BBuf / tensorrt-llm-moeLinks
☆33Updated 8 months ago
Alternatives and similar repositories for tensorrt-llm-moe
Users that are interested in tensorrt-llm-moe are comparing it to the libraries listed below
Sorting:
- 使用 cutlass 仓库在 ada 架构上实现 fp8 的 flash attention☆76Updated last year
- Standalone Flash Attention v2 kernel without libtorch dependency☆112Updated last year
- ⚡️Write HGEMM from scratch using Tensor Cores with WMMA, MMA and CuTe API, Achieve Peak⚡️ Performance.☆121Updated 5 months ago
- A standalone GEMM kernel for fp16 activation and quantized weight, extracted from FasterTransformer☆94Updated last month
- Decoding Attention is specially optimized for MHA, MQA, GQA and MLA using CUDA core for the decoding stage of LLM inference.☆45Updated 4 months ago
- Performance of the C++ interface of flash attention and flash attention v2 in large language model (LLM) inference scenarios.☆41Updated 7 months ago
- ☆56Updated 3 months ago
- ☆100Updated last year
- ☆65Updated 6 months ago
- Multiple GEMM operators are constructed with cutlass to support LLM inference.☆20Updated 2 months ago
- ☆150Updated 9 months ago
- ☆107Updated 5 months ago
- ☆97Updated 7 months ago
- A Triton JIT runtime and ffi provider in C++☆27Updated this week
- Several optimization methods of half-precision general matrix vector multiplication (HGEMV) using CUDA core.☆67Updated last year
- Implement Flash Attention using Cute.☆96Updated 10 months ago
- ☆44Updated last year
- 使用 CUDA C++ 实现的 llama 模型推理框架☆62Updated 11 months ago
- llama INT4 cuda inference with AWQ☆55Updated 9 months ago
- 🤖FFPA: Extend FlashAttention-2 with Split-D, ~O(1) SRAM complexity for large headdim, 1.8x~3x↑🎉 vs SDPA EA.☆223Updated 2 months ago
- Benchmark code for the "Online normalizer calculation for softmax" paper☆102Updated 7 years ago
- 使用 cutlass 实现 flash-attention 精简版,具有教学意义☆50Updated last year
- A practical way of learning Swizzle☆29Updated 8 months ago
- ☆91Updated last week
- ☆14Updated 7 months ago
- Tutorials of Extending and importing TVM with CMAKE Include dependency.☆14Updated last year
- QQQ is an innovative and hardware-optimized W4A8 quantization solution for LLMs.☆144Updated 2 months ago
- ☆139Updated last year
- Triton adapter for Ascend. Mirror of https://gitee.com/ascend/triton-ascend☆76Updated 3 weeks 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