源码方式安装llama.cpp及调试

llama.cpp源码方式安装和调试配置

构建和编译

• 注意这里是cuda，且要开启debug模式

cmake -B build -DGGML_CUDA=ON -DCMAKE_BUILD_TYPE=Debug
cmake --build build --config Debug

• 正在编译：
  

配置launch.json用于调式：

要根据自己的环境路径做相应修改

{
    "version": "0.2.0",
    "configurations": [
        {
            "name": "(gdb) 启动",
            "type": "cppdbg",
            "request": "launch",
            "program": "${workspaceFolder}/build/bin/llama-simple", // 
            "args": [    // 
                "-m", "output.gguf",
                "-n", "32",
                "-ngl", "99",
                "Hello my name is" 
            ],
            "stopAtEntry": false,
            "cwd": "${workspaceFolder}",
            "environment": [],
            "externalConsole": false,
            "MIMode": "gdb", // 
            "setupCommands": [
                {
                    "description": "为 gdb 启用整齐打印",
                    "text": "-enable-pretty-printing",
                    "ignoreFailures": true
                },
                {
                    "description": "将反汇编风格设置为 Intel",
                    "text": "-gdb-set disassembly-flavor intel",
                    "ignoreFailures": true
                }
            ],
            "miDebuggerPath": "/usr/bin/gdb" // 
        }

    ]
}

转换模型为gguf格式

python convert_hf_to_gguf.py --outtype f16 --outfile "output.gguf" "/raid/home/huafeng/models/Meta-Llama-3-8B-Instruct"

运行第一个程序

调试程序（llama.cpp/examples/simple/simple.cpp）