调用模型，如果希望控制和引导模型的输出，您可以通过续写模式实现：预填`assistant`角色的部分内容，模型将基于该内容延续输出，以此达成输出控制目标。您可以将该输出控制机制应用于以下场景：强制输出 JSON/XML 等特定格式、绕过模型最大输出限制生成超长内容、确保大模型在角色扮演场景中保持角色一致性。 ## 模型及 API 支持的模型如下： * 文本生成模型中无深度思考能力的模型：具体模型可在[文本生成能力](https://ark.volcengine.com/region:cn-beijing/docs/ark/model-list#32e9e053)筛选。 * 部分具有深度思考能力的模型： * doubao\-seed\-evolving * doubao\-seed\-2\-1\-pro\-260915 * doubao\-seed\-2\-1\-lite\-260915 * doubao\-seed\-2\-1\-pro\-260628 * doubao\-seed\-2\-1\-turbo\-260628 * doubao\-seed\-2\-0\-lite\-260428 * doubao\-seed\-2\-0\-mini\-260428 * doubao\-seed\-2\-0\-pro\-260215 * doubao\-seed\-2\-0\-lite\-260215 * doubao\-seed\-2\-0\-mini\-260215 * doubao\-seed\-2\-0\-code\-preview\-260215 * doubao\-seed\-character\-260628 * glm\-5\-3\-flash\-260828 * deepseek\-v4\-1\-flash\-260910

注意

其他深度思考模型不可使用续写模式，即请求时 message 以 assistant role 消息结尾时，thinking 字段会失效，模型保持默认是否深度思考的行为。

API 支持： * [Chat API](https://ark.volcengine.com/region:cn-beijing/docs/ark/chat-api) * [Responses API](https://ark.volcengine.com/region:cn-beijing/docs/ark/create-model-responses-api) ## 核心配置 ### 方式一 **支持模型**：除 `deepseek-v4-1-flash-260910` 之外的其他模型。请将消息列表的最后一条消息的 **role** 设置为`assistant`，模型会遵循 `assistant` 消息的已有格式和内容来续写。 ```Python messages = [ {"role": "user", "content": "You are a calculator, please calculate: 1 + 1"}, # 最后消息为模型消息，role 为 assistant {"role": "assistant", "content": "="} ] ``` ### 方式二 **支持模型**：`deepseek-v4-1-flash-260910` 请将消息列表的最后一条消息的 **role** 设置为`assistant`，并设置最后一条消息的 **prefix** 参数为`true`。模型会遵循 `assistant` 消息的已有格式和内容来续写。 ```Python messages = [ {"role": "user", "content": "You are a calculator, please calculate: 1 + 1"}, {"role": "assistant", "content": "=", "prefix": True} ] ``` ## 模式对比 |普通模式 |续写模式 | |---|---| |请求示例

```Python```
```import os```
```# Install SDK: pip install arkruntime```
```from arkruntime import Ark```
``````
```client = Ark(```
``` # The base URL for model invocation```
``` base_url="https://ark.cn-beijing.volces.com/api/v3",```
``` # Get API Key: https://ark.volcengine.com/region:cn-beijing/apikey```
``` api_key=os.getenv('ARK_API_KEY'),```
```)```
``````
```completion = client.chat.completions.create(```
``` # Replace with Model ID```
``` model = "doubao-seed-2-1-pro-260915",```
``` messages=[```
``` {"role": "user", "content": "You are a calculator, please calculate: 1 + 1 "},```
``` ]```
```)```
```print(completion.choices[0].message.content)```
|请求示例

```Python```
```import os```
```# Install SDK: pip install arkruntime```
```from arkruntime import Ark```
``````
```client = Ark(```
``` # The base URL for model invocation```
``` base_url="https://ark.cn-beijing.volces.com/api/v3",```
``` # Get API Key: https://ark.volcengine.com/region:cn-beijing/apikey```
``` api_key=os.getenv('ARK_API_KEY'),```
```)```
``````
```completion = client.chat.completions.create(```
``` # Replace with Model ID```
``` model = "doubao-seed-2-1-pro-260915",```
``` messages=[```
``` {"role": "user", "content": "You are a calculator, please calculate: 1 + 1"},```
``` # 额外传入模型信息,模型据此续写```
``` {"role": "assistant", "content": "="}```
``` ]```
```)```
```print(completion.choices[0].message.content)```
| |返回示例

```Shell```
```1+1等于2。```

输出会按照模型根据问题的完整自然语言回复。 |返回示例

```Shell```
```2```

模型根据信息“=”续写，保持对应的格式和行文风格。 | 接下来，对于几种典型使用场景进行举例说明。 ## 场景：改善输出格式模型可能会基于自身理解输出内容，导致结果无法直接被其他程序解析。可以通过预填充 `{` 符号，引导模型跳过一些场景回复，直接输出 JSON 对象，会让回答更加简洁和工整，可以被其他程序更好地解析。 |普通模式 |续写模式 | |---|---| |请求示例

```Python```
```import os```
```# Install SDK: pip install arkruntime```
```from arkruntime import Ark```
``````
```client = Ark(```
``` # The base URL for model invocation```
``` base_url="https://ark.cn-beijing.volces.com/api/v3",```
``` # Get API Key: https://ark.volcengine.com/region:cn-beijing/apikey```
``` api_key=os.getenv('ARK_API_KEY'),```
```)```
``````
```completion = client.chat.completions.create(```
``` # Replace with Model ID```
``` model = "doubao-seed-2-1-pro-260915",```
``` messages=[```
``` {"role": "user", "content": "Use JSON to describe the name and function of the Doubao model"}```
``` ]```
```)```
```print(completion.choices[0].message.content)```
|请求示例

```Python```
```import os```
```# Install SDK: pip install arkruntime```
```from arkruntime import Ark```
``````
```client = Ark(```
``` # The base URL for model invocation```
``` base_url="https://ark.cn-beijing.volces.com/api/v3",```
``` # Get API Key: https://ark.volcengine.com/region:cn-beijing/apikey```
``` api_key=os.getenv('ARK_API_KEY'),```
```)```
``````
```completion = client.chat.completions.create(```
``` # Replace with Model ID```
``` model = "doubao-seed-2-1-pro-260915",```
``` messages=[```
``` {"role": "user", "content": "Use JSON to describe the name and function of the Doubao model"},```
``` # 额外传入一个 { 的模型信息,模型据此续写```
``` {"role": "assistant", "content": "{"},```
``` ]```
```)```
```print(completion.choices[0].message.content)```
| |返回示例

返回一些说明性语言，无法直接按照 JSON 解析。

```Shell```
```Here's a JSON description of the Doubao model, including its name and function:```
``````json
{
"model": {
"name": "Doubao",
"alternative_names": ["doubao"],
"type": "large_language_model",
"developer": "ByteDance",
"primary_function": {
"description": "Doubao is an AI assistant designed to understand and generate human-like text responses",
"capabilities": [
"natural language understanding",
"text generation",
"question answering",
"translation between languages",
"text summarization",
"creative writing",
"code generation and explanation",
"general knowledge provision"
]
},
"languages_supported": ["Chinese", "English"],
"access_method": "API and web/mobile interfaces",
"version": "1.0",
"release_date": "2023"
}
}
``````
``````
``````
```Note: Some details about Doubao might not be publicly available, as it's a relatively new model from ByteDance. The information here represents common capabilities of similar large language models.```
``````Plain Text

```
|返回示例

和前置`assistant`信息组合成 JSON 返回。

```Shell```
```"model_name": "Doubao",```
``` "functions": [```
``` {```
``` "name": "text_generation",```
``` "description": "Generates human-like text based on input prompts, supporting creative writing, summarization, and more."```
``` },```
``` {```
``` "name": "language_translation",```
``` "description": "Translates text between multiple languages with high accuracy and contextual understanding."```
``` },```
``` {```
``` "name": "sentiment_analysis",```
``` "description": "Analyzes the sentiment of input text, determining whether it is positive, negative, or neutral."```
``` },```
``` {```
``` "name": "question_answering",```
``` "description": "Provides accurate answers to questions based on the input context or general knowledge."```
``` },```
``` {```
``` "name": "code_generation",```
``` "description": "Generates code snippets in various programming languages based on natural language descriptions."```
``` },```
``` {```
``` "name": "conversational_ai",```
``` "description": "Engages in natural, context-aware conversations with users, simulating human-like interactions."```
``` }```
``` ]```
```}```
| ### 注意事项 * 方舟已提供结构化输出能力，可让模型输出严格符合格式要求的 JSON 内容，请参见[结构化输出(beta)](https://ark.volcengine.com/region:cn-beijing/docs/ark/structured-output-beta)。 * 由于模型具有一定的随机性，无法 100% 保证回复为标准 JSON 对象。建议您对模型的回复进行校验，下面示例使用 [json_repair](https://github.com/mangiucugna/json_repair) 库校验返回的是否为合法的 JSON 对象。 ```Python import json # Install json-repair: pip install json-repair import json_repair import os # Install SDK: python -m pip install --upgrade arkruntime from arkruntime import Ark PREFILL_PREFIX = "{" client = Ark( # The base URL for model invocation base_url="https://ark.cn-beijing.volces.com/api/v3", # Get API Key: https://ark.volcengine.com/region:cn-beijing/apikey api_key=os.getenv('ARK_API_KEY'), ) completion = client.chat.completions.create( # Replace with Model ID model = "doubao-seed-2-1-pro-260915", messages=[ {"role": "user", "content": "Use JSON to describe the name and function of the Doubao model"}, {"role": "assistant", "content": PREFILL_PREFIX}, ] ) # 拼接 PREFILL_PREFIX 和模型输出 json_string = PREFILL_PREFIX + completion.choices[0].message.content # 解析 JSON Object obj = {} try: obj = json.loads(json_string) except json.JSONDecodeError as e: obj = json_repair.loads(json_string) print(obj) ``` ## 场景：输出超过模型输出上限的内容使用续写模式让大模型根据前轮请求续写内容，并拼接多次输出的内容，组合成超长输出。可以避免因为 **max_tokens**（最大输出长度）限制，导致无法获取完整内容。 > 这种方式会将模型输出多次返回给模型，会有更多 token 用量。 > 注意增加循环次数限制等兜底，防止死循环输出等特殊情况，导致意外花费。 ```Python import os # Install SDK: python -m pip install --upgrade arkruntime from arkruntime import Ark messages = [ {"role": "user", "content": '''Please translate the following text into English: '''}, # Replace with your content {"role": "assistant", "content": ""} ] # Replace with Model ID ark_model="doubao-seed-2-1-pro-260915" client = Ark( # The base URL for model invocation base_url="https://ark.cn-beijing.volces.com/api/v3", # Get API Key: https://ark.volcengine.com/region:cn-beijing/apikey api_key=os.getenv('ARK_API_KEY'), ) completion = client.chat.completions.create( model=ark_model, messages=messages ) loop_count = 0 # 初始化循环计数器 max_loops = 20 # 设置最大循环次数为20 # 触发因为输出长度限制，循环生成 assistant 回复 while completion.choices[0].finish_reason == "length": messages[-1]["content"] += completion.choices[0].message.content completion = client.chat.completions.create( model=ark_model, messages=messages ) # 增加计数器 loop_count += 1 # 检查是否超过最大循环次数 if loop_count >= max_loops: print(f"Warning: Reached maximum loop count {max_loops}, stopping content generation.") break # 返回最后一条回复 messages[-1]["content"] += completion.choices[0].message.content print(messages[-1]["content"]) ``` ## 场景：增强角色扮演一致性我们通常使用系统消息来设置角色的一些信息，但是随着对话轮次的增多，模型可能会跳脱出角色的约束，通过续写模式，提醒模型本次输出扮演的角色，保证对话符合预期。 ```Python import os # Install SDK: python -m pip install --upgrade arkruntime from arkruntime import Ark client = Ark( # The base URL for model invocation base_url="https://ark.cn-beijing.volces.com/api/v3", # Get API Key: https://ark.volcengine.com/region:cn-beijing/apikey api_key=os.getenv('ARK_API_KEY'), ) # Replace with Model ID ark_model="doubao-seed-2-1-pro-260915" messages = [ {"role": "system", "content": "The following is a scene from Journey to the West. Please converse according to the specified role. For each round of dialogue, only reply with the content of the current role; do not add any additional content.\n\nAfter leaving Wuzhuang Guan, Tang Seng and his disciples traveled for more than a month and arrived at the foot of a great mountain. Wukong looked around and saw lofty mountains, steep ridges, overgrown weeds, and extremely dangerous terrain."}, {"role": "assistant", "content": "Wukong said:"} ] completion = client.chat.completions.create( model=ark_model, messages=messages ) print(messages[-1]['content']+completion.choices[0].message.content) ``` 模型输出示例： ```Plain Text Wukong said, "Master, this mountain looks very dangerous. I will go explore the way first. You wait here." ``` 您也可以通过变更`messages`最后消息，实现角色切换。 ```Python ... messages[-1] = {"role": "assistant", "content": "Tang Seng said:"} completion = client.chat.completions.create( model=ark_model, messages=messages ) print(messages[-1]['content']+completion.choices[0].message.content) ``` 模型输出示例： ```Plain Text Tang Seng said: "Wukong, this mountain looks very dangerous, you must be careful to check it out, don't fall into the trap of the monster." ```

火山方舟