Refactored base class config for llms (#3241)

2025-08-05 15:42:06 -07:00
parent f59ef3f2e2
commit 6e8c6c1cb7
25 changed files with 942 additions and 329 deletions
@@ -1,17 +1,37 @@
 import os
-from typing import Dict, List, Optional
+from typing import Dict, List, Optional, Union

 try:
    import anthropic
 except ImportError:
    raise ImportError("The 'anthropic' library is required. Please install it using 'pip install anthropic'.")

+from mem0.configs.llms.anthropic import AnthropicConfig
 from mem0.configs.llms.base import BaseLlmConfig
 from mem0.llms.base import LLMBase


 class AnthropicLLM(LLMBase):
-    def __init__(self, config: Optional[BaseLlmConfig] = None):
+    def __init__(self, config: Optional[Union[BaseLlmConfig, AnthropicConfig, Dict]] = None):
+        # Convert to AnthropicConfig if needed
+        if config is None:
+            config = AnthropicConfig()
+        elif isinstance(config, dict):
+            config = AnthropicConfig(**config)
+        elif isinstance(config, BaseLlmConfig) and not isinstance(config, AnthropicConfig):
+            # Convert BaseLlmConfig to AnthropicConfig
+            config = AnthropicConfig(
+                model=config.model,
+                temperature=config.temperature,
+                api_key=config.api_key,
+                max_tokens=config.max_tokens,
+                top_p=config.top_p,
+                top_k=config.top_k,
+                enable_vision=config.enable_vision,
+                vision_details=config.vision_details,
+                http_client_proxies=config.http_client,
+            )
+
        super().__init__(config)

        if not self.config.model:
@@ -26,6 +46,7 @@ class AnthropicLLM(LLMBase):
        response_format=None,
        tools: Optional[List[Dict]] = None,
        tool_choice: str = "auto",
+        **kwargs,
    ):
        """
        Generate a response based on the given messages using Anthropic.
@@ -35,6 +56,7 @@ class AnthropicLLM(LLMBase):
            response_format (str or object, optional): Format of the response. Defaults to "text".
            tools (list, optional): List of tools that the model can call. Defaults to None.
            tool_choice (str, optional): Tool choice method. Defaults to "auto".
+            **kwargs: Additional Anthropic-specific parameters.

        Returns:
            str: The generated response.
@@ -48,14 +70,16 @@ class AnthropicLLM(LLMBase):
            else:
                filtered_messages.append(message)

-        params = {
-            "model": self.config.model,
-            "messages": filtered_messages,
-            "system": system_message,
-            "temperature": self.config.temperature,
-            "max_tokens": self.config.max_tokens,
-            "top_p": self.config.top_p,
-        }
+        # Get common parameters
+        params = self._get_common_params(**kwargs)
+        params.update(
+            {
+                "model": self.config.model,
+                "messages": filtered_messages,
+                "system": system_message,
+            }
+        )
+
        if tools:  # TODO: Remove tools if no issues found with new memory addition logic
            params["tools"] = tools
            params["tool_choice"] = tool_choice
@@ -1,16 +1,36 @@
 import json
 import os
-from typing import Dict, List, Optional
+from typing import Dict, List, Optional, Union

 from openai import AzureOpenAI

+from mem0.configs.llms.azure import AzureOpenAIConfig
 from mem0.configs.llms.base import BaseLlmConfig
 from mem0.llms.base import LLMBase
 from mem0.memory.utils import extract_json


 class AzureOpenAILLM(LLMBase):
-    def __init__(self, config: Optional[BaseLlmConfig] = None):
+    def __init__(self, config: Optional[Union[BaseLlmConfig, AzureOpenAIConfig, Dict]] = None):
+        # Convert to AzureOpenAIConfig if needed
+        if config is None:
+            config = AzureOpenAIConfig()
+        elif isinstance(config, dict):
+            config = AzureOpenAIConfig(**config)
+        elif isinstance(config, BaseLlmConfig) and not isinstance(config, AzureOpenAIConfig):
+            # Convert BaseLlmConfig to AzureOpenAIConfig
+            config = AzureOpenAIConfig(
+                model=config.model,
+                temperature=config.temperature,
+                api_key=config.api_key,
+                max_tokens=config.max_tokens,
+                top_p=config.top_p,
+                top_k=config.top_k,
+                enable_vision=config.enable_vision,
+                vision_details=config.vision_details,
+                http_client_proxies=config.http_client,
+            )
+
        super().__init__(config)

        # Model name should match the custom deployment name chosen for it.
@@ -68,6 +88,7 @@ class AzureOpenAILLM(LLMBase):
        response_format=None,
        tools: Optional[List[Dict]] = None,
        tool_choice: str = "auto",
+        **kwargs,
    ):
        """
        Generate a response based on the given messages using Azure OpenAI.
@@ -77,6 +98,7 @@ class AzureOpenAILLM(LLMBase):
            response_format (str or object, optional): Format of the response. Defaults to "text".
            tools (list, optional): List of tools that the model can call. Defaults to None.
            tool_choice (str, optional): Tool choice method. Defaults to "auto".
+            **kwargs: Additional Azure OpenAI-specific parameters.

        Returns:
            str: The generated response.
@@ -88,23 +110,29 @@ class AzureOpenAILLM(LLMBase):

        messages[-1]["content"] = user_prompt

-        common_params = {
-            "model": self.config.model,
-            "messages": messages,
-        }
+        # Get common parameters
+        params = self._get_common_params(**kwargs)
+        params.update(
+            {
+                "model": self.config.model,
+                "messages": messages,
+            }
+        )

        if self.config.model in {"o3-mini", "o1-preview", "o1"}:
-            params = common_params
+            # Use common params for these models
+            pass
        else:
-            params = {
-                **common_params,
-                "temperature": self.config.temperature,
-                "max_tokens": self.config.max_tokens,
-                "top_p": self.config.top_p,
-            }
-        if response_format:
-            params["response_format"] = response_format
-        if tools:  # TODO: Remove tools if no issues found with new memory addition logic
+            # Add additional parameters for other models
+            params.update(
+                {
+                    "temperature": self.config.temperature,
+                    "max_tokens": self.config.max_tokens,
+                    "top_p": self.config.top_p,
+                }
+            )
+
+        if tools:
            params["tools"] = tools
            params["tool_choice"] = tool_choice

@@ -1,23 +1,49 @@
 from abc import ABC, abstractmethod
-from typing import Dict, List, Optional
+from typing import Dict, List, Optional, Union

 from mem0.configs.llms.base import BaseLlmConfig


 class LLMBase(ABC):
-    def __init__(self, config: Optional[BaseLlmConfig] = None):
+    """
+    Base class for all LLM providers.
+    Handles common functionality and delegates provider-specific logic to subclasses.
+    """
+
+    def __init__(self, config: Optional[Union[BaseLlmConfig, Dict]] = None):
        """Initialize a base LLM class

-        :param config: LLM configuration option class, defaults to None
-        :type config: Optional[BaseLlmConfig], optional
+        :param config: LLM configuration option class or dict, defaults to None
+        :type config: Optional[Union[BaseLlmConfig, Dict]], optional
        """
        if config is None:
            self.config = BaseLlmConfig()
+        elif isinstance(config, dict):
+            # Handle dict-based configuration (backward compatibility)
+            self.config = BaseLlmConfig(**config)
        else:
            self.config = config

+        # Validate configuration
+        self._validate_config()
+
+    def _validate_config(self):
+        """
+        Validate the configuration.
+        Override in subclasses to add provider-specific validation.
+        """
+        if not hasattr(self.config, "model"):
+            raise ValueError("Configuration must have a 'model' attribute")
+
+        if not hasattr(self.config, "api_key") and not hasattr(self.config, "api_key"):
+            # Check if API key is available via environment variable
+            # This will be handled by individual providers
+            pass
+
    @abstractmethod
-    def generate_response(self, messages, tools: Optional[List[Dict]] = None, tool_choice: str = "auto"):
+    def generate_response(
+        self, messages: List[Dict[str, str]], tools: Optional[List[Dict]] = None, tool_choice: str = "auto", **kwargs
+    ):
        """
        Generate a response based on the given messages.

@@ -25,8 +51,27 @@ class LLMBase(ABC):
            messages (list): List of message dicts containing 'role' and 'content'.
            tools (list, optional): List of tools that the model can call. Defaults to None.
            tool_choice (str, optional): Tool choice method. Defaults to "auto".
+            **kwargs: Additional provider-specific parameters.

        Returns:
-            str: The generated response.
+            str or dict: The generated response.
        """
        pass
+
+    def _get_common_params(self, **kwargs) -> Dict:
+        """
+        Get common parameters that most providers use.
+
+        Returns:
+            Dict: Common parameters dictionary.
+        """
+        params = {
+            "temperature": self.config.temperature,
+            "max_tokens": self.config.max_tokens,
+            "top_p": self.config.top_p,
+        }
+
+        # Add provider-specific parameters from kwargs
+        params.update(kwargs)
+
+        return params
@@ -1,16 +1,36 @@
 import json
 import os
-from typing import Dict, List, Optional
+from typing import Dict, List, Optional, Union

 from openai import OpenAI

 from mem0.configs.llms.base import BaseLlmConfig
+from mem0.configs.llms.deepseek import DeepSeekConfig
 from mem0.llms.base import LLMBase
 from mem0.memory.utils import extract_json


 class DeepSeekLLM(LLMBase):
-    def __init__(self, config: Optional[BaseLlmConfig] = None):
+    def __init__(self, config: Optional[Union[BaseLlmConfig, DeepSeekConfig, Dict]] = None):
+        # Convert to DeepSeekConfig if needed
+        if config is None:
+            config = DeepSeekConfig()
+        elif isinstance(config, dict):
+            config = DeepSeekConfig(**config)
+        elif isinstance(config, BaseLlmConfig) and not isinstance(config, DeepSeekConfig):
+            # Convert BaseLlmConfig to DeepSeekConfig
+            config = DeepSeekConfig(
+                model=config.model,
+                temperature=config.temperature,
+                api_key=config.api_key,
+                max_tokens=config.max_tokens,
+                top_p=config.top_p,
+                top_k=config.top_k,
+                enable_vision=config.enable_vision,
+                vision_details=config.vision_details,
+                http_client_proxies=config.http_client,
+            )
+
        super().__init__(config)

        if not self.config.model:
@@ -56,6 +76,7 @@ class DeepSeekLLM(LLMBase):
        response_format=None,
        tools: Optional[List[Dict]] = None,
        tool_choice: str = "auto",
+        **kwargs,
    ):
        """
        Generate a response based on the given messages using DeepSeek.
@@ -65,17 +86,19 @@ class DeepSeekLLM(LLMBase):
            response_format (str or object, optional): Format of the response. Defaults to "text".
            tools (list, optional): List of tools that the model can call. Defaults to None.
            tool_choice (str, optional): Tool choice method. Defaults to "auto".
+            **kwargs: Additional DeepSeek-specific parameters.

        Returns:
            str: The generated response.
        """
-        params = {
-            "model": self.config.model,
-            "messages": messages,
-            "temperature": self.config.temperature,
-            "max_tokens": self.config.max_tokens,
-            "top_p": self.config.top_p,
-        }
+        # Get common parameters
+        params = self._get_common_params(**kwargs)
+        params.update(
+            {
+                "model": self.config.model,
+                "messages": messages,
+            }
+        )

        if tools:
            params["tools"] = tools
@@ -1,13 +1,35 @@
-from typing import Dict, List, Optional
+import json
+from typing import Dict, List, Optional, Union

 from openai import OpenAI

 from mem0.configs.llms.base import BaseLlmConfig
+from mem0.configs.llms.lmstudio import LMStudioConfig
 from mem0.llms.base import LLMBase
+from mem0.memory.utils import extract_json


 class LMStudioLLM(LLMBase):
-    def __init__(self, config: Optional[BaseLlmConfig] = None):
+    def __init__(self, config: Optional[Union[BaseLlmConfig, LMStudioConfig, Dict]] = None):
+        # Convert to LMStudioConfig if needed
+        if config is None:
+            config = LMStudioConfig()
+        elif isinstance(config, dict):
+            config = LMStudioConfig(**config)
+        elif isinstance(config, BaseLlmConfig) and not isinstance(config, LMStudioConfig):
+            # Convert BaseLlmConfig to LMStudioConfig
+            config = LMStudioConfig(
+                model=config.model,
+                temperature=config.temperature,
+                api_key=config.api_key,
+                max_tokens=config.max_tokens,
+                top_p=config.top_p,
+                top_k=config.top_k,
+                enable_vision=config.enable_vision,
+                vision_details=config.vision_details,
+                http_client_proxies=config.http_client,
+            )
+
        super().__init__(config)

        self.config.model = (
@@ -18,12 +40,43 @@ class LMStudioLLM(LLMBase):

        self.client = OpenAI(base_url=self.config.lmstudio_base_url, api_key=self.config.api_key)

+    def _parse_response(self, response, tools):
+        """
+        Process the response based on whether tools are used or not.
+
+        Args:
+            response: The raw response from API.
+            tools: The list of tools provided in the request.
+
+        Returns:
+            str or dict: The processed response.
+        """
+        if tools:
+            processed_response = {
+                "content": response.choices[0].message.content,
+                "tool_calls": [],
+            }
+
+            if response.choices[0].message.tool_calls:
+                for tool_call in response.choices[0].message.tool_calls:
+                    processed_response["tool_calls"].append(
+                        {
+                            "name": tool_call.function.name,
+                            "arguments": json.loads(extract_json(tool_call.function.arguments)),
+                        }
+                    )
+
+            return processed_response
+        else:
+            return response.choices[0].message.content
+
    def generate_response(
        self,
        messages: List[Dict[str, str]],
-        response_format: dict = {"type": "json_object"},
+        response_format=None,
        tools: Optional[List[Dict]] = None,
        tool_choice: str = "auto",
+        **kwargs,
    ):
        """
        Generate a response based on the given messages using LM Studio.
@@ -33,21 +86,32 @@ class LMStudioLLM(LLMBase):
            response_format (str or object, optional): Format of the response. Defaults to "text".
            tools (list, optional): List of tools that the model can call. Defaults to None.
            tool_choice (str, optional): Tool choice method. Defaults to "auto".
+            **kwargs: Additional LM Studio-specific parameters.

        Returns:
            str: The generated response.
        """
-        params = {
-            "model": self.config.model,
-            "messages": messages,
-            "temperature": self.config.temperature,
-            "max_tokens": self.config.max_tokens,
-            "top_p": self.config.top_p,
-        }
-        if response_format:
-            params["response_format"] = response_format
-        if self.config.lmstudio_response_format is not None:
+        # Get common parameters
+        params = self._get_common_params(**kwargs)
+        params.update(
+            {
+                "model": self.config.model,
+                "messages": messages,
+            }
+        )
+
+        # Handle response format - LM Studio defaults to json_object
+        if self.config.lmstudio_response_format:
            params["response_format"] = self.config.lmstudio_response_format
+        elif response_format:
+            params["response_format"] = response_format
+        else:
+            # Default to json_object for LM Studio
+            params["response_format"] = {"type": "json_object"}
+
+        if tools:
+            params["tools"] = tools
+            params["tool_choice"] = tool_choice

        response = self.client.chat.completions.create(**params)
-        return response.choices[0].message.content
+        return self._parse_response(response, tools)
@@ -1,4 +1,4 @@
-from typing import Dict, List, Optional
+from typing import Dict, List, Optional, Union

 try:
    from ollama import Client
@@ -6,25 +6,37 @@ except ImportError:
    raise ImportError("The 'ollama' library is required. Please install it using 'pip install ollama'.")

 from mem0.configs.llms.base import BaseLlmConfig
+from mem0.configs.llms.ollama import OllamaConfig
 from mem0.llms.base import LLMBase


 class OllamaLLM(LLMBase):
-    def __init__(self, config: Optional[BaseLlmConfig] = None):
+    def __init__(self, config: Optional[Union[BaseLlmConfig, OllamaConfig, Dict]] = None):
+        # Convert to OllamaConfig if needed
+        if config is None:
+            config = OllamaConfig()
+        elif isinstance(config, dict):
+            config = OllamaConfig(**config)
+        elif isinstance(config, BaseLlmConfig) and not isinstance(config, OllamaConfig):
+            # Convert BaseLlmConfig to OllamaConfig
+            config = OllamaConfig(
+                model=config.model,
+                temperature=config.temperature,
+                api_key=config.api_key,
+                max_tokens=config.max_tokens,
+                top_p=config.top_p,
+                top_k=config.top_k,
+                enable_vision=config.enable_vision,
+                vision_details=config.vision_details,
+                http_client_proxies=config.http_client,
+            )
+
        super().__init__(config)

        if not self.config.model:
            self.config.model = "llama3.1:70b"
-        self.client = Client(host=self.config.ollama_base_url)
-        self._ensure_model_exists()

-    def _ensure_model_exists(self):
-        """
-        Ensure the specified model exists locally. If not, pull it from Ollama.
-        """
-        local_models = self.client.list()["models"]
-        if not any(model.get("name") == self.config.model for model in local_models):
-            self.client.pull(self.config.model)
+        self.client = Client(host=self.config.ollama_base_url)

    def _parse_response(self, response, tools):
        """
@@ -39,22 +51,18 @@ class OllamaLLM(LLMBase):
        """
        if tools:
            processed_response = {
-                "content": response["message"]["content"],
+                "content": response["message"]["content"] if isinstance(response, dict) else response.message.content,
                "tool_calls": [],
            }

-            if response["message"].get("tool_calls"):
-                for tool_call in response["message"]["tool_calls"]:
-                    processed_response["tool_calls"].append(
-                        {
-                            "name": tool_call["function"]["name"],
-                            "arguments": tool_call["function"]["arguments"],
-                        }
-                    )
-
+            # Ollama doesn't support tool calls in the same way, so we return the content
            return processed_response
        else:
-            return response["message"]["content"]
+            # Handle both dict and object responses
+            if isinstance(response, dict):
+                return response["message"]["content"]
+            else:
+                return response.message.content

    def generate_response(
        self,
@@ -62,33 +70,37 @@ class OllamaLLM(LLMBase):
        response_format=None,
        tools: Optional[List[Dict]] = None,
        tool_choice: str = "auto",
+        **kwargs,
    ):
        """
-        Generate a response based on the given messages using OpenAI.
+        Generate a response based on the given messages using Ollama.

        Args:
            messages (list): List of message dicts containing 'role' and 'content'.
            response_format (str or object, optional): Format of the response. Defaults to "text".
            tools (list, optional): List of tools that the model can call. Defaults to None.
            tool_choice (str, optional): Tool choice method. Defaults to "auto".
+            **kwargs: Additional Ollama-specific parameters.

        Returns:
            str: The generated response.
        """
+        # Build parameters for Ollama
        params = {
            "model": self.config.model,
            "messages": messages,
-            "options": {
-                "temperature": self.config.temperature,
-                "num_predict": self.config.max_tokens,
-                "top_p": self.config.top_p,
-            },
        }
-        if response_format:
-            params["format"] = "json"

-        if tools:
-            params["tools"] = tools
+        # Add options for Ollama (temperature, num_predict, top_p)
+        options = {
+            "temperature": self.config.temperature,
+            "num_predict": self.config.max_tokens,
+            "top_p": self.config.top_p,
+        }
+        params["options"] = options
+
+        # Remove OpenAI-specific parameters that Ollama doesn't support
+        params.pop("max_tokens", None)  # Ollama uses different parameter names

        response = self.client.chat(**params)
        return self._parse_response(response, tools)
@@ -1,17 +1,36 @@
 import json
 import os
-import warnings
-from typing import Dict, List, Optional
+from typing import Dict, List, Optional, Union

 from openai import OpenAI

 from mem0.configs.llms.base import BaseLlmConfig
+from mem0.configs.llms.openai import OpenAIConfig
 from mem0.llms.base import LLMBase
 from mem0.memory.utils import extract_json


 class OpenAILLM(LLMBase):
-    def __init__(self, config: Optional[BaseLlmConfig] = None):
+    def __init__(self, config: Optional[Union[BaseLlmConfig, OpenAIConfig, Dict]] = None):
+        # Convert to OpenAIConfig if needed
+        if config is None:
+            config = OpenAIConfig()
+        elif isinstance(config, dict):
+            config = OpenAIConfig(**config)
+        elif isinstance(config, BaseLlmConfig) and not isinstance(config, OpenAIConfig):
+            # Convert BaseLlmConfig to OpenAIConfig
+            config = OpenAIConfig(
+                model=config.model,
+                temperature=config.temperature,
+                api_key=config.api_key,
+                max_tokens=config.max_tokens,
+                top_p=config.top_p,
+                top_k=config.top_k,
+                enable_vision=config.enable_vision,
+                vision_details=config.vision_details,
+                http_client_proxies=config.http_client,
+            )
+
        super().__init__(config)

        if not self.config.model:
@@ -26,18 +45,7 @@ class OpenAILLM(LLMBase):
            )
        else:
            api_key = self.config.api_key or os.getenv("OPENAI_API_KEY")
-            base_url = (
-                self.config.openai_base_url
-                or os.getenv("OPENAI_API_BASE")
-                or os.getenv("OPENAI_BASE_URL")
-                or "https://api.openai.com/v1"
-            )
-            if os.environ.get("OPENAI_API_BASE"):
-                warnings.warn(
-                    "The environment variable 'OPENAI_API_BASE' is deprecated and will be removed in the 0.1.80. "
-                    "Please use 'OPENAI_BASE_URL' instead.",
-                    DeprecationWarning,
-                )
+            base_url = self.config.openai_base_url or os.getenv("OPENAI_BASE_URL") or "https://api.openai.com/v1"

            self.client = OpenAI(api_key=api_key, base_url=base_url)

@@ -77,6 +85,7 @@ class OpenAILLM(LLMBase):
        response_format=None,
        tools: Optional[List[Dict]] = None,
        tool_choice: str = "auto",
+        **kwargs,
    ):
        """
        Generate a JSON response based on the given messages using OpenAI.
@@ -86,17 +95,19 @@ class OpenAILLM(LLMBase):
            response_format (str or object, optional): Format of the response. Defaults to "text".
            tools (list, optional): List of tools that the model can call. Defaults to None.
            tool_choice (str, optional): Tool choice method. Defaults to "auto".
+            **kwargs: Additional OpenAI-specific parameters.

        Returns:
            json: The generated response.
        """
-        params = {
-            "model": self.config.model,
-            "messages": messages,
-            "temperature": self.config.temperature,
-            "max_tokens": self.config.max_tokens,
-            "top_p": self.config.top_p,
-        }
+        # Get common parameters
+        params = self._get_common_params(**kwargs)
+        params.update(
+            {
+                "model": self.config.model,
+                "messages": messages,
+            }
+        )

        if os.getenv("OPENROUTER_API_KEY"):
            openrouter_params = {}
@@ -1,16 +1,36 @@
 import json
 import os
-from typing import Dict, List, Optional
+from typing import Dict, List, Optional, Union

 from openai import OpenAI

 from mem0.configs.llms.base import BaseLlmConfig
+from mem0.configs.llms.vllm import VllmConfig
 from mem0.llms.base import LLMBase
 from mem0.memory.utils import extract_json


 class VllmLLM(LLMBase):
-    def __init__(self, config: Optional[BaseLlmConfig] = None):
+    def __init__(self, config: Optional[Union[BaseLlmConfig, VllmConfig, Dict]] = None):
+        # Convert to VllmConfig if needed
+        if config is None:
+            config = VllmConfig()
+        elif isinstance(config, dict):
+            config = VllmConfig(**config)
+        elif isinstance(config, BaseLlmConfig) and not isinstance(config, VllmConfig):
+            # Convert BaseLlmConfig to VllmConfig
+            config = VllmConfig(
+                model=config.model,
+                temperature=config.temperature,
+                api_key=config.api_key,
+                max_tokens=config.max_tokens,
+                top_p=config.top_p,
+                top_k=config.top_k,
+                enable_vision=config.enable_vision,
+                vision_details=config.vision_details,
+                http_client_proxies=config.http_client,
+            )
+
        super().__init__(config)

        if not self.config.model:
@@ -18,8 +38,7 @@ class VllmLLM(LLMBase):

        self.config.api_key = self.config.api_key or os.getenv("VLLM_API_KEY") or "vllm-api-key"
        base_url = self.config.vllm_base_url or os.getenv("VLLM_BASE_URL")
-
-        self.client = OpenAI(base_url=base_url, api_key=self.config.api_key)
+        self.client = OpenAI(api_key=self.config.api_key, base_url=base_url)

    def _parse_response(self, response, tools):
        """
@@ -57,6 +76,7 @@ class VllmLLM(LLMBase):
        response_format=None,
        tools: Optional[List[Dict]] = None,
        tool_choice: str = "auto",
+        **kwargs,
    ):
        """
        Generate a response based on the given messages using vLLM.
@@ -66,20 +86,19 @@ class VllmLLM(LLMBase):
            response_format (str or object, optional): Format of the response. Defaults to "text".
            tools (list, optional): List of tools that the model can call. Defaults to None.
            tool_choice (str, optional): Tool choice method. Defaults to "auto".
+            **kwargs: Additional vLLM-specific parameters.

        Returns:
            str: The generated response.
        """
-        params = {
-            "model": self.config.model,
-            "messages": messages,
-            "temperature": self.config.temperature,
-            "max_tokens": self.config.max_tokens,
-            "top_p": self.config.top_p,
-        }
-
-        if response_format:
-            params["response_format"] = response_format
+        # Get common parameters
+        params = self._get_common_params(**kwargs)
+        params.update(
+            {
+                "model": self.config.model,
+                "messages": messages,
+            }
+        )

        if tools:
            params["tools"] = tools