# Wikipedia pre-training

A LoRA-adapted model based on google/gemma-2-9b, further pre-trained with Turkish Wikipedia data to enhance Turkish text processing capabilities

This is a Tamil instruction fine-tuned version of Google's Gemma 2B model, supporting bilingual text generation in English and Tamil.

A Japanese BART large model pre-trained on Japanese Wikipedia, suitable for text generation and natural language processing tasks.

SHerbert is a SentenceBERT implementation based on the Polish HerBERT model, designed to generate semantically meaningful sentence embeddings, supporting sentence similarity comparison via cosine similarity.

A streamlined version of bert-base-multilingual-cased, supporting only 5 languages (English, French, Spanish, German, and Chinese), 30% smaller than the original while maintaining the same representation quality for these languages.

BERT base model trained on Japanese Wikipedia dataset, suitable for masked language modeling tasks in Japanese text

Italian BERT base version fine-tuned on Italian SQuAD for Q&A downstream tasks

Thai-specific pre-trained model based on BERT-Base architecture, optimized for Thai tokenization characteristics, providing superior performance compared to multilingual BERT

ALBERT base model pre-trained on French Wikipedia, supports case sensitivity, suitable for French NLP tasks.

This is a Swedish pre-trained model based on the RoBERTa architecture, suitable for various natural language processing tasks.

This is a RoBERTa model pre-trained on Thai Wikipedia text, using character-level embeddings to adapt to BertTokenizerFast.

BERT model pre-trained on Chinese Wikipedia texts for POS tagging and dependency parsing.

This is a BERT model pre-trained on Japanese Wikipedia text, derived from bert-large-japanese-char, with enhanced character embedding capabilities to support more Kanji characters.