Added extra model routes for if overloaded

app.py

@@ -4,11 +4,11 @@ import requests
 import pytz
 import yaml
 import os
-import uuid
 from datasets import Dataset
 from huggingface_hub import HfApi
 from openai import OpenAI
 from tools.final_answer import FinalAnswerTool
+from huggingface_hub import InferenceClient
 
 from Gradio_UI import GradioUI
 
@@ -318,6 +318,60 @@ model = HfApiModel(
     custom_role_conversions=None
 )
 
+# Add fallback logic that only activates if the primary model fails
+def try_model_call_with_fallbacks(prompt):
+    """Try to use the primary model first, fall back to alternatives if it fails."""
+    # First attempt with primary model
+    try:
+        return model(prompt)
+    except Exception as primary_error:
+        print(f"Primary model call failed: {str(primary_error)}")
+        print("Trying fallback models...")
+        
+        # List of fallback models
+        fallbacks = [
+            {
+                "provider": "sambanova",
+                "model_name": "Qwen/Qwen2.5-Coder-32B-Instruct",
+                "display_name": "Qwen 2.5 Coder 32B"
+            },
+            {
+                "provider": "hf-inference",
+                "model_name": "deepseek-ai/DeepSeek-R1-Distill-Qwen-32B",
+                "display_name": "DeepSeek R1 Distill Qwen 32B"
+            }
+        ]
+        
+        # Get API key
+        api_key = os.getenv("HF_API_KEY") or os.getenv("HUGGINGFACE_API_KEY")
+        if not api_key:
+            raise ValueError("No Hugging Face API key found in environment variables")
+        
+        # Try each fallback model in sequence
+        for fallback in fallbacks:
+            try:
+                print(f"Trying fallback model: {fallback['display_name']}")
+                client = InferenceClient(provider=fallback["provider"], api_key=api_key)
+                messages = [{"role": "user", "content": prompt}]
+                completion = client.chat.completions.create(
+                    model=fallback["model_name"],
+                    messages=messages,
+                    max_tokens=2096,
+                    temperature=0.5
+                )
+                print(f"Successfully used fallback model: {fallback['display_name']}")
+                return completion.choices[0].message.content
+            except Exception as e:
+                print(f"Fallback model {fallback['display_name']} failed: {str(e)}")
+                continue
+        
+        # If all fallbacks fail, re-raise the original error
+        raise primary_error
+
+# Monkey patch the model's __call__ method to use our fallback logic
+original_call = model.__call__
+model.__call__ = try_model_call_with_fallbacks
+
 # Import tool from Hub
 image_generation_tool = load_tool("agents-course/text-to-image", trust_remote_code=True)