RRemove old QwenModel

e2bqwen.py

@@ -411,7 +411,7 @@ class E2BVisionAgent(CodeAgent):
                 and previous_memory_step.step_number == current_step - 1
             ):
                 if previous_memory_step.tool_calls[0].arguments == memory_step.tool_calls[0].arguments:
-                    memory_step.observations += "\nWARNING: You've executed the same action several times in a row. MAKE SURE TO NOT USELESSLY REPEAT ACTIONS."
+                    memory_step.observations += "\nWARNING: You've executed the same action several times in a row. MAKE SURE TO NOT UNNECESSARILY REPEAT ACTIONS."
 
         # Add to the current memory step
         memory_step.observations_images = [image.copy()]
@@ -471,168 +471,3 @@ class QwenVLAPIModel(Model):
             return message
         except Exception as e:
             raise Exception(f"Both endpoints failed. Last error: {e}")
-
-# class QwenVLAPIModel(Model):
-#     """Model wrapper for Qwen2.5VL API with fallback mechanism"""
-    
-#     def __init__(
-#         self, 
-#         model_path: str = "Qwen/Qwen2.5-VL-72B-Instruct",
-#         provider: str = "hyperbolic",
-#         hf_token: str = None,
-#         hf_base_url: str = "https://n5wr7lfx6wp94tvl.us-east-1.aws.endpoints.huggingface.cloud"
-#     ):
-#         super().__init__()
-#         self.model_path = model_path
-#         self.model_id = model_path
-#         self.provider = provider
-#         self.hf_token = hf_token
-#         self.hf_base_url = hf_base_url
-        
-#         # Initialize hyperbolic client
-#         self.hyperbolic_client = InferenceClient(
-#             provider=self.provider,
-#         )
-
-#         assert not self.hf_base_url.endswith("/v1/"), "Enter your base url without '/v1/' suffix."
-
-#         # Initialize HF OpenAI-compatible client if token is provided
-#         self.hf_client = None
-#         from openai import OpenAI
-#         self.hf_client = OpenAI(
-#             base_url=self.hf_base_url + "/v1/",
-#             api_key=self.hf_token
-#         )
-        
-#     def __call__(
-#         self, 
-#         messages: List[Dict[str, Any]], 
-#         stop_sequences: Optional[List[str]] = None, 
-#         **kwargs
-#     ) -> ChatMessage:
-#         """Convert a list of messages to an API request with fallback mechanism"""
-        
-#         # Format messages once for both APIs
-#         formatted_messages = self._format_messages(messages)
-        
-#         # First try the HF endpoint if available - THIS ALWAYS FAILS SO SKIPPING
-#         try:
-#             completion = self._call_hf_endpoint(
-#                 formatted_messages, 
-#                 stop_sequences, 
-#                 **kwargs
-#             )
-#             print("SUCCESSFUL call of inference endpoint")
-#             return ChatMessage(role=MessageRole.ASSISTANT, content=completion)
-#         except Exception as e:
-#             print(f"HF endpoint failed with error: {e}. Falling back to hyperbolic.")
-#             # Continue to fallback
-        
-#         # Fallback to hyperbolic
-#         try:
-#             return self._call_hyperbolic(formatted_messages, stop_sequences, **kwargs)
-#         except Exception as e:
-#             raise Exception(f"Both endpoints failed. Last error: {e}")
-    
-#     def _format_messages(self, messages: List[Dict[str, Any]]):
-#         """Format messages for API requests - works for both endpoints"""
-        
-#         formatted_messages = []
-        
-#         for msg in messages:
-#             role = msg["role"]
-#             content = []
-            
-#             if isinstance(msg["content"], list):
-#                 for item in msg["content"]:
-#                     if item["type"] == "text":
-#                         content.append({"type": "text", "text": item["text"]})
-#                     elif item["type"] == "image":
-#                         # Handle image path or direct image object
-#                         if isinstance(item["image"], str):
-#                             # Image is a path
-#                             with open(item["image"], "rb") as image_file:
-#                                 base64_image = base64.b64encode(image_file.read()).decode("utf-8")
-#                         else:
-#                             # Image is a PIL image or similar object
-#                             img_byte_arr = BytesIO()
-#                             base64_image = base64.b64encode(img_byte_arr.getvalue()).decode("utf-8")
-
-#                         content.append({
-#                             "type": "image_url",
-#                             "image_url": {
-#                                 "url": f"data:image/png;base64,{base64_image}"
-#                             }
-#                         })
-#             else:
-#                 # Plain text message
-#                 content = [{"type": "text", "text": msg["content"]}]
-            
-#             formatted_messages.append({"role": role, "content": content})
-        
-#         return formatted_messages
-
-#     def _call_hf_endpoint(self, formatted_messages, stop_sequences=None, **kwargs):
-#         """Call the Hugging Face OpenAI-compatible endpoint"""
-
-#         # Extract parameters with defaults
-#         max_tokens = kwargs.get("max_new_tokens", 4096)
-#         temperature = kwargs.get("temperature", 0.7)
-#         top_p = kwargs.get("top_p", 0.9)
-#         stream = kwargs.get("stream", False)
-        
-#         completion = self.hf_client.chat.completions.create(
-#             model="tgi",  # Model name for the endpoint
-#             messages=formatted_messages,
-#             max_tokens=max_tokens,
-#             temperature=temperature,
-#             top_p=top_p,
-#             stream=stream,
-#             stop=stop_sequences
-#         )
-        
-#         if stream:
-#             # For streaming responses, return a generator
-#             def stream_generator():
-#                 for chunk in completion:
-#                     yield chunk.choices[0].delta.content or ""
-#             return stream_generator()
-#         else:
-#             # For non-streaming, return the full text
-#             return completion.choices[0].message.content
-    
-#     def _call_hyperbolic(self, formatted_messages, stop_sequences=None, **kwargs):
-#         """Call the hyperbolic API"""
-        
-#         completion = self.hyperbolic_client.chat.completions.create(
-#             model=self.model_path,
-#             messages=formatted_messages,
-#             max_tokens=kwargs.get("max_new_tokens", 4096),
-#             temperature=kwargs.get("temperature", 0.7),
-#             top_p=kwargs.get("top_p", 0.9),
-#             stop=stop_sequences
-#         )
-        
-#         # Extract the response text
-#         output_text = completion.choices[0].message.content
-        
-#         return ChatMessage(role=MessageRole.ASSISTANT, content=output_text)
-    
-#     def to_dict(self) -> Dict[str, Any]:
-#         """Convert the model to a dictionary"""
-#         return {
-#             "class": self.__class__.__name__,
-#             "model_path": self.model_path,
-#             "provider": self.provider,
-#             "hf_base_url": self.hf_base_url,
-#             # We don't save the API keys for security reasons
-#         }
-    
-#     @classmethod
-#     def from_dict(cls, data: Dict[str, Any]) -> "QwenVLAPIModel":
-#         """Create a model from a dictionary"""
-#         return cls(
-#             model_path=data.get("model_path", "Qwen/Qwen2.5-VL-72B-Instruct"),
-#             provider=data.get("provider", "hyperbolic"),
-#             hf_base_url=data.get("hf_base_url", "https://s41ydkv0iyjeokyj.us-east-1.aws.endpoints.huggingface.cloud"),
-#         )