|
--- |
|
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, 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") |
|
``` |
|
|
|
|
