gefedya's picture
import streamlit as st
from datasets import load_dataset
from transformers import AutoModelForSequenceClassification, AutoTokenizer, AutoModel, Trainer, TrainingArguments, LineByLineTextDataset
# import json
import torch
def get_model():
model = AutoModelForSequenceClassification.from_pretrained("siebert/sentiment-roberta-large-english", num_labels=2)
model.load_state_dict(torch.load('model_cached_2.pth', map_location=torch.device('cpu')))
return model
def get_tokenizer():
tokenizer = AutoTokenizer.from_pretrained("siebert/sentiment-roberta-large-english")
return tokenizer
def make_prediction(to_analyze):
model = get_model()
tokenizer = get_tokenizer()
to_return = model(**tokenizer([to_analyze,], return_tensors='pt'))
return to_return
st.header("Sentiment analysis on twitter datasets")
st.markdown("Here is a sentiment model further trained on a slice of a twitter dataset")
# st.markdown("""
# <img width=700px src=''>
# """, unsafe_allow_html=True)
<img width=700px
src=''>""", unsafe_allow_html=True)
text = st.markdown("Try typing something here! \n You will see how much better our model is compared to the base model! No kidding")
# ^-- показать текстовое поле. В поле text лежит строка, которая находится там в данный момент
### Loading and tokenizing data
# data = load_dataset("carblacac/twitter-sentiment-analysis")
# tokenizer = AutoTokenizer.from_pretrained("siebert/sentiment-roberta-large-english")
# dataset = xs: tokenizer(xs["text"], truncation=True, padding='max_length'))
# dataset = dataset.rename_column("feeling", "labels")
with st.form(key='input_form'):
to_analyze = st.text_input(label='Input text to be analyzed')
button = st.form_submit_button(label='Analyze')
if button:
if to_analyze:
pred = make_prediction(to_analyze)
st.markdown("Negative" if torch.argmax(pred.logits).item() == 0 else "Positive")
st.markdown("Empty request. Please resubmit")
# classifier = pipeline('sentiment-analysis', model="distilbert-base-uncased-finetuned-sst-2-english")
# raw_predictions = classifier(text)
# тут уже знакомый вам код с huggingface.transformers -- его можно заменить на что угодно от fairseq до catboost