Step by step implementation of a fast softmax kernel in CUDA
☆70Jan 6, 2025Updated last year
Alternatives and similar repositories for FastSoftmax
Users that are interested in FastSoftmax are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.
Sorting:
- ☆98May 30, 2026Updated last month
- A CUDA kernel for NHWC GroupNorm for PyTorch☆23Nov 15, 2024Updated last year
- Tensara's GPU programming problems☆21Apr 23, 2026Updated 3 months ago
- learn TensorRT from scratch🥰☆18Sep 29, 2024Updated last year
- ☆18Mar 12, 2025Updated last year
- Managed Kubernetes at scale on DigitalOcean • AdDigitalOcean Kubernetes includes the control plane, bandwidth allowance, container registry, automatic updates, and more for free.
- The most comprehensive community resource for the NVIDIA CMP 170HX.☆43Jul 13, 2026Updated 2 weeks ago
- Hand-Rolled GPU communications library☆96Nov 25, 2025Updated 8 months ago
- Row-wise block scaling for fp8 quantization matrix multiplication. Solution to GPU mode AMD challenge.☆19Feb 9, 2026Updated 5 months ago
- A collection of GPU experiments and benchmarks for my personal understanding and research.☆34Updated this week
- Flash Attention in ~100 lines of CUDA (forward pass only)☆12Jun 10, 2024Updated 2 years ago
- ACPO: Python ML training/inference framework for LLVM optimizations☆13Jul 2, 2025Updated last year
- paper-read-notes☆13Sep 26, 2024Updated last year
- Standalone commandline CLI tool for compiling Triton kernels☆20Sep 13, 2024Updated last year
- Personal solutions to the Triton Puzzles☆21Jul 18, 2024Updated 2 years ago
- Deploy to Railway using AI coding agents - Free Credits Offer • AdUse Claude Code, Codex, OpenCode, and more. Autonomous software development now has the infrastructure to match with Railway.
- Fast CUDA matrix multiplication from scratch☆1,262Sep 2, 2025Updated 10 months ago
- ☆14Nov 3, 2025Updated 8 months ago
- Super fast FP32 matrix multiplication on RDNA3☆92Mar 30, 2025Updated last year
- A GPT-2 inference engine written from scratch in CUDA and C++. Implements custom CUDA kernels for tiled matrix multiplication, LayerNorm,…☆43May 17, 2026Updated 2 months ago
- TensorRT-in-Action 是一个 GitHub 代码库，提供了使用 TensorRT 的代码示例，并有对应 Jupyter Notebook。☆15Jun 1, 2023Updated 3 years ago
- 搜藏的希望的代码片段☆13Jun 6, 2023Updated 3 years ago
- HunyuanDiT with TensorRT and libtorch☆18May 22, 2024Updated 2 years ago
- a ai infra framework for edge device base on nndeploy☆18Nov 27, 2025Updated 8 months ago
- 使用mnn-llm对GOT-OCR2.0进行推理☆14Oct 2, 2024Updated last year
- Bare Metal GPUs on DigitalOcean Gradient AI • AdPurpose-built for serious AI teams training foundational models, running large-scale inference, and pushing the boundaries of what's possible.
- ☆173Updated this week
- Flash Attention in raw Cuda C beating PyTorch☆39May 14, 2024Updated 2 years ago
- Implementation of a histogram equalization program using CUDA. Histogram equalization is a technique for adjusting image intensities to e…☆13Jan 3, 2021Updated 5 years ago
- A high-performance attention mechanism that computes softmax normalization in a single streaming pass using running accumulators (online …☆31Jun 3, 2026Updated last month
- This repository includes code to reproduce the tables in "Loss Landscapes are All You Need: Neural Network Generalization Can Be Explaine…☆39Mar 2, 2023Updated 3 years ago
- ☆19Jun 6, 2025Updated last year
- segmentation algorithm yolact use tensorrt deploy☆14May 7, 2022Updated 4 years ago
- ☆19Jan 4, 2024Updated 2 years ago
- Quantized LLM training in pure CUDA/C++.☆251Updated this week
- Deploy to Railway using AI coding agents - Free Credits Offer • AdUse Claude Code, Codex, OpenCode, and more. Autonomous software development now has the infrastructure to match with Railway.
- Gemstones: A Model Suite for Multi-Faceted Scaling Laws (NeurIPS 2025)☆35Sep 28, 2025Updated 9 months ago
- A lightweight, single-header C++11 Jinja2 template engine for LLM chat templates.☆20Mar 4, 2026Updated 4 months ago
- 🤖FFPA: Extends FA-2/3 via Split-D for large headdims, 1.5x~6×↑🎉 vs SDPA, up to 513~535 TFLOPS🎉 on NVIDIA H200.☆316Updated this week
- Inference Llama 2 in one file of pure Cuda☆17Aug 20, 2023Updated 2 years ago
- arkit demo☆11Aug 20, 2018Updated 7 years ago
- 一个轻量化的大模型推理框架☆23May 26, 2025Updated last year
- IMMA-based **FP8-as-storage** GEMM experiments for Ampere (sm_86 / RTX 3090 Ti).☆24Jan 30, 2026Updated 5 months ago