FlashInfer-Bench is a benchmark suite and production workflow designed to build a virtuous cycle of self-improving AI systems.
It is part of a broader initiative to build the virtuous cycle of AI improving AI systems — enabling AI agents and engineers to collaboratively optimize the very kernels that power large language models.
Install FlashInfer-Bench with pip:
pip install flashinfer-benchImport FlashInfer-Bench:
import flashinfer_bench as fib
print(fib.__version__)This guide shows you how to use FlashInfer-Bench python module with the FlashInfer-Trace dataset.
We provide an official dataset called FlashInfer-Trace with kernels and workloads in real-world AI system deployment environments. FlashInfer-Bench can use this dataset to measure and compare the performance of kernels. It follows the FlashInfer Trace Schema.
The official dataset is on HuggingFace: https://huggingface.co/datasets/flashinfer-ai/flashinfer-trace
Our collaborators include:
