IndicTriMix: Developing Language Identification Datasets and Models for Tri-Language Code-Mixing
Researchers release IndicTriMix benchmarks and fine-tuned MuRIL and XLM-RoBERTa models for token-level language identification in tri-language code-mixed text.