flash-linear-attention
🚀 Efficient implementations for emerging model architectures
GraphCanon updated 2d · GitHub synced 2d · 31 views this month
Decision brief
Flash-linear-attention accelerates linear attention mechanisms in large language models, using CUDA for optimal performance.
Good fit when
- High-performance requirements with Nvidia GPUs where CUDA can offer significant speed-ups
- Need to benchmark against multiple operations like chunk_retention and flash_attn
Avoid when
- Limited GPU hardware or no support for backend flavors like CUDA, ROCM, XPU, NPU, or CPU
- Do not require linear attention mechanism in modeling large language models or sequence data
Observed Jul 12, 2026 · Source: enrich:decision_facts
Verify the decision
Maintenance and security
Full trust report- Maintenance
- Very active (0d since push)
- As of 2d
- Provenance
- Not a fork · Organization account
- As of 2d
- Security (OSV)
- No lockfile
- As of 1mo
Public GitHub metadata and optional OSV scans. Signals, not a guarantee. Trust methodology.
Install
pip install flash-linear-attentionHow it fits your stack(11)
Typed graph edges - alternatives, integrations, successors, and dependencies. Ranked by relationship type, not raw GitHub stars.
Integrates
Depends on
Related
Relationship graph
Optional deeper exploration of typed edges and category neighbours.
Similar tools
Same-category neighbours not already linked as typed edges.
Evidence and technical details
Sourced facts, taxonomy, compatibility claims, README excerpt, and machine-readable endpoints.
Overview
The repository fla-org/flash-linear-attention provides efficient implementations of linear attention mechanisms, focusing on large language models and sequence modeling within machine learning systems.
Capability facts
- Languages
- python
Source: github.language+pyproject.toml · Aug 17, 2026
Categories
Compatibility
Sourced claims from the README excerpt - not unsourced marketing copy.
Source: README excerpt (regex_v1, Aug 17, 2026)
$ python -m benchmarks.ops.run --op chunk_retention chunk_gla chunk_gdn flash_attnSource link
Tags
README
Installation
torch lives in a backend extra ([cuda] / [rocm] / [xpu] / [npu] / [cpu]). CUDA is one command; other backends are two so torch (and the right triton flavor that torch pulls transitively) come from the PyTorch wheel index instead of PyPI:
---
# you might have to first install `fla` via `pip install -e .` to enable its import
$ python -m benchmarks.ops.run --op chunk_retention chunk_gla chunk_gdn flash_attn
=================================================================================
Machine: NVIDIA GB200 | CUDA 12.9 | PyTorch 2.9.0+cu129.msh
=================================================================================
fwd B T H D op main[0a484709](ms)
-----------------------------------------------------------------------
1 8192 96 128 chunk_retention 0.787
chunk_gla 1.765
chunk_gdn 1.265
flash_attn 3.753
-----------------------------------------------------------------------
2 16384 16 128 chunk_retention 0.792
chunk_gla 1.445
chunk_gdn 1.029
flash_attn 5.035
-----------------------------------------------------------------------
4 2048 16 128 chunk_retention 0.559
chunk_gla 0.514
chunk_gdn 0.753
flash_attn 0.346
-----------------------------------------------------------------------
4 4096 64 128 chunk_retention 0.997
chunk_gla 2.251
chunk_gdn 1.581
flash_attn 2.560
-----------------------------------------------------------------------
8 1024 8 64 chunk_retention 0.425
chunk_gla 0.358
chunk_gdn 0.631
flash_attn 0.157
-----------------------------------------------------------------------
8 2048 32 256 chunk_retention 1.174
chunk_gla 2.897
chunk_gdn 1.831
flash_attn 1.408
=================================================================================
fwdbwd B T H D op main[0a484709](ms)
-----------------------------------------------------------------------
1 8192 96 128 chunk_retention 2.618
chunk_gla 7.670
chunk_gdn 4.738
flash_attn 15.371
-----------------------------------------------------------------------
2 16384 16 128 chunk_retention 2.122
chunk_
For agents
This page has a .md twin and JSON over the API.