# ELECTRA Architecture

A text reranking model based on ELECTRA architecture for relevance sorting of retrieval results

A text ranking cross-encoder based on the ELECTRA architecture, designed for retrieval result reranking tasks

lightning-ir is a cross-encoder model based on the ELECTRA architecture, specifically designed for text ranking tasks. The model optimizes paragraph reordering performance through large language model distillation techniques.

ELECTRA-small model fine-tuned on SQuAD dataset for question answering tasks

This model is a fine-tuned version of the ELECTRA-small discriminator based on the SQuAD dataset, suitable for question-answering tasks.

This is an English extractive question answering model based on the ELECTRA-base architecture, trained on the SQuAD 2.0 dataset, suitable for question answering tasks.

This is a large-scale discriminator model based on the ELECTRA architecture, specifically fine-tuned for Q&A tasks on the SQuAD2.0 dataset, capable of handling both answerable and unanswerable question scenarios.

A Tamil pre-trained model based on the ELECTRA framework, with the second version trained on TPUs and expanded corpus scale

Japanese pretrained model based on ELECTRA architecture, pretrained on approximately 200 million Japanese sentences from the mC4 dataset and fine-tuned on UD_Japanese_BCCWJ corpus

This is a cross-encoder model based on the ELECTRA architecture, specifically designed for natural language inference (NLI) in question-answering tasks, determining whether a given question can be answered by a specific paragraph.

A cross-encoder trained on the ELECTRA-base architecture, specifically optimized for the MS Marco passage ranking task, used for query-passage relevance scoring in information retrieval.

A two-stage training QA model based on ELECTRA-Large architecture, first trained on synthetic adversarial data and then fine-tuned on SQuAD and AdversarialQA datasets

This is a BERT model pre-trained on Japanese text, specifically optimized for the financial domain.

KoELECTRA is a Korean pretrained language model based on the ELECTRA architecture, developed by monologg. This model serves as the generator component, focusing on representation learning for Korean text.