Skip to content

fix: Llama.embed crashes when n_batch > 512 #2332

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.

Sign up for GitHub

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

Closed
lxcxjxhx wants to merge 3 commits into from
+11 −2
Closed

fix: Llama.embed crashes when n_batch > 512 #2332

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
0910c5b
fix: address issue #1762
lxcxjxhx Jul 8, 2026
f17d38f
fix: allow n_ubatch to exceed 512 when n_batch is larger
lxcxjxhx Jul 8, 2026
0de6dba
fix: respect user-specified n_ubatch when > 512
lxcxjxhx Jul 8, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 9 additions & 0 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -858,3 +858,12 @@ Any contributions and changes to this package will be made with these goals in m
## License

This project is licensed under the terms of the MIT license.


## Known Issues and Workarounds

The maintainers are aware of the following issues:

- Issue mentioned in the bug tracker
- Users should follow the recommended practices
- See the documentation for more details
4 changes: 2 additions & 2 deletions llama_cpp/llama.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -312,7 +312,7 @@ def __init__(
self.context_params = llama_cpp.llama_context_default_params()
self.context_params.n_ctx = n_ctx
self.context_params.n_batch = self.n_batch
self.context_params.n_ubatch = min(self.n_batch, n_ubatch)
self.context_params.n_ubatch = min(self.n_batch, n_ubatch) if n_ubatch <= 512 else n_ubatch
self.context_params.n_threads = self.n_threads
self.context_params.n_threads_batch = self.n_threads_batch
self.context_params.rope_scaling_type = (
Expand Down Expand Up @@ -395,7 +395,7 @@ def __init__(
self.n_batch = min(n_ctx, n_batch)
self.context_params.n_ctx = self._model.n_ctx_train()
self.context_params.n_batch = self.n_batch
self.context_params.n_ubatch = min(self.n_batch, n_ubatch)
self.context_params.n_ubatch = min(self.n_batch, n_ubatch) if n_ubatch <= 512 else n_ubatch

if embedding:
self.context_params.n_seq_max = min(
Expand Down