deploy: 623ceaf

en/_sources/tutorial/206-prompt.md.txt

@@ -300,6 +300,56 @@ print(prompt)
 ]
 ```
 
+
+### LiteLLMChatWrapper
+
+`LiteLLMChatWrapper` encapsulates the litellm chat API, which takes a list of
+messages as input. The litellm supports different types of models, and each model
+might need to obey different formats. To simplify the usage, we provide a format
+that could be compatible with most models. If more specific formats are needed,
+you can refer to the specific model you use as well as the
+[litellm](https://github.com/BerriAI/litellm) documentation to customize your
+own format function for your model.
+
+
+- format all the messages in the chat history, into a single message with `"user"` as `role`
+
+#### Prompt Strategy
+
+- Messages will consist dialogue history in the `user` message prefixed by the system message and "## Dialogue History".
+
+```python
+from agentscope.models import LiteLLMChatWrapper
+
+model = LiteLLMChatWrapper(
+    config_name="", # empty since we directly initialize the model wrapper
+    model_name="gpt-3.5-turbo",
+)
+
+prompt = model.format(
+  Msg("system", "You are a helpful assistant", role="system"),
+  [
+      Msg("user", "What is the weather today?", role="user"),
+      Msg("assistant", "It is sunny today", role="assistant"),
+  ],
+)
+
+print(prompt)
+```
+
+```bash
+[
+    {
+        "role": "user",
+        "content": (
+            "You are a helpful assistant\n\n"
+            "## Dialogue History\nuser: What is the weather today?\n"
+            "assistant: It is sunny today"
+        ),
+    },
+]
+```
+
 ### OllamaChatWrapper
 
 `OllamaChatWrapper` encapsulates the Ollama chat API, which takes a list of

en/tutorial/206-prompt.html

@@ -403,6 +403,55 @@ <h4>Prompt Strategy<a class="headerlink" href="#id2" title="Link to this heading
 </div>
 </section>
 </section>
+<section id="litellmchatwrapper">
+<h3>LiteLLMChatWrapper<a class="headerlink" href="#litellmchatwrapper" title="Link to this heading"></a></h3>
+<p><code class="docutils literal notranslate"><span class="pre">LiteLLMChatWrapper</span></code> encapsulates the litellm chat API, which takes a list of
+messages as input. The litellm supports different types of models, and each model
+might need to obey different formats. To simplify the usage, we provide a format
+that could be compatible with most models. If more specific formats are needed,
+you can refer to the specific model you use as well as the
+<a class="reference external" href="https://github.com/BerriAI/litellm">litellm</a> documentation to customize your
+own format function for your model.</p>
+<ul class="simple">
+<li><p>format all the messages in the chat history, into a single message with <code class="docutils literal notranslate"><span class="pre">&quot;user&quot;</span></code> as <code class="docutils literal notranslate"><span class="pre">role</span></code></p></li>
+</ul>
+<section id="id3">
+<h4>Prompt Strategy<a class="headerlink" href="#id3" title="Link to this heading"></a></h4>
+<ul class="simple">
+<li><p>Messages will consist dialogue history in the <code class="docutils literal notranslate"><span class="pre">user</span></code> message prefixed by the system message and “## Dialogue History”.</p></li>
+</ul>
+<div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="kn">from</span> <span class="nn">agentscope.models</span> <span class="kn">import</span> <span class="n">LiteLLMChatWrapper</span>
+
+<span class="n">model</span> <span class="o">=</span> <span class="n">LiteLLMChatWrapper</span><span class="p">(</span>
+    <span class="n">config_name</span><span class="o">=</span><span class="s2">&quot;&quot;</span><span class="p">,</span> <span class="c1"># empty since we directly initialize the model wrapper</span>
+    <span class="n">model_name</span><span class="o">=</span><span class="s2">&quot;gpt-3.5-turbo&quot;</span><span class="p">,</span>
+<span class="p">)</span>
+
+<span class="n">prompt</span> <span class="o">=</span> <span class="n">model</span><span class="o">.</span><span class="n">format</span><span class="p">(</span>
+  <span class="n">Msg</span><span class="p">(</span><span class="s2">&quot;system&quot;</span><span class="p">,</span> <span class="s2">&quot;You are a helpful assistant&quot;</span><span class="p">,</span> <span class="n">role</span><span class="o">=</span><span class="s2">&quot;system&quot;</span><span class="p">),</span>
+  <span class="p">[</span>
+      <span class="n">Msg</span><span class="p">(</span><span class="s2">&quot;user&quot;</span><span class="p">,</span> <span class="s2">&quot;What is the weather today?&quot;</span><span class="p">,</span> <span class="n">role</span><span class="o">=</span><span class="s2">&quot;user&quot;</span><span class="p">),</span>
+      <span class="n">Msg</span><span class="p">(</span><span class="s2">&quot;assistant&quot;</span><span class="p">,</span> <span class="s2">&quot;It is sunny today&quot;</span><span class="p">,</span> <span class="n">role</span><span class="o">=</span><span class="s2">&quot;assistant&quot;</span><span class="p">),</span>
+  <span class="p">],</span>
+<span class="p">)</span>
+
+<span class="nb">print</span><span class="p">(</span><span class="n">prompt</span><span class="p">)</span>
+</pre></div>
+</div>
+<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span><span class="o">[</span>
+<span class="w">    </span><span class="o">{</span>
+<span class="w">        </span><span class="s2">&quot;role&quot;</span>:<span class="w"> </span><span class="s2">&quot;user&quot;</span>,
+<span class="w">        </span><span class="s2">&quot;content&quot;</span>:<span class="w"> </span><span class="o">(</span>
+<span class="w">            </span><span class="s2">&quot;You are a helpful assistant\n\n&quot;</span>
+<span class="w">            </span><span class="s2">&quot;## Dialogue History\nuser: What is the weather today?\n&quot;</span>
+<span class="w">            </span><span class="s2">&quot;assistant: It is sunny today&quot;</span>
+<span class="w">        </span><span class="o">)</span>,
+<span class="w">    </span><span class="o">}</span>,
+<span class="o">]</span>
+</pre></div>
+</div>
+</section>
+</section>
 <section id="ollamachatwrapper">
 <h3>OllamaChatWrapper<a class="headerlink" href="#ollamachatwrapper" title="Link to this heading"></a></h3>
 <p><code class="docutils literal notranslate"><span class="pre">OllamaChatWrapper</span></code> encapsulates the Ollama chat API, which takes a list of
@@ -413,8 +462,8 @@ <h3>OllamaChatWrapper<a class="headerlink" href="#ollamachatwrapper" title="Link
 <code class="docutils literal notranslate"><span class="pre">&quot;system&quot;</span></code>, or <code class="docutils literal notranslate"><span class="pre">&quot;assistant&quot;</span></code>.</p></li>
 <li><p>An optional <code class="docutils literal notranslate"><span class="pre">images</span></code> field can be added to the message</p></li>
 </ul>
-<section id="id3">
-<h4>Prompt Strategy<a class="headerlink" href="#id3" title="Link to this heading"></a></h4>
+<section id="id4">
+<h4>Prompt Strategy<a class="headerlink" href="#id4" title="Link to this heading"></a></h4>
 <ul class="simple">
 <li><p>If the role field of the first input message is <code class="docutils literal notranslate"><span class="pre">&quot;system&quot;</span></code>,
 it will be treated as system prompt and the other messages will consist
@@ -455,8 +504,8 @@ <h4>Prompt Strategy<a class="headerlink" href="#id3" title="Link to this heading
 <h3>OllamaGenerationWrapper<a class="headerlink" href="#ollamagenerationwrapper" title="Link to this heading"></a></h3>
 <p><code class="docutils literal notranslate"><span class="pre">OllamaGenerationWrapper</span></code> encapsulates the Ollama generation API, which
 takes a string prompt as input without any constraints (updated to 2024/03/22).</p>
-<section id="id4">
-<h4>Prompt Strategy<a class="headerlink" href="#id4" title="Link to this heading"></a></h4>
+<section id="id5">
+<h4>Prompt Strategy<a class="headerlink" href="#id5" title="Link to this heading"></a></h4>
 <p>If the role field of the first message is <code class="docutils literal notranslate"><span class="pre">&quot;system&quot;</span></code>, a system prompt will be created. The rest of the messages will be combined into dialogue history in string format.</p>
 <div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="kn">from</span> <span class="nn">agentscope.models</span> <span class="kn">import</span> <span class="n">OllamaGenerationWrapper</span>
 <span class="kn">from</span> <span class="nn">agentscope.message</span> <span class="kn">import</span> <span class="n">Msg</span>
@@ -501,8 +550,8 @@ <h3><code class="docutils literal notranslate"><span class="pre">GeminiChatWrapp
 an agent may act as many different roles and speak continuously.
 Therefore, we decide to convert the list of messages into a user message
 in our built-in <code class="docutils literal notranslate"><span class="pre">format</span></code> function.</p>
-<section id="id5">
-<h4>Prompt Strategy<a class="headerlink" href="#id5" title="Link to this heading"></a></h4>
+<section id="id6">
+<h4>Prompt Strategy<a class="headerlink" href="#id6" title="Link to this heading"></a></h4>
 <p>If the role field of the first message is <code class="docutils literal notranslate"><span class="pre">&quot;system&quot;</span></code>, a system prompt will be added in the beginning. The other messages will be combined into dialogue history.</p>
 <p><strong>Note</strong> sometimes the <code class="docutils literal notranslate"><span class="pre">parts</span></code> field may contain image urls, which is not
 supported in <code class="docutils literal notranslate"><span class="pre">format</span></code> function. We recommend developers to customize the
@@ -546,8 +595,8 @@ <h3><code class="docutils literal notranslate"><span class="pre">ZhipuAIChatWrap
 <code class="docutils literal notranslate"><span class="pre">&quot;system&quot;</span></code> or <code class="docutils literal notranslate"><span class="pre">&quot;assistant&quot;</span></code>.</p></li>
 <li><p>There must be at least one <code class="docutils literal notranslate"><span class="pre">user</span></code> message.</p></li>
 </ul>
-<section id="id6">
-<h4>Prompt Strategy<a class="headerlink" href="#id6" title="Link to this heading"></a></h4>
+<section id="id7">
+<h4>Prompt Strategy<a class="headerlink" href="#id7" title="Link to this heading"></a></h4>
 <p>If the role field of the first message is <code class="docutils literal notranslate"><span class="pre">&quot;system&quot;</span></code>, it will be converted into a single message with the <code class="docutils literal notranslate"><span class="pre">role</span></code> field as <code class="docutils literal notranslate"><span class="pre">&quot;system&quot;</span></code> and the <code class="docutils literal notranslate"><span class="pre">content</span></code> field as the system message. The rest of the messages will be converted into a message with the <code class="docutils literal notranslate"><span class="pre">role</span></code> field as <code class="docutils literal notranslate"><span class="pre">&quot;user&quot;</span></code> and the <code class="docutils literal notranslate"><span class="pre">content</span></code> field as the dialogue history.</p>
 <p>An example is shown below:</p>
 <div class="highlight-python notranslate"><div class="highlight"><pre><span></span><span class="kn">from</span> <span class="nn">agentscope.models</span> <span class="kn">import</span> <span class="n">ZhipuAIChatWrapper</span>

zh_CN/_sources/tutorial/206-prompt.md.txt

@@ -271,6 +271,46 @@ print(prompt)
 ]
 ```
 
+### LiteLLMChatWrapper
+
+`LiteLLMChatWrapper`封装了litellm聊天API，它接受消息列表作为输入。Litellm支持不同类型的模型，每个模型可能需要遵守不同的格式。为了简化使用，我们提供了一种与大多数模型兼容的格式。如果需要更特定的格式，您可以参考您所使用的特定模型以及[litellm](https://github.com/BerriAI/litellm)文档，来定制适合您模型的格式函数。
+- 格式化聊天历史中的所有消息，将其整合成一个以`"user"`作为`role`的单一消息
+#### 提示策略
+- 消息将包括对话历史，`user`消息由系统消息(system message)和"## Dialog History"前缀。
+
+
+```python
+from agentscope.models import LiteLLMChatWrapper
+
+model = LiteLLMChatWrapper(
+    config_name="", # empty since we directly initialize the model wrapper
+    model_name="gpt-3.5-turbo",
+)
+
+prompt = model.format(
+  Msg("system", "You are a helpful assistant", role="system"),
+  [
+      Msg("user", "What is the weather today?", role="user"),
+      Msg("assistant", "It is sunny today", role="assistant"),
+  ],
+)
+
+print(prompt)
+```
+
+```bash
+[
+    {
+        "role": "user",
+        "content": (
+            "You are a helpful assistant\n\n"
+            "## Dialogue History\nuser: What is the weather today?\n"
+            "assistant: It is sunny today"
+        ),
+    },
+]
+```
+
 ### `OllamaChatWrapper`
 
 `OllamaChatWrapper`封装了Ollama聊天API，它接受消息列表作为输入。消息必须遵守以下规则(更新于2024/03/22)：