Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

evalscope perf 测试sglang 部署的openai api server 无法输出结果 #128

Open
hetian127 opened this issue Sep 12, 2024 · 1 comment
Assignees

Comments

@hetian127
Copy link

hetian127 commented Sep 12, 2024

版本:evalscope 0.5.3
sglang 0.3.0

在本地起了一个sglang的openai api server,命令如下:
CUDA_VISIBLE_DEVICES=4,5,6,7 python -m sglang.launch_server --model-path /local/models/Qwen2-72B-Instruct --tp 4 --host 0.0.0.0
能够正常访问,用下面的语句进行benchmark测试:
evalscope perf --url 'http://localhost:30000/v1/chat/completions' --parallel 128 --model '/local/models/Qwen2-72B-Instruct' --log-every-n-query 10 --read-timeout=120 --dataset-path './dataset/open_qa.jsonl' -n 50 --max-prompt-length 128000 --api openai --stream --stop '<|im_end|>' --dataset openqa --debug

server端有输出:
image

测试脚本到这个位置就不动了
image

同样的命令在测试vllm时可以正常输出结果,但是测试sglang时无法生成结果。
麻烦帮忙看看问题出在哪里,多谢

@Yunnglin
Copy link
Collaborator

我们后面复现一下

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants