Skip to content
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

Feat/unload model api #97

Closed
wants to merge 5 commits into from
Closed

Feat/unload model api #97

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
a4b7226
Add api to unload model
thunhu99 Oct 31, 2023
3d81378
set context and model back to nullptr after delete
thunhu99 Oct 31, 2023
b617433
Merge branch 'janhq:main' into feat/unload-model-api
thunhuanh Nov 2, 2023
af3bdb9
fix wrong api endpoind - handler mapping
thunhu99 Nov 2, 2023
29411b5
Merge branch 'feat/unload-model-api' of https://github.com/thunhuanh/…
thunhu99 Nov 2, 2023
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
13 changes: 13 additions & 0 deletions controllers/llamaCPP.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -344,3 +344,16 @@ void llamaCPP::loadModel(
warmupModel();
callback(resp);
}

void inferences::llamaCPP::unloadModel(const HttpRequestPtr &req, std::function<void(const HttpResponsePtr &)> &&callback)
{
Json::Value jsonResp;
if (model_loaded) {
llama.unloadModel();
Copy link
Contributor

Choose a reason for hiding this comment

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

As I tested, the server stops working after this line.
The below lines do not execute to return result

Copy link
Author

Choose a reason for hiding this comment

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

Oh, sorry my mistake. The API endpoint map to the wrong handler:
it should be METHOD_ADD(llamaCPP::unloadModel, "unloadmodel", Delete);
instead of METHOD_ADD(llamaCPP::loadModel, "unloadmodel", Delete);

model_loaded = false;
jsonResp["message"] = "Model unloaded successfully";
}
jsonResp["message"] = "No model loaded";
auto resp = nitro_utils::nitroHttpJsonResponse(jsonResp);
callback(resp);
}
11 changes: 11 additions & 0 deletions controllers/llamaCPP.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -260,6 +260,13 @@ struct llama_server_context {
return true;
}

void unloadModel() {
llama_free(ctx);
llama_free_model(model);
ctx = nullptr;
model = nullptr;
}

std::vector<llama_token> tokenize(const json &json_prompt,
bool add_bos) const {
// If `add_bos` is true, we only add BOS, when json_prompt is a string,
Expand Down Expand Up @@ -1272,6 +1279,7 @@ class llamaCPP : public drogon::HttpController<llamaCPP> {
METHOD_ADD(llamaCPP::chatCompletion, "chat_completion", Post);
METHOD_ADD(llamaCPP::embedding, "embedding", Post);
METHOD_ADD(llamaCPP::loadModel, "loadmodel", Post);
METHOD_ADD(llamaCPP::loadModel, "unloadmodel", Delete);
// PATH_ADD("/llama/chat_completion", Post);
METHOD_LIST_END
void chatCompletion(const HttpRequestPtr &req,
Expand All @@ -1282,6 +1290,9 @@ class llamaCPP : public drogon::HttpController<llamaCPP> {
std::function<void(const HttpResponsePtr &)> &&callback);
void warmupModel();

void unloadModel(const HttpRequestPtr &req,
std::function<void(const HttpResponsePtr &)> &&callback);

private:
llama_server_context llama;
bool model_loaded = false;
Expand Down