| language: | |
| - "th" | |
| tags: | |
| - "thai" | |
| - "masked-lm" | |
| - "wikipedia" | |
| license: "apache-2.0" | |
| pipeline_tag: "fill-mask" | |
| mask_token: "<mask>" | |
| widget: | |
| - text: "แผนกนี้กำลัง<mask>กับความท้าทายใหม่" | |
| # roberta-base-thai-syllable | |
| ## Model Description | |
| This is a RoBERTa model pre-trained on Thai Wikipedia texts, derived from [wangchanberta-base-wiki-syllable](https://huggingface.co/airesearch/wangchanberta-base-wiki-syllable). Character-embeddings are modified to use BertTokenizerFast. You can fine-tune `roberta-base-thai-syllable` for downstream tasks, such as [POS-tagging](https://huggingface.co/KoichiYasuoka/roberta-base-thai-syllable-upos), [dependency-parsing](https://huggingface.co/KoichiYasuoka/roberta-base-thai-syllable-ud-goeswith), and so on. | |
| ## How to Use | |
| ```py | |
| from transformers import AutoTokenizer,AutoModelForMaskedLM | |
| tokenizer=AutoTokenizer.from_pretrained("KoichiYasuoka/roberta-base-thai-syllable") | |
| model=AutoModelForMaskedLM.from_pretrained("KoichiYasuoka/roberta-base-thai-syllable") | |
| ``` | |