ShaYeBuHui01 / flash_attention_inference
Performance of the C++ interface of flash attention and flash attention v2 in large language model (LLM) inference scenarios.
☆14Updated last year
Alternatives and similar repositories for flash_attention_inference:
Users that are interested in flash_attention_inference are comparing it to the libraries listed below
- ☆26Updated 10 months ago
- Standalone Flash Attention v2 kernel without libtorch dependency☆103Updated 5 months ago
- CPU Memory Compiler and Parallel programing☆25Updated 2 months ago
- ☆142Updated last month
- Several optimization methods of half-precision general matrix vector multiplication (HGEMV) using CUDA core.☆55Updated 5 months ago
- ☆33Updated 3 weeks ago
- ☆35Updated 4 months ago
- ☆97Updated 2 months ago
- This project is about convolution operator optimization on GPU, include GEMM based (Implicit GEMM) convolution.☆25Updated last month
- 使用 cutlass 仓库在 ada 架构上实现 fp8 的 flash attention☆53Updated 6 months ago
- An extension library of WMMA API (Tensor Core API)☆87Updated 7 months ago
- Performance of the C++ interface of flash attention and flash attention v2 in large language model (LLM) inference scenarios.☆34Updated 5 months ago
- Examples of CUDA implementations by Cutlass CuTe☆137Updated last week
- We invite you to visit and follow our new repository at https://github.com/microsoft/TileFusion. TiledCUDA is a highly efficient kernel …☆175Updated 2 weeks ago
- ⚡️Write HGEMM from scratch using Tensor Cores with WMMA, MMA and CuTe API, Achieve Peak⚡️ Performance.☆51Updated last week
- Decoding Attention is specially optimized for multi head attention (MHA) using CUDA core for the decoding stage of LLM inference.☆29Updated 3 months ago
- llama INT4 cuda inference with AWQ☆50Updated 3 weeks ago
- ☆108Updated 10 months ago
- ☆80Updated last year
- Multiple GEMM operators are constructed with cutlass to support LLM inference.☆16Updated 4 months ago
- Implement Flash Attention using Cute.☆69Updated last month
- A standalone GEMM kernel for fp16 activation and quantized weight, extracted from FasterTransformer☆88Updated 11 months ago
- play gemm with tvm☆86Updated last year
- Benchmark code for the "Online normalizer calculation for softmax" paper☆66Updated 6 years ago
- ☆57Updated 2 months ago
- QQQ is an innovative and hardware-optimized W4A8 quantization solution for LLMs.☆100Updated 2 months ago
- ☆108Updated 10 months ago
- ☆180Updated 7 months ago
- ☆69Updated last month