Update app.py

app.py

@@ -1,5 +1,5 @@
 import streamlit as st
-from transformers import pipeline, AutoModelForCausalLM, AutoTokenizer, GPT2Tokenizer, GPT2LMHeadModel, MT5Model, AutoModelForSeq2SeqLM, BartForConditionalGeneration
+from transformers import pipeline, AutoModelForCausalLM, AutoTokenizer, GPT2Tokenizer, GPT2LMHeadModel, MT5Model, AutoModelForSeq2SeqLM, BartForConditionalGeneration, BartTokenizer
 
 # Frasi di esempio
 examples = [
@@ -51,16 +51,16 @@ input_text = st.text_area(
 #model_name = "morenolq/bart-it"
 model_name = "Cicciokr/BART-la-s"
 model = BartForConditionalGeneration.from_pretrained(model_name)
-tokenizer = AutoTokenizer.from_pretrained(model_name, use_fast=False)
-tokenizer.pad_token_id = tokenizer.eos_token_id
+tokenizer = BartTokenizer.from_pretrained(model_name)
+#tokenizer.pad_token_id = tokenizer.eos_token_id
 
 model_name_mio = "Cicciokr/BART-CC100-la"
 model_mio = BartForConditionalGeneration.from_pretrained(model_name_mio)
-tokenizer_mio = AutoTokenizer.from_pretrained(model_name_mio, use_fast=False)
-tokenizer_mio.pad_token_id = tokenizer_mio.eos_token_id
+tokenizer_mio = BartTokenizer.from_pretrained(model_name_mio)
+#tokenizer_mio.pad_token_id = tokenizer_mio.eos_token_id
 
-generator = pipeline("text2text-generation", model=model, tokenizer=tokenizer)
-generator_mio = pipeline("text2text-generation", model=model_mio, tokenizer=tokenizer_mio)
+#generator = pipeline("text2text-generation", model=model, tokenizer=tokenizer)
+#generator_mio = pipeline("text2text-generation", model=model_mio, tokenizer=tokenizer_mio)
 
 # Se l'utente ha inserito (o selezionato) un testo
 if input_text:
@@ -73,19 +73,15 @@ if input_text:
     #generated_text = output[0]["generated_text"]
     if 'input_text_value_correct' in st.session_state:
         st.write(f" Parola corretta: {st.session_state['input_text_value_correct']}\n")
+        st.write(f" Parola corretta: {input_text}\n")
+        st.write(f" -----------------------------------------------------------\n")
     st.subheader("Risultato BART TheLatinLibrary:")
     st.write(f" Frase predetta: {generated_text}\n")
     #st.write(f" Frase predetta: {tokenizer.decode(output[0], skip_special_tokens=True)}\n")
     #print(output)
-
+    st.write(f" -----------------------------------------------------------\n")
     inputs_mio = tokenizer_mio(input_text, return_tensors="pt")
-    output_mio = model_mio.generate(
-        **inputs_mio,
-        max_length=64,
-        do_sample=True, 
-        num_beams=1,
-        forced_bos_token_id=tokenizer.bos_token_id
-    )
+    output_mio = model_mio.generate(**inputs_mio, forced_bos_token_id=tokenizer.bos_token_id)
     print(output_mio)
     generated_text_mio = tokenizer_mio.decode(output[0], skip_special_tokens=True)
     #generated_text_mio = output_mio[0]["generated_text"]