Skip to content

server : fix kv cache management #3588

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

Merged
ggerganov merged 1 commit into from
Oct 12, 2023
Merged

server : fix kv cache management #3588

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
10 changes: 7 additions & 3 deletions examples/server/server.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -405,13 +405,17 @@ struct llama_server_context
// compare the evaluated prompt with the new prompt
n_past = common_part(embd, prompt_tokens);
embd = prompt_tokens;

if (n_past == num_prompt_tokens)
{
// we have to evaluate at least 1 token to generate logits.
printf("we have to evaluate at least 1 token to generate logits\n");
n_past--;
}

// since #3228 we now have to manually manage the KV cache
llama_kv_cache_seq_rm(ctx, 0, n_past, -1);

LOG_VERBOSE("prompt ingested", {
{"n_past", n_past},
{"cached", tokens_to_str(ctx, embd.cbegin(), embd.cbegin() + n_past)},
Expand Down Expand Up @@ -461,16 +465,16 @@ struct llama_server_context
// compare the evaluated prompt with the new prompt
n_past = common_part(embd, prompt_tokens);

// since #3228 we now have to manually manage the KV cache
llama_kv_cache_seq_rm(ctx, 0, n_past, -1);

embd = prompt_tokens;
if (n_past == num_prompt_tokens)
{
// we have to evaluate at least 1 token to generate logits.
n_past--;
}

// since #3228 we now have to manually manage the KV cache
llama_kv_cache_seq_rm(ctx, 0, n_past, -1);

LOG_VERBOSE("prompt ingested", {
{"n_past", n_past},
{"cached", tokens_to_str(ctx, embd.cbegin(), embd.cbegin() + n_past)},
Expand Down