added max_new_tokens to translation

app.py

@@ -12,26 +12,26 @@ def main ():
     print(f"Running main")
 
     #print(subprocess.Popen('pip freeze > requirements_hug.txt', shell=True))
-    # text = "Tenho uma proposta para a Caixa Geral de Depositos, para 3 consultores outsystems, 300 euros por dia e um periodo de seis meses."
-    # st.write(text)
+    original = "Tenho uma proposta para a Caixa Geral de Depositos, para 3 consultores Outsystems, 300 euros por dia e um periodo de seis meses."
+    st.write(f"Original: {original}")
     # traducao = get_translation(text_to_translate=text, languageCode="pt")
     # st.write(traducao)
-    s2t = init_model_trans()
-    # ner = init_model_ner() #async
 
+    translation = translate(original) 
+    st.write(f"Translation: {translation}")  
     print("Rendering UI...")
     start_render = set_start()
     wav_audio_data = st_audiorec()
     audit_elapsedtime(function="Rendering UI", start=start_render)
 
-    if wav_audio_data is not None and s2t is not None:
+    if wav_audio_data is not None:
+        s2t = init_model_trans()
+        # ner = init_model_ner()
         print("Loading data...")
         start_loading = set_start()
         st.audio(wav_audio_data, format='audio/wav')
         original = transcribe(wav_audio_data, s2t)
         st.write(f"Original: {original}")
-        print("next is translation")
-        print("translating audio...")
         translation = translate(original)
         st.write(f"Transcription: {translation}")
 

translation.py

@@ -13,9 +13,10 @@ def translate(text_to_translate: str) -> str:
     
     start = set_start()
     print("Initiating translation model...")
+    text_size = len(text_to_translate)*2
     tokenizer = AutoTokenizer.from_pretrained("unicamp-dl/translation-pt-en-t5")
     model = AutoModelForSeq2SeqLM.from_pretrained("unicamp-dl/translation-pt-en-t5")
-    pten_pipeline = pipeline('text2text-generation', model=model, tokenizer=tokenizer)
+    pten_pipeline = pipeline('text2text-generation', kwargs={"max_new_tokens": text_size},model=model, tokenizer=tokenizer)
     translated_text = pten_pipeline(text_to_translate)
 
     audit_elapsedtime(function="Finished translation", start=start)