M5Module-LLM/docs/en.md

M5Module-LLM Arduino API

M5Module-LLM Arduino Driver Library API Documentation.

M5ModuleLLM Class

M5ModuleLLM is used to initialize the LLM Module and provides internal members for quick initialization of various LLM units, making it easier to build applications according to your needs.

class M5ModuleLLM {
public:
    bool begin(Stream * targetPort);
    bool checkConnection();
    void update();

    m5_module_llm::ApiSys sys;
    m5_module_llm::ApiLlm llm;
    m5_module_llm::ApiAudio audio;
    m5_module_llm::ApiTts tts;
    m5_module_llm::ApiTts melotts;
    m5_module_llm::ApiKws kws;
    m5_module_llm::ApiAsr asr;
    m5_module_llm::ApiAsr yolo;
    m5_module_llm::ApiVad vad;
    m5_module_llm::ApiWhisper whisper;
    m5_module_llm::ApiDepthAnything depthanything;
    m5_module_llm::ModuleMsg msg;
    m5_module_llm::ModuleComm comm;
private:
};

begin

Function Prototype:

bool begin(Stream* targetPort);

Function Description:

• Initializes the LLM Module UART interface configuration.

Parameters:

• Stream* targetPort:
  • Pass the Serial pointer.

Return Value:

• bool:
  • true: Initialization successful
  • false: Initialization failed

checkConnection

Function Prototype:

bool checkConnection();

Function Description:

• Sends the sys.ping command to check the connection status of the LLM Module.

Parameters:

• None

Return Value:

• bool:
  • true: Module responds
  • false: No response from module

update

Function Prototype:

void update();

Function Description:

• Pulls the LLM Module UART response data, this API should be included in the Loop and executed continuously.

Parameters:

• None

Return Value:

• None

ApiSys Class

The internal member ApiSys sys of M5ModuleLLM is used to control the SYS unit, enabling operations like system reset.

ping

Function Prototype:

int ping();

Function Description:

• Sends the sys.ping command to check the connection status of the LLM Module.

Parameters:

• None

Return Value:

• int:
  • MODULE_LLM_OK / Error Code

reset

Function Prototype:

int reset(bool waitResetFinish = true);

Function Description:

• Sends the sys.reset command to reset the software service.

Parameters:

• bool waitResetFinish:
  • true: Blocks and waits for reset to finish
  • false: Performs reset without blocking

Return Value:

• int:
  • MODULE_LLM_OK / Error Code

reboot

Function Prototype:

int reboot();

Function Description:

• Sends the sys.reboot command to reboot the system.

Parameters:

• None

Return Value:

• int:
  • MODULE_LLM_OK / Error Code

ApiAudio Class

Note: This function has been deprecated in version 1.3 and later, and is now automatically configured internally.

The internal member ApiAudio audio of M5ModuleLLM is used to control the initialization and configuration of the Audio unit.

setup

Function prototype:

String setup(ApiAudioSetupConfig_t config = ApiAudioSetupConfig_t(), String request_id = "audio_setup");

Function description:

• Initializes the Audio unit and activates the system sound card. (This feature must be enabled before using KWS and TTS)

Parameters:

ApiAudioSetupConfig_t config:

• LLM unit initialization configuration:
• String request_id:
  • Session ID, default can be used.

struct ApiAudioSetupConfig_t {
    int capcard      = 0;
    int capdevice    = 0;
    float capVolume  = 0.5;
    int playcard     = 0;
    int playdevice   = 1;
    float playVolume = 0.15;
};

Parameter	Description	Input Values
capcard	Microphone sound card index	Default sound card: 0
capdevice	Microphone device index	Onboard silicon microphone: 0
capVolume	Input volume	0.0～10.0 (1<volume increases gain, default is 0.5)
playcard	Speaker sound card index	Default sound card: 0
playdevice	Speaker device index	Onboard speaker: 1
playVolume	Output volume	0.0～10.0 (1<volume increases gain, default is 0.5)

Return Value:

• String:
  • audio_work_id: audio unit work_id

ApiCamera Class

The internal member ApiCamera camera of M5ModuleLLM is used to control the initialization and configuration of the Camera unit.

setup

Function prototype:

String setup(ApiCameraSetupConfig_t config = ApiCameraSetupConfig_t(), String request_id = "camera_setup");

Function description:

• Initializes the Camera unit and activates camera input. (This feature must be enabled before using UVC)

Parameters:

ApiCameraSetupConfig_t config:

• Camera unit initialization configuration:
• String request_id:
  • Session ID, default can be used.

struct ApiCameraSetupConfig_t {
    String response_format = "camera.raw";
    String input           = "/dev/video0";
    bool enoutput          = false;
    int frame_width        = 320;
    int frame_height       = 320;
};

Parameter	Description	Input Values
input	UVC index	"/dev/video0"
enoutput	Whether to output image data via serial	Enable: true Disable: false
frame_width	Image width	320
frame_height	Image height	320

Return Value:

• String:
  • camera_work_id: camera unit work_id

ApiKws Class

The internal member ApiKws kws of M5ModuleLLM is used to control the initialization and configuration of the KWS unit.

setup

Function Prototype:

String setup(ApiKwsSetupConfig_t config = ApiKwsSetupConfig_t(), String request_id = "kws_setup",
             String language = "en_US");```

**Function Description:**

- Initializes the KWS unit and configures the wake-up keyword.

**Parameters:**

ApiKwsSetupConfig_t config:

- KWS unit initialization configuration:
- String request_id:
    - Session ID, default can be used.

```cpp
struct ApiKwsSetupConfig_t {
    String kws             = "HELLO";
    String model           = "sherpa-onnx-kws-zipformer-gigaspeech-3.3M-2024-01-01";
    String response_format = "kws.bool";
    String input           = "sys.pcm";
    bool enoutput          = true;
};

Parameter	Description	Input Values
model	Conversion Model	English Model: "sherpa-onnx-kws-zipformer-gigaspeech-3.3M-2024-01-01" Chinese Model: "sherpa-onnx-kws-zipformer-wenetspeech-3.3M-2024-01-01"
kws	KWS Wake-up Word Text	No mixing of Chinese and English; English must be in uppercase
enoutput	Enable UART Output	Enable: true Disable: false

Return Value:

• String:
  • kws_work_id: KWS unit work_id

ApiVad Class

The internal member ApiVad vad of M5ModuleLLM is used to control the initialization and configuration of the VAD unit.

setup

Function Prototype:

String setup(ApiVadSetupConfig_t config = ApiVadSetupConfig_t(), String request_id = "vad_setup");

Function Description:

• Initializes the VAD unit.

Parameters:

ApiVadSetupConfig_t config:

• VAD unit initialization configuration:
• String request_id:
  • Session ID, default can be used.

struct ApiKwsSetupConfig_t {
    String model           = "silero-vad";
    String response_format = "vad.bool";
    String input           = {"sys.pcm", "kws.1000"};
    bool enoutput          = true;
};

Parameter	Description	Input Values
model	Conversion Model	Model: "silero-vad"
input	Input	KWS Wake-up Input: "kws.xxx" (input the KWS unit's work_id) Onboard Microphone Input: "sys.pcm" UART Stream Input: "vad.wav.stream.base64"
enoutput	Enable UART Output	Enable: true Disable: false

Return Value:

• String:
  • vad_work_id: VAD unit work_id

ApiAsr Class

The internal member ApiAsr asr of M5ModuleLLM is used to control the initialization and configuration of the ASR unit.

setup

Function Prototype:

String setup(ApiAsrSetupConfig_t config = ApiAsrSetupConfig_t(), String request_id = "asr_setup",
             String language = "en_US");

Function Description:

• Initializes the ASR unit and enables speech-to-text functionality.

Input Parameters:

ApiAsrSetupConfig_t config:

• ASR unit initialization configuration:
• String request_id:
  • Session ID, the default can be used.

struct ApiAsrSetupConfig_t {
    String model           = "sherpa-ncnn-streaming-zipformer-20M-2023-02-17";
    String response_format = "asr.utf-8.stream";
    String input           = ["sys.pcm", "kws.1000"];
    bool enoutput          = true;
    float rule1            = 2.4;
    float rule2            = 1.2;
    float rule3            = 30.0;
};

Parameter	Description	Input Values
model	Conversion model	English Model: "sherpa-ncnn-streaming-zipformer-20M-2023-02-17" Chinese Model: "sherpa-ncnn-streaming-zipformer-zh-14M-2023-02-23"
response_format	Output format	Normal output: "asr.utf-8" Stream output: "asr.utf-8.stream"
input	Input	KWS wake input: "kws.xxx" (input kws unit work_id) Onboard microphone input: "sys.pcm" UART stream input: "asr.wav.stream.base64"
rule1	Timeout for unrecognized content wake	Unit: seconds
rule2	Maximum recognition interval	Unit: seconds
rule3	Maximum recognition timeout	Unit: seconds
enoutput	Enable UART output	Enable: true Disable: false

Return Value:

• String:
  • asr_work_id: ASR unit work_id

ApiWhisper Class

The internal member ApiWhisper whisper of M5ModuleLLM is used to control the initialization and configuration of the Whisper unit.

setup

Function Prototype:

String setup(ApiWhisperSetupConfig_t config = ApiWhisperSetupConfig_t(), String request_id = "asr_setup",

Function Description:

• Initializes the Whisper unit and enables speech-to-text functionality.

Input Parameters:

ApiWhisperSetupConfig_t config:

• Whisper unit initialization configuration:
• String request_id:
  • Session ID, the default can be used.

struct ApiAsrSetupConfig_t {
    String model           = "whisper-tiny";
    String response_format = "asr.utf-8";
    String input           = [ "sys.pcm", "kws.1000", "vad.1001" ];
    String language        = "en";
    bool enoutput          = true;
};

Parameter	Description	Input Values
model	Conversion model	Model: "whisper-tiny"
response_format	Output format	Normal output: "asr.utf-8"
input	Input	KWS wake input: "kws.xxx" (input kws unit work_id) Onboard microphone input: "sys.pcm" UART stream input: "asr.wav.stream.base64"
language	Language used for language recognition	Default: "en" Optional: "zh", "ja"
enoutput	Enable UART output	Enable: true Disable: false

Return Value:

• String:
  • whisper_work_id: Whisper unit work_id

ApiLlm Class

The internal member ApiLlm llm of M5ModuleLLM is used to control the initialization and configuration of the LLM unit.

setup

Function prototype:

String setup(ApiLlmSetupConfig_t config = ApiLlmSetupConfig_t(), String request_id = "llm_setup");

Function Description:

• Initializes the LLM unit and supports configuring the input and output data format for the LLM unit.

Parameters:

• ApiLlmSetupConfig_t config:
  • LLM unit initialization configuration
• String request_id:
  • Session ID, the default value can be used.

struct ApiLlmSetupConfig_t {
    String prompt;
    String model           = "qwen2.5-0.5B-prefill-20e";
    String response_format = "llm.utf-8.stream";
    String input           = ["llm.utf-8", "kws.1000"];
    bool enoutput          = true;
    int max_token_len      = 127;
};

Parameter	Description	Input Values
model	Model used for conversion	Predefined model "qwen2.5-0.5B-prefill-20e"
response_format	Output format	Normal output: "llm.utf-8" Streaming output: "llm.utf-8.stream"
input	Input format	ASR input: "asr.xxx" (work_id of the ASR unit) UART input: "llm.utf-8" KWS wake-up interruption: "kws.xxx" (work_id of the KWS unit)
max_length	Configures the maximum output token length (maximum returned inference text length)	Maximum value: 1023
prompt	Model initialization system prompt	String
enoutput	Enable UART output	Enable: true Disable: false

Return Value:

• String:
  • llm_work_id: LLM unit work ID

inference

Function prototype:

int inference(String work_id, String input, String request_id = "llm_inference");

Function Description:

• Sends input data to start inference. The result will be placed in the responseMsgList container in M5ModuleLLM.msg.

Parameters:

• String work_id:
  • The LLM unit's work ID being called
• String input:
  • Input text
• String request_id:
  • Session ID, used to differentiate when multiple sessions exist.

Return Value:

• int:
  • MODULE_LLM_OK / Error Code

inferenceAndWaitResult

Function prototype:

int inferenceAndWaitResult(String work_id, String input, std::function<void(String&)> onResult, uint32_t timeout = 5000, String request_id = "llm_inference");

Function Description:

• Sends input data to start inference, blocks while waiting for the result, then calls the callback function.

Parameters:

• String work_id:
  • The LLM unit's work ID being called
• String input:
  • Input text
• void onResult(String&)
  • Callback function for inference result
• uint32_t timeout:
  • Timeout for waiting for inference result
• String request_id:
  • Session ID, used to differentiate when multiple sessions exist.

Return Value:

• int:
  • MODULE_LLM_OK / Error Code

ApiVlm Class

The internal member ApiVlm vlm of M5ModuleLLM is used to control the initialization and configuration of the VLM unit.

setup

Function prototype:

String setup(ApiVlmSetupConfig_t config = ApiVlmSetupConfig_t(), String request_id = "vlm_setup");

Function Description:

• Initializes the VLM unit and supports configuring the input and output data format for the VLM unit.

Parameters:

• ApiVlmSetupConfig_t config:
  • VLM unit initialization configuration
• String request_id:
  • Session ID, the default value can be used.

struct ApiVlmSetupConfig_t {
    String prompt;
    String model           = "internvl2.5-1B-ax630c";
    String response_format = "vlm.utf-8.stream";
    String input           = ["vlm.utf-8", "kws.1000"];
    bool enoutput          = true;
    int max_token_len      = 1023;
};

Parameter	Description	Input Values
model	Model used for conversion	Predefined model "internvl2.5-1B-ax630c"
response_format	Output format	Normal output: "vlm.utf-8" Streaming output: "vlm.utf-8.stream"
input	Input format	ASR input: "asr.xxx" (work_id of the ASR unit) UART input: "llm.utf-8" KWS wake-up interruption: "kws.xxx" (work_id of the KWS unit)
max_length	Configures the maximum output token length (maximum returned inference text length)	Maximum value: 1023
prompt	Model initialization system prompt	String
enoutput	Enable UART output	Enable: true Disable: false

Return Value:

• String:
  • vlm_work_id: VLM unit work ID

inference

Function prototype:

int inference(String work_id, String input, String request_id = "vlm_inference");

Function Description:

• Sends input data to start inference. The result will be placed in the responseMsgList container in M5ModuleLLM.msg.

Parameters:

• String work_id:
  • The VLM unit's work ID being called
• String input:
  • Input text
• String request_id:
  • Session ID, used to differentiate when multiple sessions exist.

Return Value:

• int:
  • MODULE_LLM_OK / Error Code

inferenceAndWaitResult

Function prototype:

int inferenceAndWaitResult(String work_id, String input, std::function<void(String&)> onResult,
                           uint32_t timeout = 5000, String request_id = "vlm_inference");

Function Description:

• Sends input data to start inference, blocks while waiting for the result, then calls the callback function.

Parameters:

• String work_id:
  • The VLM unit's work ID being called
• String input:
  • Input text
• void onResult(String&)
  • Callback function for inference result
• uint32_t timeout:
  • Timeout for waiting for inference result
• String request_id:
  • Session ID, used to differentiate when multiple sessions exist.

Return Value:

• int:
  • MODULE_LLM_OK / Error Code

ApiTts Class

The internal member ApiTts tts of M5ModuleLLM is used to control the initialization and configuration of the TTS unit.

setup

Function prototype:

String setup(ApiTtsSetupConfig_t config = ApiTtsSetupConfig_t(), String request_id = "tts_setup");

Function description:

• Initializes the TTS unit and enables the text-to-speech functionality.

Parameters:

• ApiTtsSetupConfig_t config:
  • LLM unit initialization configuration:
• String request_id:
  • Session ID, use the default if not needed.

struct ApiTtsSetupConfig_t {
    String model           = "single_speaker_english_fast";
    String response_format = "sys.pcm";
    String input           = ["tts.utf-8.stream", "kws.1000"];
    bool enoutput          = false;
    bool enaudio           = true;
};

Parameter	Description	Input values
model	Conversion model	English model: "single_speaker_english_fast" Chinese model: "single_speaker_fast"
input	Input	LLM input: "llm.xxx" (input LLM unit's work_id) UART input: "tts.utf-8" UART stream input: "tts.utf-8.stream" KWS wake-up interrupt: "kws.xxx" (input KWS unit's work_id)
enoutput	Enable UART output	Enable: true Disable: false
enaudio	Enable speaker playback	Enable: true Disable: true

Return value:

• String:
  • tts_work_id: TTS unit work_id

inference

Function prototype:

int inference(String work_id, String input, uint32_t timeout = 0, String request_id = "tts_inference");

Function description:

• Input data and start the inference conversion. After completion, the speaker will automatically play.

Parameters:

• String work_id:
  • Work ID of the TTS unit to be called.
• String input:
  • Input text.
• uint32_t timeout:
  • Timeout for waiting for inference.
• String request_id:
  • Session ID, used to distinguish between multiple sessions.

Return value:

• int:
  • MODULE_LLM_OK / Error Code

ApiMelotts Class

The internal member ApiMelotts melotts of M5ModuleLLM is used to control the initialization and configuration of the Melotts unit.

setup

Function prototype:

String setup(ApiMelottsSetupConfig_t config = ApiMelottsSetupConfig_t(), String request_id = "melotts_setup", 
             String language = "en_US");

Function description:

• Initializes the Melotts unit and enables the text-to-speech functionality.

Parameters:

• ApiMelottsSetupConfig_t config:
  • Melotts unit initialization configuration:
• String request_id:
  • Session ID, use the default if not needed.

struct ApiMelottsSetupConfig_t {
    String model              = "melotts_zh-cn";
    String response_format    = "sys.pcm";
    std::vector<String> input = {"tts.utf-8.stream"};
    bool enoutput             = false;
    bool enaudio              = true;
};

Parameter	Description	Input values
model	Conversion model	Chinese and English model: "melotts_zh-cn" Chinese model: "single_speaker_fast"
input	Input	LLM input: "llm.xxx" (input LLM unit's work_id) UART input: "melotts.utf-8" UART stream input: "melotts.utf-8.stream"
enoutput	Enable UART output	Enable: true Disable: false
enaudio	Enable speaker playback	Enable: true Disable: true

Return value:

• String:
  • melotts_work_id: Melotts unit work_id

inference

Function prototype:

int inference(String work_id, String input, uint32_t timeout = 0, String request_id = "tts_inference");

Function description:

• Input data and start the inference conversion. After completion, the speaker will automatically play.

Parameters:

• String work_id:
  • Work ID of the Melotts unit to be called.
• String input:
  • Input text.
• uint32_t timeout:
  • Timeout for waiting for inference.
• String request_id:
  • Session ID, used to distinguish between multiple sessions.

Return value:

• int:
  • MODULE_LLM_OK / Error Code

ApiYolo Class

The internal member ApiYolo yolo of M5ModuleLLM is used to control the initialization and configuration of the Yolo unit.

setup

Function prototype:

String setup(ApiYoloSetupConfig_t config = ApiYoloSetupConfig_t(), String request_id = "yolo_setup");

Function description:

• Initializes the Yolo unit and enables image detection functionality.

Parameters:

• ApiYoloSetupConfig_t config:
  • Yolo unit initialization configuration:
• String request_id:
  • Session ID, use the default if not needed.

struct ApiYoloSetupConfig_t {
    String model              = "yolo11n";
    String response_format    = "yolo.box.stream";
    std::vector<String> input = {"yolo.jpeg.base64"};
    bool enoutput             = true;
};

Parameter	Description	Input values
model	Conversion model	Detection model: "yolo11n" Pose model: "yolo11n-pose" Hand pose model: "yolo11n-hand-pose"
response_format	Output format	Detection output: "yolo.box.stream" Pose output: "yolo.pose.stream"
input	Input	UVC input: "camera.xxx" (input camera unit's work_id) UART stream input: "yolo.jpeg.base64.stream"
enoutput	Enable UART output	Enable: true Disable: false

Return value:

• String:
  • yolo_work_id: Yolo unit work_id

ModuleMsg Class

The internal member ModuleMsg msg of M5ModuleLLM provides a container responseMsgList used to cache various information returned from the LLM Module. Refer to the following example, where the main loop iterates to retrieve the results.

void loop()
{
    module_llm.update();

    // Handle response msg
    for (auto& msg : module_llm.msg.responseMsgList) {
        // KWS msg
        if (msg.work_id == kws_work_id) {
            Serial.printf(">> Keyword detected\n");
        }

        // ASR msg
        if (msg.work_id == asr_work_id) {
            if (msg.object == "asr.utf-8.stream") {
                // Parse and get asr result
                JsonDocument doc;
                deserializeJson(doc, msg.raw_msg);
                String asr_result = doc["data"]["delta"].as<String>();
                Serial.printf(">> %s\n", asr_result.c_str());
            }
        }
    }
    module_llm.msg.responseMsgList.clear();
}

VoiceAssistant Class

M5ModuleLLM_VoiceAssistant is used to quickly create an LLM voice assistant instance, achieving a fast implementation of KWS (keyword spotting) -> ASR (speech-to-text) -> LLM (large model inference) -> TTS (text-to-speech).

• During initialization, simply pass the M5ModuleLLM instance to the constructor, and register the corresponding event callback functions to complete the voice assistant setup.

/*
 * SPDX-FileCopyrightText: 2024 M5Stack Technology CO LTD
 *
 * SPDX-License-Identifier: MIT
 */
#include <Arduino.h>
#include <M5Unified.h>
#include <M5ModuleLLM.h>

M5ModuleLLM module_llm;
M5ModuleLLM_VoiceAssistant voice_assistant(&module_llm);

/* On ASR data callback */
void on_asr_data_input(String data, bool isFinish, int index)
{
    M5.Display.setTextColor(TFT_GREEN, TFT_BLACK);
    M5.Display.printf(">> %s\n", data.c_str());

    /* If ASR data is finish */
    if (isFinish) {
        M5.Display.setTextColor(TFT_YELLOW, TFT_BLACK);
        M5.Display.print(">> ");
    }
};

/* On LLM data callback */
void on_llm_data_input(String data, bool isFinish, int index)
{
    M5.Display.print(data);

    /* If LLM data is finish */
    if (isFinish) {
        M5.Display.print("\n");
    }
};

void setup()
{
    M5.begin();
    M5.Display.setTextSize(2);
    M5.Display.setTextScroll(true);

    /* Init module serial port */
    Serial2.begin(115200, SERIAL_8N1, 16, 17);  // Basic
    // Serial2.begin(115200, SERIAL_8N1, 13, 14);  // Core2
    // Serial2.begin(115200, SERIAL_8N1, 18, 17);  // CoreS3

    /* Init module */
    module_llm.begin(&Serial2);

    /* Make sure module is connected */
    M5.Display.printf(">> Check ModuleLLM connection..\n");
    while (1) {
        if (module_llm.checkConnection()) {
            break;
        }
    }

    /* Begin voice assistant preset */
    M5.Display.printf(">> Begin voice assistant..\n");
    int ret = voice_assistant.begin("HELLO");
    if (ret != MODULE_LLM_OK) {
        while (1) {
            M5.Display.setTextColor(TFT_RED);
            M5.Display.printf(">> Begin voice assistant failed\n");
        }
    }

    /* Register on ASR data callback function */
    voice_assistant.onAsrDataInput(on_asr_data_input);

    /* Register on LLM data callback function */
    voice_assistant.onLlmDataInput(on_llm_data_input);

    M5.Display.printf(">> Voice assistant ready\n");
}

void loop()
{
    /* Keep voice assistant preset update */
    voice_assistant.update();
}

Error Code

enum ModuleLLMErrorCode_t {
    MODULE_LLM_OK                              = 0,
    MODULE_LLM_RESET_WARN                      = -1,
    MODULE_LLM_JSON_FORMAT_ERROR               = -2,
    MODULE_LLM_ACTION_MATCH_FAILED             = -3,
    MODULE_LLM_INFERENCE_DATA_PUSH_FAILED      = -4,
    MODULE_LLM_MODEL_LOADING_FAILED            = -5,
    MODULE_LLM_UNIT_NOT_EXIST                  = -6,
    MODULE_LLM_UNKNOWN_OPERATION               = -7,
    MODULE_LLM_UNIT_RESOURCE_ALLOCATION_FAILED = -8,
    MODULE_LLM_UNIT_CALL_FAILED                = -9,
    MODULE_LLM_MODEL_INIT_FAILED               = -10,
    MODULE_LLM_MODEL_RUN_FAILED                = -11,
    MODULE_LLM_MODULE_NOT_INITIALISED          = -12,
    MODULE_LLM_MODULE_ALREADY_WORKING          = -13,
    MODULE_LLM_MODULE_NOT_WORKING              = -14,
    MODULE_LLM_NO_UPDATEABLE_MODULES           = -15,
    MODULE_LLM_NO_MODULES_AVAILABLE_FOR_UPDATE = -16,
    MODULE_LLM_FILE_OPEN_FAILED                = -17,
    MODULE_LLM_WAIT_RESPONSE_TIMEOUT           = -97,
    MODULE_LLM_RESPONSE_PARSE_FAILED           = -98,
    MODULE_LLM_ERROR_NONE                      = -99,
};