How to use this model directly from the
from transformers import AutoTokenizer, AutoModel tokenizer = AutoTokenizer.from_pretrained("monsoon-nlp/tamillion") model = AutoModel.from_pretrained("monsoon-nlp/tamillion")
This is the second version of a Tamil language model trained with Google Research's ELECTRA.
Tokenization and pre-training CoLab: https://colab.research.google.com/drive/1Pwia5HJIb6Ad4Hvbx5f-IjND-vCaJzSE?usp=sharing
V1: small model with GPU; 190,000 steps;
V2 (current): base model with TPU and larger corpus; 224,000 steps
Sudalai Rajkumar's Tamil-NLP page contains classification and regression tasks: https://www.kaggle.com/sudalairajkumar/tamil-nlp
The model outperformed mBERT on news classification: (Random: 16.7%, mBERT: 53.0%, TaMillion: 75.1%)
The model slightly outperformed mBERT on movie reviews: (RMSE - mBERT: 0.657, TaMillion: 0.626)
Equivalent accuracy on the Tirukkural topic task.
I didn't find a Tamil-language question answering dataset, but this model could be finetuned to train a QA model. See Hindi and Bengali examples here: https://colab.research.google.com/drive/1i6fidh2tItf_-IDkljMuaIGmEU6HT2Ar
Included as vocab.txt in the upload