-
Notifications
You must be signed in to change notification settings - Fork 936
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
42221 segmentation fault (core dumped) ./mpt #404
Comments
It returns gibberish after a few tokens and then crashes |
Try setting the --ctx-size parameter to 1024. It must be higher than the -n parameter. |
|
Do you still get segmentation fault? Please paste the whole output including the command you are using. |
Yep, something is off. With
I see
without timing information (event viewer shows crash) and with
everything seems fine
So probably a problem if the number of tokens to predict is near the context size. |
Running in a debugger: crashes in
due to |
Interesting.
crashes in
with |
7 prompt tokens + 16 predicted > 16 n_ctx I think we need to cut down the value of the -n parameter if too high so the tokens wont overflow the ctx. |
Replacing
with
seems to work for me. I checked with a prompt longer than the context size, but didn't consider testing batch size larger than context size. |
Isn't mpt storywriter meant to be used with large contexts? Is that not possible with ggml? |
It is possible AFAIK. But there exist some restrictions regarding the parameters which maybe are not thoroughly enforced everywhere. Therefore a more complete description of how you are calling |
I'm trying to pass the first paragraph of a book |
Yes it is possible to use --ctx-size up to 64k. But be aware that token evaluation time is increasing for each new token predicted. |
Maybe better to restrict n_predict instead. Something like:
|
My issue is that it generates nonsense:
|
only happens when giving it long context. it can generate that amount just fine without becoming nonsense |
But considering the core dumped issue has been fixed by --ctx-size, ill close this issue |
This line of the output tells you how many tokens there is in the prompt: the -n parameter sets how many tokens to predict = N now try setting --ctx-size to something higher than X + N |
The text was updated successfully, but these errors were encountered: