import streamlit as st from transformers import AutoTokenizer, AutoModelForSequenceClassification, Trainer, TextClassificationPipeline import operator import matplotlib.pyplot as plt import pandas as pd def get_sentiment(out): d = dict() for k in out: print(k) label = k['label'] score = k['score'] d[label] = score winning_lab = max(d.items(), key=operator.itemgetter(1))[0] winning_score = d[winning_lab] df = pd.DataFrame.from_dict(d, orient = 'index') return df #winning_lab, winning_score model_name = "mrm8488/distilroberta-finetuned-financial-news-sentiment-analysis" model = AutoModelForSequenceClassification.from_pretrained(model_name) tokenizer = AutoTokenizer.from_pretrained(model_name) pipe = TextClassificationPipeline(model=model, tokenizer=tokenizer, return_all_scores=True) text = st.text_area(f'Ciao! This app uses {model_name}.\nEnter your text to test it ❤️') if text: out = pipe(text) df = get_sentiment(out[0]) fig, ax = plt.subplots() c = ['#C34A36', '#FFC75F', '#008F7A'] ax.bar(df.index, df[0], color=c, width=0.4) st.pyplot(fig) #st.json(get_sentiment(out[0][0]))