XLM-T-Sent-Politics

This is an "extension" of the multilingual twitter-xlm-roberta-base-sentiment model (model, original paper) with a focus on sentiment from politicians' tweets. The original sentiment fine-tuning was done on 8 languages (Ar, En, Fr, De, Hi, It, Sp, Pt) but further training was done using tweets from Members of Parliament from UK (English), Spain (Spanish) and Greece (Greek).

• Reference Paper: Politics, Sentiment and Virality: A Large-Scale Multilingual Twitter Analysis in Greece, Spain and United Kingdom.
• Git Repo: https://github.com/cardiffnlp/politics-and-virality-twitter.

Full classification example

from transformers import AutoModelForSequenceClassification
from transformers import TFAutoModelForSequenceClassification
from transformers import AutoTokenizer
import numpy as np
from scipy.special import softmax

MODEL = f"cardiffnlp/xlm-twitter-politics-sentiment"

tokenizer = AutoTokenizer.from_pretrained(MODEL)

# PT
model = AutoModelForSequenceClassification.from_pretrained(MODEL)

text = "Good night 😊"
text = preprocess(text)
encoded_input = tokenizer(text, return_tensors='pt')
output = model(**encoded_input)
scores = output[0][0].detach().numpy()
scores = softmax(scores)

# # TF
# model = TFAutoModelForSequenceClassification.from_pretrained(MODEL)
# model.save_pretrained(MODEL)

# text = "Good night 😊"
# encoded_input = tokenizer(text, return_tensors='tf')
# output = model(encoded_input)
# scores = output[0][0].numpy()
# scores = softmax(scores)

# Print labels and scores
ranking = np.argsort(scores)
for i in range(scores.shape[0]):
    s = scores[ranking[i]]
    print(i, s)

Output:

0 0.0048229103
1 0.03117284
2 0.9640044