Update app.py

app.py

@@ -12,10 +12,22 @@ model.to("cpu")
 def generate_text(prompt):
     inputs = tokenizer(prompt, return_tensors="pt")
     inputs = {key: value.to("cpu") for key, value in inputs.items()}
-    outputs = model.generate(**inputs, max_length=100, do_sample=True, top_p=0.95, top_k=50)
-    generated_text = tokenizer.decode(outputs[0], skip_special_tokens=True)
+    
+    prompt_length = inputs["input_ids"].shape[1]
+    
+    outputs = model.generate(
+        **inputs,
+        max_length=prompt_length + 50,  
+        min_length=prompt_length + 1,  
+        do_sample=True,
+        top_p=0.95,
+        top_k=50
+    )
+    
+    generated_text = tokenizer.decode(outputs[0][prompt_length:], skip_special_tokens=True)
     return generated_text
 
+
 demo = gr.Interface(
     fn=generate_text,
     inputs="text",