Update README.md

README.md

@@ -8,11 +8,11 @@ tags:
 - text2text-generation
 base_model: google/deplot
 ---
-# **Ko-Deplot**
+# **ko-deplot**
 
-Ko-Deplot is a korean Visual-QA model based on the Google's Pix2Struct architecture. It was fine-tuned from [Deplot](https://huggingface.co/google/deplot), using korean chart image-text pairs.
+ko-deplot is a korean Visual-QA model based on the Google's Pix2Struct architecture. It was fine-tuned from [Deplot](https://huggingface.co/google/deplot), using korean chart image-text pairs.
 
-Ko-Deplot은 Google의 Pix2Struct 구조를 기반으로 한 한국어 Visual-QA 모델입니다. [Deplot](https://huggingface.co/google/deplot) 모델을 한국어 차트 이미지-텍스트 쌍 데이터셋을 이용하여 파인튜닝하였습니다.
+ko-deplot은 Google의 Pix2Struct 구조를 기반으로 한 한국어 Visual-QA 모델입니다. [Deplot](https://huggingface.co/google/deplot) 모델을 한국어 차트 이미지-텍스트 쌍 데이터셋을 이용하여 파인튜닝하였습니다.
 
 - **Developed by:** [NUUA](https://www.nuua.ai/en/)
 - **Model type:** Visual Question Answering
@@ -28,8 +28,8 @@ You can run a prediction by querying an input image together with a question as
 from transformers import Pix2StructProcessor, Pix2StructForConditionalGeneration
 from PIL import Image
 
-processor = Pix2StructProcessor.from_pretrained('nuua/Ko-Deplot')
-model = Pix2StructForConditionalGeneration.from_pretrained('nuua/Ko-Deplot')
+processor = Pix2StructProcessor.from_pretrained('nuua/ko-deplot')
+model = Pix2StructForConditionalGeneration.from_pretrained('nuua/ko-deplot')
 
 IMAGE_PATH = "LOCAL_PATH_TO_IMAGE"
 image = Image.open(IMAGE_PATH)
@@ -39,6 +39,19 @@ predictions = model.generate(**inputs, max_new_tokens=512)
 print(processor.decode(predictions[0], skip_special_tokens=True))
 ```
 
+# **Tokenizer Details**
+The model's tokenizer vocab was extended from 50,344 to 65,536 tokens using the following:
+
+- Complete Korean Jamo
+- [Additional Korean Jamo](http://koreantypography.org/wp-content/uploads/2016/02/kst_12_7_2_06.pdf)
+- Ko-Electra tokens
+
+모델의 tokenizer vocab을 50344개에서 65536개로 아래를 이용하여 확장시킨 후 학습을 진행하였습니다:
+
+- 완성형 한글 자모
+- [추가 완성형 한글 자모](http://koreantypography.org/wp-content/uploads/2016/02/kst_12_7_2_06.pdf)
+- Ko-Electra 한글 토큰
+
 # **Training Details**
 
 ## Training Data
@@ -61,7 +74,7 @@ The model was first exposed to a short warmup stage, following its [original pap
 
 ## Hardware
 
-Ko-Deplot was trained by using A100 80G. 
+ko-deplot was trained by using A100 80G. 
 
 A100 80G GPU를 이용하여 학습하였습니다.