Update neuroprompt_deep.py

neuroprompt_deep.py

@@ -10,7 +10,7 @@ logging.basicConfig(level=logging.INFO)
 logger = logging.getLogger(__name__)
 
 class NeuroPromptDeep:
-    def __init__(self, model_name: str = "gpt2"):  # Changed to smaller GPT-2 model
+    def __init__(self, model_name: str = "gpt2"):
         """
         Initialize the generative AI engine with a local language model.
         """
@@ -74,20 +74,20 @@ class NeuroPromptDeep:
                 cache_dir=self.cache_dir
             )
             
-            # CPU-optimized configuration
+            # SIMPLIFIED CPU-ONLY CONFIGURATION
             self.model = AutoModelForCausalLM.from_pretrained(
                 self.model_name,
-                cache_dir=self.cache_dir,
-                device_map="auto",  # Will automatically use CPU
-                low_cpu_mem_usage=True  # Reduces memory footprint
+                cache_dir=self.cache_dir
             )
             
-            # CPU-based pipeline
+            # Explicitly move model to CPU
+            self.model = self.model.to('cpu')
+            
+            # CPU-based pipeline without device conflict
             self.generator = pipeline(
                 "text-generation",
                 model=self.model,
-                tokenizer=self.tokenizer,
-                device=-1  # Force CPU usage
+                tokenizer=self.tokenizer
             )
             
             logger.info("Model loaded successfully for CPU!")
@@ -117,7 +117,7 @@ class NeuroPromptDeep:
             # Format the full prompt
             full_prompt = f"{system_prompt}\n<|user|>\n{prompt}</s>\n<|assistant|>\n"
             
-            # Generate response
+            # Generate response with CPU-friendly settings
             outputs = self.generator(
                 full_prompt,
                 max_new_tokens=settings["max_length"],