metadata

language:
  - es
  - en
  - multilingual
license: mit
tags:
  - codeswitching
  - spanish-english
  - sentiment-analysis
datasets:
  - lince

codeswitch-spaeng-sentiment-analysis-lince

This is a pretrained model for Sentiment Analysis of spanish-english code-mixed data used from LinCE

This model is trained for this below repository.

https://github.com/sagorbrur/codeswitch

To install codeswitch:

pip install codeswitch

Sentiment Analysis of Spanish-English Code-Mixed Data

Method-1


from transformers import AutoTokenizer, AutoModelForSequenceClassification, pipeline

tokenizer = AutoTokenizer.from_pretrained("sagorsarker/codeswitch-spaeng-sentiment-analysis-lince")

model = AutoModelForSequenceClassification.from_pretrained("sagorsarker/codeswitch-spaeng-sentiment-analysis-lince")

nlp = pipeline('sentiment-analysis', model=model, tokenizer=tokenizer)
sentence = "El perro le ladraba a La Gatita .. .. lol #teamlagatita en las playas de Key Biscayne este Memorial day"
nlp(sentence)

Method-2

from codeswitch.codeswitch import SentimentAnalysis
sa = SentimentAnalysis('spa-eng')
sentence = "El perro le ladraba a La Gatita .. .. lol #teamlagatita en las playas de Key Biscayne este Memorial day"
result = sa.analyze(sentence)
print(result)