happyz
/
gemma.cpp
mirror of https://github.com/google/gemma.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Add entropy expectations for Griffin-2b model in gemma_test and make sure it passes. 

PiperOrigin-RevId: 675564389
This commit is contained in:
Daniel Keysers 2024-09-17 07:45:22 -07:00 committed by Copybara-Service
parent e4ba93412a
commit 760a69449e
1 changed files with 19 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

26
evals/gemma_test.cc
View File

@ -29,7 +29,8 @@
// To run the test, pass the following flags:
// --model <model> --tokenizer <tokenizer_path> --weights <weights_path>
// It should pass for the following models:
// 2b-it (v1 and v1.1), 7b-it (v1 and v1.1), gemma2-2b-it, 9b-it, 27b-it
// Gemma1: 2b-it (v1 and v1.1), 7b-it (v1 and v1.1), gr2b-it,
// Gemma2: gemma2-2b-it, 9b-it, 27b-it,
namespace gcpp {
namespace {
@ -46,14 +47,15 @@ class GemmaTest : public ::testing::Test {
s_env->MutableConfig().temperature = 0.0f; // deterministic
s_env->MutableConfig().verbosity = 0;
// Using the turn structure worsens results sometimes.
// However, gemma-2 27B seems to need the turn structure to work.
// However, some models need the turn structure to work.
// It would be good to make these tests more consistent.
if (s_env->GetModel()->Info().model == Model::GEMMA2_27B) {
if (s_env->GetModel()->Info().model == Model::GEMMA2_27B ||
s_env->GetModel()->Info().model == Model::GRIFFIN_2B) {
std::string mutable_prompt = prompt;
auto [response, n] = s_env->QueryModel(mutable_prompt); // Uses turns.
return response;
}
// Otherwise, don't use turn structure.
// Otherwise, do not use turn structure.
const std::vector<int> tokens = s_env->TokenizeAndPrependBOS(prompt);
auto [response, n] = s_env->QueryModel(tokens);
return response;
@ -66,15 +68,16 @@ class GemmaTest : public ::testing::Test {
s_env->MutableConfig().verbosity = 0;
std::vector<std::string> replies;
// Using the turn structure worsens results sometimes.
// However, gemma-2 27B seems to need the turn structure to work.
// However, some models need the turn structure to work.
// It would be good to make these tests more consistent.
if (s_env->GetModel()->Info().model == Model::GEMMA2_27B) {
if (s_env->GetModel()->Info().model == Model::GEMMA2_27B ||
s_env->GetModel()->Info().model == Model::GRIFFIN_2B) {
for (auto [response, n] : s_env->BatchQueryModel(inputs)) {
replies.push_back(response);
}
return replies;
}
// Not Gemma-2 27B. Do not use turn structure.
// Otherwise, do not use turn structure.
std::vector<std::vector<int>> prompts_vector;
prompts_vector.reserve(inputs.size());
for (const auto& input_string : inputs) {
@ -243,6 +246,9 @@ TEST_F(GemmaTest, CrossEntropySmall) {
// 7B v.1 and v.1.1 produce slightly different results.
EXPECT_NEAR(entropy, 2.8f, 0.2f);
break;
case gcpp::Model::GRIFFIN_2B:
EXPECT_NEAR(entropy, 2.25f, 0.02f);
break;
case gcpp::Model::GEMMA2_2B:
EXPECT_NEAR(entropy, 1.14f, 0.02f);
break;
@ -271,6 +277,9 @@ TEST_F(GemmaTest, CrossEntropyJingleBells) {
// 7B v.1 and v.1.1 produce slightly different results.
EXPECT_NEAR(entropy, 1.07f, 0.05f);
break;
case gcpp::Model::GRIFFIN_2B:
EXPECT_NEAR(entropy, 1.95f, 0.02f);
break;
case gcpp::Model::GEMMA2_2B:
EXPECT_NEAR(entropy, 0.49f, 0.02f);
break;
@ -299,6 +308,9 @@ TEST_F(GemmaTest, CrossEntropyGettysburg) {
// 7B v.1 and v.1.1 produce slightly different results.
EXPECT_NEAR(entropy, 0.75f, 0.1f);
break;
case gcpp::Model::GRIFFIN_2B:
EXPECT_NEAR(entropy, 0.82f, 0.02f);
break;
case gcpp::Model::GEMMA2_2B:
EXPECT_NEAR(entropy, 0.20f, 0.02f);
break;