Update app.py

app.py

@@ -7,7 +7,7 @@ tokenizer = AutoTokenizer.from_pretrained("ping98k/typhoon-7b-rag-instruct-th")
 model = AutoModelForCausalLM.from_pretrained("ping98k/typhoon-7b-rag-instruct-th", device_map={"": 0})
 
 @spaces.GPU(duration=120)
-def response(instruction, history, inputText):
+'''def response(instruction, history, inputText):
     inp = f"""### Instruction:
 {instruction}
 
@@ -18,7 +18,7 @@ def response(instruction, history, inputText):
 
 ### Response:"""
     print(f"Ask: {instruction}")
-
+    
     input_ids = tokenizer(inp, return_tensors='pt').to("cuda")
     beam_output = model.generate(**input_ids, max_new_tokens=300)
     outputText = tokenizer.decode(beam_output[0], skip_special_token=True)
@@ -38,4 +38,19 @@ gr.ChatInterface(
     additional_inputs=[
         gr.Textbox(defaultInput, label="Input Text"),
     ],
-).launch()
+).launch()'''
+
+def response(inp):
+    input_ids = tokenizer(inp, return_tensors='pt').to("cuda")
+    beam_output = model.generate(**input_ids, max_new_tokens=300)
+    outputText = tokenizer.decode(beam_output[0], skip_special_token=True)
+    outputText = outputText.replace(inp,"").replace("<s>","").replace("</s>","")
+
+    print(f"Ans: {outputText}")
+    
+    return outputText
+
+gradio_interface = gradio.Interface(
+  fn=response
+)
+gradio_interface.launch()