Skip to content

Use the full dataset for performance benchmarks #1069

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
farook-edev merged 1 commit into from
Nov 11, 2025
+2 −26
Merged

Use the full dataset for performance benchmarks #1069

Show file tree
Hide file tree
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
2 changes: 0 additions & 2 deletions flutter/cpp/datasets/ifeval.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -77,8 +77,6 @@ std::vector<uint8_t> IFEval::ProcessOutput(const int sample_idx,
sample_output_tokens_[sample_idx] = output_tokens;
used_sample_ids_.insert(sample_idx);

LOG(INFO) << "Processed " << std::to_string(used_sample_ids_.size()) << "/29";

return {1};
}

Expand Down
2 changes: 1 addition & 1 deletion flutter/cpp/datasets/ifeval.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -46,7 +46,7 @@ class IFEval : public Dataset {

size_t TotalSampleCount() override { return samples_.size(); }

size_t PerformanceSampleCount() override { return 1; }
size_t PerformanceSampleCount() override { return samples_.size(); }

void LoadSamplesToRam(const std::vector<QuerySampleIndex>& samples) override;

Expand Down
2 changes: 0 additions & 2 deletions flutter/cpp/datasets/ifeval_utils/common.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -51,8 +51,6 @@ inline bool contains_string(const std::string& text,
inline bool contains_word(const std::string& text, const std::string& word) {
if (word.empty()) return false;

LOG(INFO) << "searching for '" << word << "'...";

auto to_lower_ascii = [](std::string s) {
for (char& c : s) c = std::tolower(static_cast<unsigned char>(c));
return s;
Expand Down
20 changes: 0 additions & 20 deletions flutter/cpp/datasets/mmlu_gen.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,8 +7,6 @@
#include "tensorflow/core/example/example.pb.h"
#include "tensorflow/core/example/feature_util.h"

#define li LOG(INFO) << "li:" << __FILE__ << ":" << __LINE__ << "@" << __func__

namespace mlperf {
namespace mobile {

Expand All @@ -18,8 +16,6 @@ MmluGen::MmluGen(Backend* backend, const std::string& input_tfrecord,
sp_processor = std::unique_ptr<sentencepiece::SentencePieceProcessor>(
LoadSentencePieceProcessor(sp_path));

li;

// Load all TFRecord samples into memory
// NOTE this can be moved to LoadSamplesToRam, but will cause delays between
// queries due to IO reads happening between them
Expand All @@ -41,8 +37,6 @@ MmluGen::MmluGen(Backend* backend, const std::string& input_tfrecord,
std::string preface = input.substr(0, input.find("\n\n") + 2);

input = preface + "Question: " + question_formatted;

LOG(INFO) << input;
}

// std::string input_formatted = FormatLlamaUserPrompt(input, "Provide only
Expand All @@ -59,7 +53,6 @@ MmluGen::MmluGen(Backend* backend, const std::string& input_tfrecord,
samples_.push_back(std::move(sample));
sample_output_tokens_.push_back(std::vector<int>());
}
li;
}

void MmluGen::LoadSamplesToRam(const std::vector<QuerySampleIndex>& samples) {
Expand All @@ -78,8 +71,6 @@ void MmluGen::UnloadSamplesFromRam(
std::vector<void*> MmluGen::GetData(int sample_idx) {
std::vector<void*> data;

li;
LOG(INFO) << "Sample ID: " << std::to_string(sample_idx);
if (sample_idx < samples_.size()) {
data.push_back(reinterpret_cast<void*>(
const_cast<std::vector<int>*>(&(samples_[sample_idx]->input_tokens))));
Expand All @@ -92,18 +83,12 @@ std::vector<uint8_t> MmluGen::ProcessOutput(const int sample_idx,
const std::vector<void*>& outputs) {
if (sample_idx >= samples_.size() || outputs.empty()) return {0};

li;
const auto& output_tokens =
*(reinterpret_cast<std::vector<int>*>(outputs[0]));

li;
sample_output_tokens_[sample_idx] = output_tokens;
used_sample_ids_.insert(sample_idx);

li;
LOG(INFO) << "Processed " << std::to_string(used_sample_ids_.size())
<< "/100";

return {1};
}

Expand All @@ -117,14 +102,9 @@ bool MmluGen::ComputeSampleAccuracy(const int sample_idx) {
std::string prediction;
sp_processor->Decode(sample_output_tokens_[sample_idx], &prediction).ok();

li;
LOG(INFO) << "index: " << std::to_string(sample_idx) << std::endl;
LOG(INFO) << "Output: [[[" << prediction << "]]]" << std::endl;

char predicted_char = find_answer_char(prediction);
const std::string& correct = samples_[sample_idx]->answer;

li;
return (predicted_char == correct[0]);
}

Expand Down
2 changes: 1 addition & 1 deletion flutter/cpp/datasets/mmlu_gen.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,7 +26,7 @@ class MmluGen : public Dataset {

size_t TotalSampleCount() override { return samples_.size(); }

size_t PerformanceSampleCount() override { return 1; }
size_t PerformanceSampleCount() override { return samples_.size(); }

void LoadSamplesToRam(const std::vector<QuerySampleIndex>& samples) override;

Expand Down
Loading