Skip to content

Set default batch size to 16 #267

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
anasdorbani merged 3 commits into from
Jun 3, 2026
Merged

Set default batch size to 16 #267

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
75a25b9
Set default batch size to 16
anasdorbani Jun 3, 2026
b4dbd76
Add default batch size coverage
anasdorbani Jun 3, 2026
0f15219
Fix aggregate metrics flake
anasdorbani Jun 3, 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
5 changes: 1 addition & 4 deletions src/functions/aggregate/llm_rerank/implementation.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -109,10 +109,7 @@ nlohmann::json LlmRerank::SlidingWindow(nlohmann::json& tuples) {
auto carry_forward_tuples = nlohmann::json::array();
int start_index = 0;

auto batch_size = static_cast<int>(model.GetModelDetails().batch_size);
if (batch_size == 2048) {
batch_size = std::min<int>(batch_size, num_tuples);
}
auto batch_size = std::min<int>(model.GetModelDetails().batch_size, num_tuples);

@queryproc queryproc Jun 3, 2026 •
edited
Loading

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

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

We should make sure that we take the min of 2: (user_batch, default_batch = 16). Using this batch size, we then expand the data section of the prompt while not going beyond the max input context window size.

@queryproc queryproc Jun 3, 2026

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

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

Alternatively, if the user provided a batch_size, it overrides the default size and then we expand till we reach the max input context window size.


if (batch_size <= 0) {
throw std::runtime_error("Batch size must be greater than zero");
Expand Down
2 changes: 2 additions & 0 deletions src/include/flock/model_manager/repository.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,6 +8,8 @@

namespace flock {

inline constexpr int DEFAULT_BATCH_SIZE = 16;

struct ModelDetails {
std::string provider_name;
std::string model_name;
Expand Down
2 changes: 1 addition & 1 deletion src/model_manager/model.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -87,7 +87,7 @@ void Model::LoadModelDetails(const nlohmann::json& model_json) {
} else if (db_model_args.contains("batch_size")) {
model_details_.batch_size = db_model_args.at("batch_size").get<int>();
} else {
model_details_.batch_size = 2048;
model_details_.batch_size = DEFAULT_BATCH_SIZE;
}
}
}
Expand Down
3 changes: 2 additions & 1 deletion test/unit/model_manager/model_manager_test.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -63,6 +63,7 @@ TEST_F(ModelManagerTest, ModelInitializationMinimal) {
EXPECT_EQ(details.model_name, "gpt-4o-test");
EXPECT_EQ(details.model, "gpt-4o");
EXPECT_EQ(details.provider_name, "openai");
EXPECT_EQ(details.batch_size, DEFAULT_BATCH_SIZE);
});
}

Expand Down Expand Up @@ -127,4 +128,4 @@ TEST_F(ModelManagerTest, GetModelDetails) {
EXPECT_EQ(details.batch_size, 10);
}

}// namespace flock
}// namespace flock
Loading