FLA

All

8 repositories

flash-linear-attention
Public
🚀 Efficient implementations of state-of-the-art linear attention models in Torch and Triton
natural-language-processing machine-learning-systems large-language-models
Python
•
MIT License
•144•2.2k•26•4•Updated Apr 4, 2025Apr 4, 2025
fla-zoo
Public
Flash-Linear-Attention models beyond language
computer-vision flash-linear-attention
Python
•1•9•0•0•Updated Apr 3, 2025Apr 3, 2025
flame
Public
🔥 A minimal training framework for scaling FLA models
sequence-modeling large-language-models flash-linear-attention
Python
•
MIT License
•15•94•0•1•Updated Apr 1, 2025Apr 1, 2025
fla-rl
Public
A minimal RL frame work for scaling FLA models on long-horizon reasoning and agentic scenarios.
MIT License
•0•4•0•0•Updated Apr 1, 2025Apr 1, 2025
ThunderKittens
Public
Tile primitives for speedy kernels
Cuda
•
MIT License
•133•2•0•0•Updated Mar 27, 2025Mar 27, 2025
native-sparse-attention
Public
🐳 Efficient Triton implementations for "Native Sparse Attention: Hardware-Aligned and Natively Trainable Sparse Attention"
Python
•
MIT License
•30•608•9•0•Updated Mar 19, 2025Mar 19, 2025
flash-hybrid-attention
Public
0•7•0•0•Updated Mar 5, 2025Mar 5, 2025
flash-bidirectional-linear-attention
Public
Triton implement of bi-directional (non-causal) linear attention
computer-vision machine-learning-systems triton-lang
Python
•
MIT License
•1•44•0•0•Updated Feb 4, 2025Feb 4, 2025