Skip to content

Add input token limit to LLM #1077

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 4 commits into from
Nov 27, 2025
+18 −2
Merged

Add input token limit to LLM #1077

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
69cb8b9
added input token limit to pipeline backend config
farook-edev Nov 17, 2025
330bd44
Merge branch 'submission-v6.0' into farook/input-token-limit
farook-edev Nov 18, 2025
7875f2a
changed input limit to be part of the dataset
farook-edev Nov 24, 2025
3b39a34
formatting
farook-edev Nov 25, 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
7 changes: 7 additions & 0 deletions flutter/cpp/datasets/ifeval.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -32,6 +32,13 @@ IFEval::IFEval(Backend* backend, const std::string& input_tfrecord,
std::vector<int> input_tokens;
sp_processor->Encode(input_formatted.c_str(), &input_tokens).ok();

// input token sanity check
if (input_tokens.size() > input_token_limit_) {
LOG(WARNING) << "Input token limit exceeded for entry "
<< std::to_string(i) << ". Ignoring.";
continue;
}

auto sample = std::make_unique<ifeval::Sample>();
sample->key = key;
sample->prompt = prompt;
Expand Down
1 change: 1 addition & 0 deletions flutter/cpp/datasets/ifeval.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -81,6 +81,7 @@ class IFEval : public Dataset {
std::unordered_set<size_t> used_sample_ids_;
std::set<int> loaded_sample_ids_;
std::unique_ptr<sentencepiece::SentencePieceProcessor> sp_processor;
static constexpr int input_token_limit_ = 1024;
static constexpr int token_limit_ = 1024;
};

Expand Down
7 changes: 7 additions & 0 deletions flutter/cpp/datasets/mmlu_gen.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -45,6 +45,13 @@ MmluGen::MmluGen(Backend* backend, const std::string& input_tfrecord,
std::vector<int> input_tokens;
sp_processor->Encode(input.c_str(), &input_tokens).ok();

// input token sanity check
if (input_tokens.size() > input_token_limit_) {
LOG(WARNING) << "Input token limit exceeded for entry "
<< std::to_string(i) << ". Ignoring.";
continue;
}

auto sample = std::make_unique<PromptSample>();
sample->input = input;
sample->input_tokens = input_tokens;
Expand Down
1 change: 1 addition & 0 deletions flutter/cpp/datasets/mmlu_gen.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -66,6 +66,7 @@ class MmluGen : public Dataset {
std::unordered_set<size_t> used_sample_ids_;
std::set<int> loaded_sample_ids_;
std::unique_ptr<sentencepiece::SentencePieceProcessor> sp_processor;
static constexpr int input_token_limit_ = 1024;
static constexpr int token_limit_ = 4;
};

Expand Down
4 changes: 2 additions & 2 deletions mobile_back_tflite/cpp/backend_tflite/llm_pipeline.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -151,8 +151,8 @@ struct LLMBackendData {
kv_cache_t kv_cache;
std::vector<int> prompt_tokens;
std::vector<int> output_tokens;
uint8_t threads = 2;
int max_output_tokens = 1024;
uint8_t threads = 8;
int max_output_tokens = 128;
std::unordered_set<int> stop_token_ids{128001, 128008, 128009};

LLMBackendData() {}
Expand Down
Loading