Skip to content

fix: llama_memory_seq_rm(mem, -1, ...) #15200

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
wants to merge 1 commit into from
Closed

fix: llama_memory_seq_rm(mem, -1, ...) #15200

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
2 changes: 1 addition & 1 deletion examples/lookahead/lookahead.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -429,7 +429,7 @@ int main(int argc, char ** argv) {

// KV cache management
// if no verification token matched, we simply remove all cells from this batch -> no fragmentation
llama_memory_seq_rm(mem, -1, n_past, -1);
llama_memory_seq_rm(mem, 0, n_past, -1);
Copy link

Copilot AI Aug 9, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The fix changes the sequence ID from -1 to 0, but this may alter the intended behavior. Please verify that sequence ID 0 is the correct target sequence for removal, as this could affect which cached tokens are cleared in the lookahead algorithm.

Suggested change
llama_memory_seq_rm(mem, 0, n_past, -1);
llama_memory_seq_rm(mem, -1, n_past, -1);

Copilot uses AI. Check for mistakes.

if (seq_id_best != 0) {
// if a verification token matched, we keep the best sequence and remove the rest
Expand Down
2 changes: 1 addition & 1 deletion tools/main/main.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -354,7 +354,7 @@ int main(int argc, char ** argv) {
}

// remove any "future" tokens that we might have inherited from the previous session
llama_memory_seq_rm(mem, -1, n_matching_session_tokens, -1);
llama_memory_seq_rm(mem, 0, n_matching_session_tokens, -1);
Copy link

Copilot AI Aug 9, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The fix changes the sequence ID from -1 to 0, but this may alter the intended behavior. Please verify that sequence ID 0 is the correct target sequence for removal, as this could affect which cached tokens are cleared.

Suggested change
llama_memory_seq_rm(mem, 0, n_matching_session_tokens, -1);
llama_memory_seq_rm(mem, -1, n_matching_session_tokens, -1);

Copilot uses AI. Check for mistakes.
}

LOG_DBG("recalculate the cached logits (check): embd_inp.size() %zu, n_matching_session_tokens %zu, embd_inp.size() %zu, session_tokens.size() %zu\n",
Expand Down
Loading