Update app.py

app.py

@@ -315,7 +315,7 @@ class Phi3MiniEducationalLLM(Runnable):
                     model_path,
                     quantization_config=quant_config,
                     device_map="auto",
-                    dtype=torch.float16,
+                    torch_dtype=torch.float16,
                     trust_remote_code=True,
                     low_cpu_mem_usage=True,
                     token=hf_token
@@ -343,11 +343,12 @@ class Phi3MiniEducationalLLM(Runnable):
         """Optimized model loading for Phi-3-mini."""
         self.model = AutoModelForCausalLM.from_pretrained(
             model_path,
-            dtype=torch.float16,  # Use float16 to save memory
+            torch_dtype=torch.float16,  # Use float16 to save memory
             device_map="auto",          # Let transformers decide placement
             trust_remote_code=True,
             low_cpu_mem_usage=True,
-            token=hf_token
+            token=hf_token,
+            revision="0a67737cc96d2554230f90338b163bc6380a2a85"  # Pin revision for security
         )
 
     def _format_chat_template(self, prompt: str) -> str:
@@ -523,7 +524,7 @@ class Educational_Agent:
         start_init_and_langgraph_time = time.perf_counter()
         current_time = datetime.now()
         
-        self.llm = Phi3MiniEducationalLLM(model_path="microsoft/Phi-3-mini-4k-instruct", use_4bit=True)
+        self.llm = Phi3MiniEducationalLLM(model_path="microsoft/Phi-3-mini-4k-instruct", use_4bit=False)
         self.tool_decision_engine = Tool_Decision_Engine(self.llm)
         
         # Create LangGraph workflow