将安全设置传递给Vertex AI模型的正确格式是使用vertexai.generative_models._generative_models模块中的SafetySettingsType。在创建Vertex类示例时，应将其作为名为safety_settings的可选参数传递。
以下是如何操作的示例：

from vertexai.generative_models._generative_models import SafetySettingsType
from llama_index.llms.vertex import Vertex

# Define your safety settings
safety_settings = SafetySettingsType(
    # Add your specific safety settings here
)

# Set up necessary variables
credentials = {
    "project_id": "INSERT_PROJECT_ID",
    "api_key": "INSERT_API_KEY",
}

# Create an instance of the Vertex class with safety settings
llm = Vertex(
    model="text-bison",
    project=credentials["project_id"],
    credentials=credentials,
    safety_settings=safety_settings
)

# Access the complete method from the instance
response = llm.complete("Hello world!")
print(str(response))

这确保了safety_settings参数被正确传递，并避免了pydantic.v1.errors.ConfigError1(https://github.com/run-llama/llama_index/blob/main/llama-index-integrations/llms/llama-index-llms-vertex/llama_index/llms/vertex/base.py)。

详细信息

关于Dosu

此响应旨在帮助您节省时间并提高效率。它不是精确解决方案，而是您自己研究的起点。
请帮助我变得更有用！如果这有帮助，请留下一个👍，如果不相关，请留下一个👎。

• 要继续对话，请提及@dosu。*

根据错误提示，需要调用update forward refs方法，因此类必须更新。解决方法是在使用之前手动调用它。

@dosu-bot, @logan-markewich

safety_settings={
                HarmCategory.HARM_CATEGORY_DANGEROUS_CONTENT: HarmBlockThreshold.BLOCK_ONLY_HIGH,
                HarmCategory.HARM_CATEGORY_HATE_SPEECH: HarmBlockThreshold.BLOCK_MEDIUM_AND_ABOVE,
                HarmCategory.HARM_CATEGORY_HARASSMENT: HarmBlockThreshold.BLOCK_LOW_AND_ABOVE,
                HarmCategory.HARM_CATEGORY_SEXUALLY_EXPLICIT: HarmBlockThreshold.BLOCK_NONE,
            }

Vertex.update_forward_refs()
llm = Vertex(
    model="gemini-1.5-pro-preview-0514", 
    safety_settings=safety_settings, 
    project=credentials.project_id, 
    credentials=credentials,
    temperature = 0.3,
    max_tokens = 8192,
)

NameError: name 'SafetySetting' is not defined. Did you mean: 'SafetySettingsType'?

要解决 NameError: name 'SafetySetting' is not defined 错误，您需要确保 SafetySetting 被正确定义或导入。然而，在您的情况中，似乎您应该使用 SafetySettingsType。
以下是如何正确定义和传递安全设置的方法：

1. 导入必要的类：

from vertexai.preview.language_models import HarmCategory, HarmBlockThreshold
from vertexai.generative_models._generative_models import SafetySettingsType
from llama_index.llms.vertex import Vertex

1. 使用 SafetySettingsType 定义您的安全设置：

safety_settings = SafetySettingsType(
    {
        HarmCategory.HARM_CATEGORY_DANGEROUS_CONTENT: HarmBlockThreshold.BLOCK_ONLY_HIGH,
        HarmCategory.HARM_CATEGORY_HATE_SPEECH: HarmBlockThreshold.BLOCK_MEDIUM_AND_ABOVE,
        HarmCategory.HARM_CATEGORY_HARASSMENT: HarmBlockThreshold.BLOCK_LOW_AND_ABOVE,
        HarmCategory.HARM_CATEGORY_SEXUALLY_EXPLICIT: HarmBlockThreshold.BLOCK_NONE,
    }
)

1. 调用 update_forward_refs 并创建 Vertex 示例：

Vertex.update_forward_refs()

llm = Vertex(
    model="gemini-1.5-pro-preview-0514",
    safety_settings=safety_settings,
    project=credentials.project_id,
    credentials=credentials,
    temperature=0.3,
    max_tokens=8192,
)

这应该解决 NameError 并正确将安全设置传递给您的 Vertex AI 模型 1(https://github.com/run-llama/llama_index/blob/main/llama-index-integrations/llms/llama-index-llms-vertex/llama_index/llms/vertex/base.py)。

• 要继续与 Dosu 对话，请提及 @dosu。*