FlagOpen / FlagGems
FlagGems is an operator library for large language models implemented in Triton Language.
☆496Updated this week
Alternatives and similar repositories for FlagGems:
Users that are interested in FlagGems are comparing it to the libraries listed below
- Distributed Triton for Parallel Systems☆451Updated 2 weeks ago
- A collection of memory efficient attention operators implemented in the Triton language.☆262Updated 10 months ago
- Puzzles for learning Triton, play it with minimal environment configuration!☆296Updated 4 months ago
- A Easy-to-understand TensorOp Matmul Tutorial☆342Updated 7 months ago
- flash attention tutorial written in python, triton, cuda, cutlass☆334Updated 3 months ago
- Dynamic Memory Management for Serving LLMs without PagedAttention☆354Updated last week
- Development repository for the Triton-Linalg conversion☆185Updated 2 months ago
- Several optimization methods of half-precision general matrix multiplication (HGEMM) using tensor core with WMMA API and MMA PTX instruct…☆393Updated 7 months ago
- ☆198Updated 9 months ago
- Domain-specific language designed to streamline the development of high-performance GPU/CPU/Accelerators kernels☆1,052Updated this week
- Shared Middle-Layer for Triton Compilation☆246Updated last week
- Materials for learning SGLang☆387Updated last month
- Yinghan's Code Sample☆323Updated 2 years ago
- ☆148Updated 3 months ago
- 📚FFPA(Split-D): Yet another Faster Flash Attention with O(1) GPU SRAM complexity large headdim, 1.8x~3x↑🎉 faster than SDPA EA.☆169Updated 2 weeks ago
- learning how CUDA works☆238Updated last month
- Disaggregated serving system for Large Language Models (LLMs).☆562Updated 2 weeks ago
- FlagScale is a large model toolkit based on open-sourced projects.☆268Updated this week
- Analyze the inference of Large Language Models (LLMs). Analyze aspects like computation, storage, transmission, and hardware roofline mod…☆448Updated 7 months ago
- A model compilation solution for various hardware☆427Updated this week
- how to learn PyTorch and OneFlow☆425Updated last year
- USP: Unified (a.k.a. Hybrid, 2D) Sequence Parallel Attention for Long Context Transformers Model Training and Inference☆477Updated this week
- Zero Bubble Pipeline Parallelism☆383Updated 2 weeks ago
- [MLSys'25] QServe: W4A8KV4 Quantization and System Co-design for Efficient LLM Serving; [MLSys'25] LServe: Efficient Long-sequence LLM Se…☆643Updated last month
- ☆131Updated last month
- AI Accelerator Benchmark focuses on evaluating AI Accelerators from a practical production perspective, including the ease of use and ver…☆236Updated last week
- BitBLAS is a library to support mixed-precision matrix multiplications, especially for quantized LLM deployment.☆594Updated 2 months ago
- A simple high performance CUDA GEMM implementation.☆361Updated last year
- Mirage: Automatically Generating Fast GPU Kernels without Programming in Triton/CUDA☆808Updated this week
- ☆115Updated 4 months ago