huggingface/text-generation-inference
5
0
Fork 0
mirror of https://github.com/huggingface/text-generation-inference.git synced 2025-06-10 03:12:07 +00:00
Code Issues Packages Projects Releases Wiki Activity

test(backend): add more unittest

Browse Source
This commit is contained in:
Morgan Funtowicz 2024-12-03 14:39:10 +01:00
parent c94b9de445
commit ad3ed0d1a1
3 changed files with 47 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
backends/trtllm/CMakeLists.txt
View File

@ -39,10 +39,10 @@ if(${CMAKE_BUILD_TYPE} STREQUAL "Debug")
endif()
# This attempt to detect if the compiler can emit warning if it can't apply return value optimization from a function
#check_cxx_compiler_flag("-Wnrvo" COMPILER_SUPPORT_WARNING_ON_NVRO)
#if(${COMPILER_SUPPORT_WARNING_ON_NVRO})
# set(CMAKE_CXX_FLAGS "{CMAKE_CXX_FLAGS} -Wnvro")
#endif()
check_cxx_compiler_flag("-Wnrvo" COMPILER_SUPPORT_WARNING_ON_NVRO)
if(${COMPILER_SUPPORT_WARNING_ON_NVRO})
set(CMAKE_CXX_FLAGS "{CMAKE_CXX_FLAGS} -Wnvro")
endif()
# Let's build TRTLLM as part of CMake
add_subdirectory("${trtllm_SOURCE_DIR}/cpp" "${trtllm_SOURCE_DIR}/..")
@ -74,12 +74,15 @@ if (${TGI_TRTLLM_BACKEND_BUILD_TESTS})
)
FetchContent_MakeAvailable(Catch2)
add_executable(tgi_trtllm_backend_tests tests/test_hardware.cpp)
add_executable(tgi_trtllm_backend_tests tests/test_hardware.cpp tests/test_backend.cpp)
target_include_directories(tgi_trtllm_backend_tests PUBLIC "${trtllm_SOURCE_DIR}/cpp/include")
target_include_directories(tgi_trtllm_backend_tests PUBLIC "csrc/")
target_link_libraries(tgi_trtllm_backend_tests PRIVATE CUDA::cudart CUDA::nvml)
target_link_libraries(tgi_trtllm_backend_tests PRIVATE Catch2::Catch2WithMain tensorrt_llm nlohmann_json::nlohmann_json spdlog::spdlog)
if(CMAKE_BUILD_TYPE MATCHES "Debug")
set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -Werror -fsanitize=undefined -fsanitize=address")
set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Werror -fsanitize=undefined -fsanitize=address")
target_link_options(tgi_trtllm_backend_tests BEFORE PUBLIC -fsanitize=undefined PUBLIC -fsanitize=address)
endif()

6
backends/trtllm/csrc/backend.hpp
View File

@ -69,10 +69,10 @@ namespace huggingface::tgi::backends::trtllm {
constexpr explicit generation_config_t(const json &config):
top_p(config.value("top_p", 1.0f)), temperature( config.value("temperature", 1.0f)), stop_words(0) {
if(config.contains("/eos_token_id"_json) && config["/eos_token_id"_json].is_array()) {
const auto& eos_token_id = config["eos_token_id"];
if(config.contains("/eos_token_id"_json_pointer) && config["/eos_token_id"_json_pointer].is_array()) {
const auto& eos_token_id = config["/eos_token_id"_json_pointer];
std::for_each(eos_token_id.begin(), eos_token_id.end(), [this](const auto token_id) {
stop_words.emplace_back(token_id.template get<int32_t>());
stop_words.emplace_back(1, token_id.template get<int32_t>());
});
SPDLOG_DEBUG("Detected {:d} predefined stop_words from generation_config.json", stop_words.size());

36
backends/trtllm/tests/test_backend.cpp Normal file
View File

@ -0,0 +1,36 @@
//
// Created by mfuntowicz on 12/3/24.
//
#include <catch2/catch_all.hpp>
#include <nlohmann/json.hpp>
#include "../csrc/backend.hpp"
using namespace huggingface::tgi::backends::trtllm;
TEST_CASE("parse generation_config.json", "[generation_config_t]")
{
const json config_j = {{"temperature", 0.6}, {"top_p", 0.95}, {"eos_token_id", {1,2,3}}};
const auto generation_config = generation_config_t(config_j);
REQUIRE_FALSE(generation_config.stop_words.empty());
REQUIRE(generation_config.stop_words.size() == config_j["/eos_token_id"_json_pointer].size());
for (auto [lhs, rhs] : std::views::zip(generation_config.stop_words, std::list<std::vector<int32_t>>{{1}, {2}, {3}}))
{
// Currently we do not support multi-tokens stop words
REQUIRE(lhs.size() == 1);
REQUIRE(rhs.size() == 1);
REQUIRE_THAT(lhs, Catch::Matchers::UnorderedEquals(rhs));
}
}
TEST_CASE("parallel_config", "[backend_workspace_t]")
{
}
TEST_CASE("executor_config", "[backend_workspace_t]")
{
}