BERT codemixed base model for spanglish (cased)

This model was built using lingualytics, an open-source library that supports code-mixed analytics.

Model description

Input for the model: Any codemixed spanglish text Output for the model: Sentiment. (0 - Negative, 1 - Neutral, 2 - Positive)

I took a bert-base-multilingual-cased model from Huggingface and finetuned it on CS-EN-ES-CORPUS dataset.

Performance of this model on the dataset

metric score
acc 0.718615
f1 0.71759
acc_and_f1 0.718103
precision 0.719302
recall 0.718615

Intended uses & limitations

Make sure to preprocess your data using these methods before using this model.

How to use

Here is how to use this model to get the features of a given text in PyTorch:

# You can include sample code which will be formatted
from transformers import BertTokenizer, BertModelForSequenceClassification
tokenizer = AutoTokenizer.from_pretrained('rohanrajpal/bert-base-en-es-codemix-cased')
model = AutoModelForSequenceClassification.from_pretrained('rohanrajpal/bert-base-en-es-codemix-cased')
text = "Replace me by any text you'd like."
encoded_input = tokenizer(text, return_tensors='pt')
output = model(**encoded_input)

and in TensorFlow:

from transformers import BertTokenizer, TFBertModel
tokenizer = BertTokenizer.from_pretrained('rohanrajpal/bert-base-en-es-codemix-cased')
model = TFBertModel.from_pretrained('rohanrajpal/bert-base-en-es-codemix-cased')
text = "Replace me by any text you'd like."
encoded_input = tokenizer(text, return_tensors='tf')
output = model(encoded_input)

Limitations and bias

Since I dont know spanish, I cant verify the quality of annotations or the dataset itself. This is a very simple transfer learning approach and I'm open to discussions to improve upon this.

Training data

I trained on the dataset on the bert-base-multilingual-cased model.

Training procedure

Followed the preprocessing techniques followed here

Eval results

BibTeX entry and citation info

    title = "{GLUEC}o{S}: An Evaluation Benchmark for Code-Switched {NLP}",
    author = "Khanuja, Simran  and
      Dandapat, Sandipan  and
      Srinivasan, Anirudh  and
      Sitaram, Sunayana  and
      Choudhury, Monojit",
    booktitle = "Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics",
    month = jul,
    year = "2020",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "",
    pages = "3575--3585"
Downloads last month
Hosted inference API
Text Classification
This model can be loaded on the Inference API on-demand.