Update app.py

app.py

@@ -301,7 +301,7 @@ class Phi2EducationalLLM(Runnable):
                     model_path,
                     quantization_config=quant_config,
                     device_map="auto",
-                    torch_dtype=torch.float16,
+                    dtype=torch.float16,
                     trust_remote_code=True,
                     low_cpu_mem_usage=True
                 )
@@ -328,7 +328,7 @@ class Phi2EducationalLLM(Runnable):
         """Optimized model loading for 16GB RAM systems."""
         self.model = AutoModelForCausalLM.from_pretrained(
             model_path,
-            torch_dtype=torch.float16,  # Use float16 to save memory
+            dtype=torch.float16,  # Use float16 to save memory
             device_map="cpu",           # Force CPU for stability
             trust_remote_code=True,
             low_cpu_mem_usage=True,
@@ -345,7 +345,7 @@ class Phi2EducationalLLM(Runnable):
         
         self.model = AutoModelForCausalLM.from_pretrained(
             fallback_model,
-            torch_dtype=torch.float16 if torch.cuda.is_available() else torch.float32,
+            dtype=torch.float16 if torch.cuda.is_available() else torch.float32,
             device_map="cpu",
             trust_remote_code=True,
             low_cpu_mem_usage=True