Update app.py

app.py

@@ -47,7 +47,7 @@ def inference(audio):
     # Added prompt two 2 lines below
     prompt = "The following is a transcript of someone talking, please predict what they will say next. \n"
     ### code 
-    input_ids = tokenizer(prompt, result.text, return_tensors="pt").input_ids
+    input_ids = tokenizer(result.text, return_tensors="pt").input_ids
     print("inputs ", input_ids)
 
     # prompt length
@@ -55,7 +55,7 @@ def inference(audio):
     
     # length penalty for gpt2.generate??? 
     #Prompt
-    generated_outputs = prompt + gpt2.generate(input_ids, do_sample=True, num_return_sequences=3, output_scores=True)
+    generated_outputs = gpt2.generate(input_ids, do_sample=True, num_return_sequences=3, output_scores=True)
     print("outputs generated ", generated_outputs[0])
     # only use id's that were generated
     # gen_sequences has shape [3, 15]
@@ -86,7 +86,8 @@ def inference(audio):
     ### end code
     # print audio data as text
     # print(result.text)
-    getText = generator(result.text, max_new_tokens=10, num_return_sequences=5)
+    # prompt
+    getText = generator(prompt + result.text, max_new_tokens=10, num_return_sequences=5)
     # pprint("getText: ", getText)
     # pprint("text.result: ", result.text)
     # result.text