Triton Python Packages

liger-kernel

Efficient Triton Kernels for LLM Training

sageattention

[ICLR2025, ICML2025, NeurIPS2025 Spotlight] Quantized Attention achieves speedup of 2-5x compared to FlashAttention, without losing end-to-end metrics across language, image, and video models.

liger-kernel-nightly

Efficient Triton Kernels for LLM Training

tritonparse

TritonParse: A Compiler Tracer, Visualizer, and Reproducer for Triton Kernels

clearml-serving

ClearML - Model-Serving Orchestration and Repository Solution

fastdeploy

Deploy DL/ ML inference pipelines with minimal extra code.

turboquant-vllm

TurboQuant KV cache compression plugin for vLLM — asymmetric K/V, 8 models validated, consumer GPUs

tokenspeed-iris

AMD RAD's multi-GPU Triton-based framework for seamless multi-GPU programming

hilbertsfc

Ultra-fast 2D & 3D Hilbert curve kernels in Python. JIT compiled, branchless, L1-cache-friendly lookup tables, loop unrolling, SIMD, and multi-threading.

dlblas

DLBlas: clean and efficient kernels

yolo26-kit

YOLO26 ↔ YOLOv8-pipeline bridge: shape adapters, decoder, letterbox, and ORT wrappers.

transformersplus

Add Some plus extra features to transformers

flag-gems

FlagGems is a function library written in Triton language.

flash-sparse-attn

Trainable fast and memory-efficient sparse attention

fdclient

Deploy DL/ ML inference pipelines with minimal extra code.

triformer

Transformer components in Triton

grammared-language

Adding Grammarly (and other) open source ML models to LanguageTool

triton-runner

Multi-Level Triton Runner supporting Python, IR, PTX, and cubin.

atlas-quantum

GPU-accelerated quantum tensor network simulator with coherence-aware VQE, adaptive MPS, molecular chemistry, VRA integration, and cuQuantum backend

fused-turboquant

Fused Triton kernels for TurboQuant KV cache compression — 2-4 bit quantization with RHT rotation. Drop-in HuggingFace & vLLM integration. Up to 4.9x KV cache compression for Llama, Qwen, Mistral, and more.

torchgw

TorchGW — Fast Sampled Gromov-Wasserstein optimal transport in pure PyTorch. GPU-accelerated with Triton fused Sinkhorn kernels. 3-175x faster than POT.

hip-attn

Training-free Post-training Efficient Sub-quadratic Complexity Attention. Implemented with OpenAI Triton.

flash-sinkhorn

Sinkhorn optimal transport kernels in PyTorch + Triton (squared Euclidean, no cost matrix materialization).

tritonllm

LLM Inference via Triton (Flexible & Modular): Focused on Kernel Optimization using CUBIN binaries, Starting from gpt-oss Model