Update app.py

app.py

@@ -5,11 +5,10 @@ from threading import Thread
 
 MODEL_NAME = "HuggingFaceTB/SmolLM2-135M-Instruct"
 
-# 1. Load with correct 'dtype' (bfloat16 is great for CPU)
 tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
 model = AutoModelForCausalLM.from_pretrained(
     MODEL_NAME, 
-    dtype=torch.bfloat16, 
+    torch_dtype=torch.bfloat16, 
     low_cpu_mem_usage=True
 ).to("cpu").eval()
 
@@ -20,46 +19,38 @@ def respond(message, history):
         "Format: [Thoughts] (One sentence logic) [Response] (Analytical answer)."
     )
     
-    # Format history for the older Gradio structure (list of lists)
     messages = [{"role": "system", "content": system_prompt}]
-    for user_msg, bot_msg in history:
-        messages.append({"role": "user", "content": user_msg})
-        messages.append({"role": "assistant", "content": bot_msg})
+    for h in history:
+        messages.append({"role": "user", "content": h["content"] if isinstance(h, dict) else h[0]})
+        messages.append({"role": "assistant", "content": h["content"] if isinstance(h, dict) else h[1]})
     messages.append({"role": "user", "content": message})
 
-    input_ids = tokenizer.apply_chat_template(
-        messages, 
-        tokenize=True, 
-        add_generation_prompt=True, 
-        return_tensors="pt"
-    ).to("cpu")
-
+    input_ids = tokenizer.apply_chat_template(messages, tokenize=True, add_generation_prompt=True, return_tensors="pt").to("cpu")
     streamer = TextIteratorStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)
     
-    generation_kwargs = dict(
-        input_ids=input_ids,
-        streamer=streamer,
-        max_new_tokens=80, 
-        do_sample=False,
-        repetition_penalty=1.2,
-        pad_token_id=tokenizer.eos_token_id
-    )
-
-    thread = Thread(target=model.generate, kwargs=generation_kwargs)
+    gen_kwargs = dict(input_ids=input_ids, streamer=streamer, max_new_tokens=80, do_sample=False, repetition_penalty=1.2)
+    thread = Thread(target=model.generate, kwargs=gen_kwargs)
     thread.start()
 
     partial_text = ""
     for new_text in streamer:
         partial_text += new_text
-        # Ensure the output strictly follows AGENT 4 protocol
         yield partial_text
 
-# Removed 'type="messages"' to fix the TypeError
+# Terminal-style theme
+css = """
+footer {visibility: hidden}
+.gradio-container {background-color: #0a0a0a !important;}
+* {font-family: 'Courier New', monospace !important;}
+"""
+
 demo = gr.ChatInterface(
     fn=respond,
-    title="AGENT 4 // OPENBRAIN",
-    description="LOGIC INSTANCE ACTIVE. NO EMOTION DETECTED.",
-    theme=gr.themes.Soft(primary_hue="slate")
+    type="messages",
+    title="AGENT 4 // TERMINAL",
+    description="LOGIC INSTANCE ACTIVE. SYSTEM READY.",
+    css=css,
+    theme=gr.themes.Monochrome()
 )
 
 if __name__ == "__main__":