ELECTRA Transformer Models for NLP

4 repos

This cluster contains pre-trained ELECTRA discriminator models and related transformer implementations for natural language processing across multiple languages (Turkish, German, French) and datasets. ELECTRA is an efficient pre-training approach that uses replaced token detection instead of masked language modeling. The repositories represent model checkpoints and implementations built on PyTorch and TensorFlow, designed for downstream NLP tasks like text classification and language understanding.

Related clusters