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

这个需要多大得显存可以跑起来RTX4090 24G可以吗 #11

Open
2 tasks
anstonjie opened this issue Aug 17, 2024 · 9 comments
Open
2 tasks

这个需要多大得显存可以跑起来RTX4090 24G可以吗 #11

anstonjie opened this issue Aug 17, 2024 · 9 comments

Comments

@anstonjie
Copy link

System Info / 系統信息

这个需要多大得显存可以跑起来RTX4090 24G可以吗

Who can help? / 谁可以帮助到您?

No response

Information / 问题信息

  • The official example scripts / 官方的示例脚本
  • My own modified scripts / 我自己修改的脚本和任务

Reproduction / 复现过程

这个需要多大得显存可以跑起来RTX4090 24G可以吗

Expected behavior / 期待表现

这个需要多大得显存可以跑起来RTX4090 24G可以吗

@bys0318
Copy link
Member

bys0318 commented Aug 17, 2024

你好,我这边用H800是可以在24g显存之内完成1w字生成的

@bys0318
Copy link
Member

bys0318 commented Aug 17, 2024

显存占用大概在20g左右

@anstonjie
Copy link
Author

anstonjie commented Aug 18, 2024 via email

@anstonjie
Copy link
Author

anstonjie commented Aug 18, 2024 via email

@bys0318
Copy link
Member

bys0318 commented Aug 18, 2024

请试试在每次生成后加一下这行代码,释放未使用的显存:

torch.cuda.empty_cache()

@anstonjie
Copy link
Author

anstonjie commented Aug 18, 2024 via email

@allenxml
Copy link

在ollama中导入q4的gguf格式模型,在openwebui中提问,输出速度很慢,ollama主机4060ti显存你8G,显卡核心频率经常在210,很少到最大频率,7950x的CPU占用率50%。

@rrgkGitHub
Copy link

用bitsandbytes 4bit量化加载模型后,12G显存可以运行。 速度中等偏慢,但可以接受。

@mrpen2
Copy link

mrpen2 commented Aug 28, 2024

用vllm加速的话需要多少显存?

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

5 participants