
- ☆48Updated 3 weeks ago
- GPTQ inference TVM kernel☆40Updated last year
- Quantized Attention on GPU☆44Updated 6 months ago
- A practical way of learning Swizzle☆19Updated 4 months ago
- DeeperGEMM: crazy optimized version☆69Updated last month