deepseek-ai/flashmla · Architecture
A static read of how deepseek-ai/flashmla is wired together — module coupling, layering, and circular dependencies — derived from its import graph.
A
Architecture grade
score 100/100
0
Import cycles
lower is better
18
Modules parsed
12 edges
100%
Layeredness
clean-layer ratio
Most-depended-on modules
These are the highest-fan-in files — change them and the blast radius is largest. Read them first.
- flash_mla/__init__.py5 importers · radius 5
- tests/kernelkit/utils.py2 importers · radius 2
- tests/kernelkit/bench.py1 importers · radius 1
- flash_mla/flash_mla_interface.py1 importers · radius 6
- tests/kernelkit/compare.py1 importers · radius 1
- tests/kernelkit/generate.py1 importers · radius 1
- tests/kernelkit/precision.py1 importers · radius 1
- tests/kernelkit/__init__.py0 importers · radius 0
Modules (3)
Source directories grouped from the import graph. Each links to a per-module page with its files and how it couples to the rest.
- tests/kernelkit6 files · 6 in-degree
- flash_mla2 files · 6 in-degree
- tests4 files · 0 in-degree
Ask AI about deepseek-ai/flashmla
Grounded in the actual source code. Pick a starter question or write your own.
What does this repo do, in one paragraph?How would I get started using it?What are the main alternatives?Show me the entry point.
Or write your own questionInformational only. RepoPilot summarises public signals at the time of analysis; they can be incomplete or stale. Not professional, security, or legal advice.