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

输入长度怎么设置 #4

Open
wzg-zhuo opened this issue May 9, 2024 · 5 comments
Open

输入长度怎么设置 #4

wzg-zhuo opened this issue May 9, 2024 · 5 comments

Comments

@wzg-zhuo
Copy link

wzg-zhuo commented May 9, 2024

GGML_ASSERT: /tmp/pip-req-build-obcizsli/third_party/ggml/src/ggml.c:2493: view_src == NULL || data_size + view_offs <= ggml_nbytes(view_src)
Could not attach to process. If your uid matches the uid of the target
process, check the setting of /proc/sys/kernel/yama/ptrace_scope, or try
again as the root user. For more details, see /etc/sysctl.d/10-ptrace.conf
ptrace: 对设备不适当的 ioctl 操作.
No stack.
The program is not being run.

max_length设置1024,推理报错,请问是转化时候要设置么

@yvonwin
Copy link
Owner

yvonwin commented May 9, 2024

抱歉,我无法复现这个错误,但是这个问题应该和转换无关,需要更多的信息才能确认。

@wzg-zhuo
Copy link
Author

wzg-zhuo commented May 9, 2024

max_length设置512不会报错,过长会报错,可以试试

@yvonwin
Copy link
Owner

yvonwin commented May 9, 2024

max_length 和 分配的 kv_cache 大小有关,如果你用的是gpu, 有可能模型太大,显存不够?

@wzg-zhuo
Copy link
Author

wzg-zhuo commented May 9, 2024

使用的是cpu,8c,模型使用的是qwen1.5-1.8-chat

@yvonwin
Copy link
Owner

yvonwin commented May 9, 2024

那就有点奇怪了,我这边即使是32b的模型也不存在该问题。我在m1, 13600k, 和一个epyc上测试过。

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

2 participants