mirror of
https://github.com/NVIDIA/TensorRT-LLM.git
synced 2026-01-14 06:27:45 +08:00
84 lines
2.2 KiB
C++
84 lines
2.2 KiB
C++
/*
|
|
* Copyright (c) 2022-2023, NVIDIA CORPORATION. All rights reserved.
|
|
*
|
|
* Licensed under the Apache License, Version 2.0 (the "License");
|
|
* you may not use this file except in compliance with the License.
|
|
* You may obtain a copy of the License at
|
|
*
|
|
* http://www.apache.org/licenses/LICENSE-2.0
|
|
*
|
|
* Unless required by applicable law or agreed to in writing, software
|
|
* distributed under the License is distributed on an "AS IS" BASIS,
|
|
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
* See the License for the specific language governing permissions and
|
|
* limitations under the License.
|
|
*/
|
|
|
|
#pragma once
|
|
|
|
#include "tensorrt_llm/common/quantization.h"
|
|
#include "tensorrt_llm/runtime/common.h"
|
|
#include "tensorrt_llm/runtime/gptModelConfig.h"
|
|
#include "tensorrt_llm/runtime/worldConfig.h"
|
|
|
|
#include <filesystem>
|
|
#include <istream>
|
|
#include <string>
|
|
#include <utility>
|
|
|
|
namespace tensorrt_llm::runtime
|
|
{
|
|
|
|
class GptJsonConfig
|
|
{
|
|
public:
|
|
GptJsonConfig(std::string name, std::string precision, SizeType worldSize, GptModelConfig const& modelConfig)
|
|
: mName(std::move(name))
|
|
, mPrecision(std::move(precision))
|
|
, mWorldSize{worldSize}
|
|
, mGptModelConfig(modelConfig)
|
|
{
|
|
}
|
|
|
|
static GptJsonConfig parse(std::string const& json);
|
|
|
|
static GptJsonConfig parse(std::istream& json);
|
|
|
|
static GptJsonConfig parse(std::filesystem::path const& path);
|
|
|
|
[[nodiscard]] GptModelConfig getModelConfig() const
|
|
{
|
|
return mGptModelConfig;
|
|
}
|
|
|
|
[[nodiscard]] std::string const& getName() const
|
|
{
|
|
return mName;
|
|
}
|
|
|
|
[[nodiscard]] std::string const& getPrecision() const
|
|
{
|
|
return mPrecision;
|
|
}
|
|
|
|
[[nodiscard]] SizeType const& getWorldSize() const
|
|
{
|
|
return mWorldSize;
|
|
}
|
|
|
|
[[nodiscard]] std::string engineFilename(WorldConfig const& worldConfig, std::string const& model) const;
|
|
|
|
[[nodiscard]] std::string engineFilename(WorldConfig const& worldConfig) const
|
|
{
|
|
return engineFilename(worldConfig, getName());
|
|
}
|
|
|
|
private:
|
|
std::string const mName;
|
|
std::string const mPrecision;
|
|
SizeType const mWorldSize;
|
|
GptModelConfig const mGptModelConfig;
|
|
};
|
|
|
|
} // namespace tensorrt_llm::runtime
|