Update app.py

app.py

@@ -1,4 +1,3 @@
-
 import gradio as gr
 from unsloth import FastLanguageModel
 import torch 
@@ -21,7 +20,7 @@ def predict(input_text):
         tokenize=True,
         add_generation_prompt=True,  
         return_tensors="pt",
-    ).to("cuda")
+    )  
 
     outputs = model.generate(
         input_ids=inputs,
@@ -30,7 +29,7 @@ def predict(input_text):
     )
     
     decoded_output = tokenizer.batch_decode(outputs, skip_special_tokens=True)
-    assistant_response = decoded_output[0].split('assistant')[-1].strip()
+    assistant_response = decoded_output[0].split('assistant\n')[-1].strip()
     
     return assistant_response
 
@@ -39,7 +38,7 @@ iface = gr.Interface(
     inputs="text",
     outputs="text",
     title="Parviz(eng) Chatbot",
-    description="A simple chatbot interface using FastLanguageModel.",
+    description="LLM Test model .",
 )
 
 iface.launch()