tlc-pack / libflash_attnLinks
Standalone Flash Attention v2 kernel without libtorch dependency
☆110Updated last year
Alternatives and similar repositories for libflash_attn
Users that are interested in libflash_attn are comparing it to the libraries listed below
Sorting:
- A standalone GEMM kernel for fp16 activation and quantized weight, extracted from FasterTransformer☆94Updated 3 weeks ago
- ⚡️Write HGEMM from scratch using Tensor Cores with WMMA, MMA and CuTe API, Achieve Peak⚡️ Performance.☆119Updated 4 months ago
- Performance of the C++ interface of flash attention and flash attention v2 in large language model (LLM) inference scenarios.☆40Updated 7 months ago
- 使用 cutlass 仓库在 ada 架构上实现 fp8 的 flash attention☆75Updated last year
- We invite you to visit and follow our new repository at https://github.com/microsoft/TileFusion. TiledCUDA is a highly efficient kernel …☆186Updated 8 months ago
- High-speed GEMV kernels, at most 2.7x speedup compared to pytorch baseline.☆116Updated last year
- ☆32Updated 8 months ago
- TileFusion is an experimental C++ macro kernel template library that elevates the abstraction level in CUDA C for tile processing.☆97Updated 3 months ago
- llama INT4 cuda inference with AWQ☆55Updated 8 months ago
- ☆98Updated last year
- Several optimization methods of half-precision general matrix vector multiplication (HGEMV) using CUDA core.☆67Updated last year
- ☆64Updated 5 months ago
- Benchmark code for the "Online normalizer calculation for softmax" paper☆101Updated 7 years ago
- GPTQ inference TVM kernel☆39Updated last year
- Implement Flash Attention using Cute.☆96Updated 9 months ago
- ☆106Updated 4 months ago
- ☆56Updated 2 months ago
- play gemm with tvm☆91Updated 2 years ago
- Decoding Attention is specially optimized for MHA, MQA, GQA and MLA using CUDA core for the decoding stage of LLM inference.☆44Updated 3 months ago
- Triton adapter for Ascend. Mirror of https://gitee.com/ascend/triton-ascend☆76Updated last week
- 🤖FFPA: Extend FlashAttention-2 with Split-D, ~O(1) SRAM complexity for large headdim, 1.8x~3x↑🎉 vs SDPA EA.☆220Updated last month
- Multiple GEMM operators are constructed with cutlass to support LLM inference.☆19Updated 2 months ago
- A Triton JIT runtime and ffi provider in C++☆25Updated 2 weeks ago
- ☆95Updated 6 months ago
- ☆150Updated 8 months ago
- ☆13Updated 7 months ago
- ☆43Updated last year
- DeeperGEMM: crazy optimized version☆71Updated 5 months ago
- CUDA 8-bit Tensor Core Matrix Multiplication based on m16n16k16 WMMA API☆32Updated 2 years ago
- This is a demo how to write a high performance convolution run on apple silicon☆54Updated 3 years ago