Update app.py

app.py

@@ -5,29 +5,41 @@ import torch
 MODEL_ID = "Qwen/Qwen2.5-Coder-3B-Instruct"
 
 tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)
+
 model = AutoModelForCausalLM.from_pretrained(
     MODEL_ID,
-    torch_dtype=torch.float32,
     device_map="cpu"
 )
 
-def generate(prompt, max_tokens=256, temperature=0.2):
-    inputs = tokenizer(prompt, return_tensors="pt")
-    outputs = model.generate(
-        **inputs,
-        max_new_tokens=max_tokens,
-        temperature=temperature,
-        do_sample=True
+# IMPORTANT: set pad token
+if tokenizer.pad_token_id is None:
+    tokenizer.pad_token = tokenizer.eos_token
+
+def generate(prompt: str):
+    if not prompt.strip():
+        return "⚠️ Empty prompt"
+
+    inputs = tokenizer(
+        prompt,
+        return_tensors="pt",
+        padding=True
     )
+
+    with torch.no_grad():
+        outputs = model.generate(
+            input_ids=inputs["input_ids"],
+            attention_mask=inputs["attention_mask"],
+            max_new_tokens=128,
+            do_sample=False,
+            use_cache=False,          # 🔥 THIS AVOIDS THE BUG
+            pad_token_id=tokenizer.eos_token_id
+        )
+
     return tokenizer.decode(outputs[0], skip_special_tokens=True)
 
 gr.Interface(
     fn=generate,
-    inputs=[
-        gr.Textbox(label="Prompt"),
-        gr.Slider(64, 512, value=256, label="Max tokens"),
-        gr.Slider(0.0, 1.0, value=0.2, label="Temperature")
-    ],
-    outputs=gr.Textbox(label="Response"),
-    title="Qwen 2.5 Coder (Free CPU)"
+    inputs=gr.Textbox(lines=8, placeholder="Enter prompt"),
+    outputs=gr.Textbox(lines=12),
+    title="Qwen 2.5 Coder – Stable CPU Mode"
 ).launch()