flagos-ai/FlagPerf

FlagPerf is an open-source software platform for benchmarking AI chips.

Stars: 379
Forks: 124
Open beginner issues: 0
Indexed issues: 7

Dominant language: Python
License: Apache-2.0
Last GitHub push: Jul 18, 2026
Latest indexed: Sep 16, 2026
Contributing guide: No contributing guide
Code of conduct: No code of conduct
Beginner labels: No beginner labels indexed
PR merge metrics: No merged PRs in 30d

Beginner friendly Everything

Open All

7 open issues indexed

单机BENCHMARK 测试基础规格的inferconnect相关用例时会报错 Open

Difficulty 3/5 1-2 days Newbie friendliness 35/100

flagos-ai/FlagPerf#797 · Opened May 15, 2025
llama2-7b评测时，/usr/local/nvidia是镜像中的文件吗？本地/usr/local/下不存在nvidia目录，rank0.out.log提示如下 Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

flagos-ai/FlagPerf#783 · Opened Feb 10, 2025
评测llama2_7b_finetune时rank0报错，是cuda版本的问题还是pytorch版本的问题 Open

Difficulty 3/5 1-2 days Newbie friendliness 45/100

flagos-ai/FlagPerf#782 · Opened Feb 7, 2025
在使用多卡训练评测时会生成rank0.out.log、rank1.out.log。。。。rankn.out.log，我们应该以哪一个为标准 Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

flagos-ai/FlagPerf#777 · Opened Jan 9, 2025
llama3_8b_mmlu:pytorch_2.1 推理测试失败 Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

flagos-ai/FlagPerf#766 · Opened Nov 15, 2024
ascend训练只支持aquila2吗 Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

flagos-ai/FlagPerf#756 · Opened Sep 26, 2024
stable_diffusion_v1_4使用torchtrt推理时报错 Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

flagos-ai/FlagPerf#754 · 2 comments · Opened Sep 24, 2024

A short digest of beginner-friendly GitHub issues.