[Frontend] Added support for HF's new continue_final_message parameter (#8942)

vllm/entrypoints/openai/serving_chat.py

@@ -140,6 +140,7 @@ class OpenAIServingChat(OpenAIServing):
                     messages=request.messages,
                     chat_template=request.chat_template or self.chat_template,
                     add_generation_prompt=request.add_generation_prompt,
+                    continue_final_message=request.continue_final_message,
                     tools=tool_dicts,
                     documents=request.documents,
                     **(request.chat_template_kwargs or {}),
@@ -150,6 +151,7 @@ class OpenAIServingChat(OpenAIServing):
                     conversation=conversation,
                     chat_template=request.chat_template or self.chat_template,
                     add_generation_prompt=request.add_generation_prompt,
+                    continue_final_message=request.continue_final_message,
                     tools=tool_dicts,
                     documents=request.documents,
                     **(request.chat_template_kwargs or {}),
@@ -361,7 +363,7 @@ class OpenAIServingChat(OpenAIServing):
 
                     # Send response to echo the input portion of the
                     # last message
-                    if request.echo:
+                    if request.echo or request.continue_final_message:
                         last_msg_content: str = ""
                         if conversation and "content" in conversation[
                                 -1] and conversation[-1].get("role") == role:
@@ -716,7 +718,7 @@ class OpenAIServingChat(OpenAIServing):
                 stop_reason=output.stop_reason)
             choices.append(choice_data)
 
-        if request.echo:
+        if request.echo or request.continue_final_message:
             last_msg_content = ""
             if conversation and "content" in conversation[-1] and conversation[
                     -1].get("role") == role: