Skip to content

server : disable speculative decoding for SWA models #13970

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
merged 2 commits into from
Jun 2, 2025
Merged

server : disable speculative decoding for SWA models #13970

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
b4f7dcf
server : use swa-full fo draft context
ggerganov Jun 2, 2025
272df3f
server : disable speculative decoding for SWA models
ggerganov Jun 2, 2025
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
8 changes: 5 additions & 3 deletions tools/server/server.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2016,6 +2016,11 @@ struct server_context {
params_base.n_cache_reuse = 0;
SRV_WRN("%s\n", "cache_reuse is not supported by this context, it will be disabled");
}

if (!params_base.speculative.model.path.empty()) {
SRV_ERR("%s\n", "err: speculative decode is not supported by this context");
return false;
}
}

return true;
Expand Down Expand Up @@ -3203,9 +3208,7 @@ struct server_context {
}
} else {
// if we don't cache the prompt, we have to remove the entire KV cache
llama_kv_self_seq_rm(ctx, slot.id, 0, -1);
slot.n_past = 0;
slot.cache_tokens.clear(); // TODO: not needed, will be cleared later via "keep_first()"
}

if (slot.n_past > 0 && slot.n_past < (int) slot.cache_tokens.size()) {
Expand All @@ -3220,7 +3223,6 @@ struct server_context {
SLT_WRN(slot, "n_past = %d, cache_tokens.size() = %d, seq_id = %d, pos_min = %d, n_swa = %d\n", slot.n_past, (int) slot.cache_tokens.size(), slot.id, pos_min, n_swa);
SLT_WRN(slot, "forcing full prompt re-processing due to lack of cache data (likely due to SWA, see %s)\n",
"https://github.com/ggml-org/llama.cpp/pull/13194#issuecomment-2868343055");
llama_kv_self_seq_rm(ctx, slot.id, 0, -1);
slot.n_past = 0;
}
}
Expand Down
Loading