AMDResearch/NPUEval

NPUEval is an LLM evaluation dataset written specifically to target AIE kernel code generation on RyzenAI hardware.

33
/ 100
Emerging

This is a dataset designed to evaluate how well Large Language Models (LLMs) can generate specialized code for AMD's RyzenAI hardware's Neural Processing Unit (NPU). It provides a set of prompts as input and allows you to test the generated code's functional correctness on AIE kernel architecture. AI software developers and researchers working with AMD's RyzenAI platform would use this to benchmark and improve LLM code generation for NPU applications.

Use this if you are developing or evaluating LLMs that generate low-level code for AMD's AI Engine (AIE) kernels on RyzenAI hardware and need a standardized way to measure their performance.

Not ideal if you are working with NPU hardware other than AMD's AIE2/AIE2P or if your focus is on high-level application development rather than kernel-level code generation.

AI Engine development NPU code generation LLM evaluation Hardware acceleration RyzenAI programming
No License No Package No Dependents
Maintenance 6 / 25
Adoption 7 / 25
Maturity 8 / 25
Community 12 / 25

How are scores calculated?

Stars

30

Forks

4

Language

C++

License

Last pushed

Nov 08, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/llm-tools/AMDResearch/NPUEval"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.