AlibabaPAI / FLASHNNLinks
☆100Updated last year
Alternatives and similar repositories for FLASHNN
Users that are interested in FLASHNN are comparing it to the libraries listed below
Sorting:
- ☆107Updated 5 months ago
- A lightweight design for computation-communication overlap.☆181Updated last week
- 使用 cutlass 仓库在 ada 架构上实现 fp8 的 flash attention☆76Updated last year
- Performance of the C++ interface of flash attention and flash attention v2 in large language model (LLM) inference scenarios.☆41Updated 7 months ago
- A standalone GEMM kernel for fp16 activation and quantized weight, extracted from FasterTransformer☆94Updated last month
- Implement Flash Attention using Cute.☆96Updated 10 months ago
- ☆65Updated 5 months ago
- NVSHMEM‑Tutorial: Build a DeepEP‑like GPU Buffer☆138Updated last month
- ☆240Updated last year
- ⚡️Write HGEMM from scratch using Tensor Cores with WMMA, MMA and CuTe API, Achieve Peak⚡️ Performance.☆121Updated 5 months ago
- PyTorch bindings for CUTLASS grouped GEMM.☆124Updated 4 months ago
- 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
- ☆141Updated 9 months ago
- ☆150Updated 9 months ago
- DeeperGEMM: crazy optimized version☆72Updated 5 months ago
- Benchmark code for the "Online normalizer calculation for softmax" paper☆101Updated 7 years ago
- Tile-based language built for AI computation across all scales☆67Updated 3 weeks ago
- An efficient GPU support for LLM inference with x-bit quantization (e.g. FP6,FP5).☆265Updated 3 months ago
- A GPU-optimized system for efficient long-context LLMs decoding with low-bit KV cache.☆60Updated last week
- Flash-LLM: Enabling Cost-Effective and Highly-Efficient Large Generative Model Inference with Unstructured Sparsity☆221Updated 2 years ago
- Triton adapter for Ascend. Mirror of https://gitee.com/ascend/triton-ascend☆76Updated 3 weeks ago
- PyTorch bindings for CUTLASS grouped GEMM.☆154Updated last week
- Standalone Flash Attention v2 kernel without libtorch dependency☆112Updated last year
- ☆148Updated 7 months ago
- High-speed GEMV kernels, at most 2.7x speedup compared to pytorch baseline.☆116Updated last year
- Sequence-level 1F1B schedule for LLMs.☆32Updated last month
- ☆150Updated 5 months ago
- Several optimization methods of half-precision general matrix vector multiplication (HGEMV) using CUDA core.☆67Updated last year
- 🤖FFPA: Extend FlashAttention-2 with Split-D, ~O(1) SRAM complexity for large headdim, 1.8x~3x↑🎉 vs SDPA EA.☆223Updated 2 months ago
- Examples of CUDA implementations by Cutlass CuTe☆241Updated 3 months ago