initial release

README.md

@@ -0,0 +1,25 @@
+---
+language:
+- "ko"
+tags:
+- "korean"
+- "masked-lm"
+license: "cc-by-sa-4.0"
+pipeline_tag: "fill-mask"
+mask_token: "[MASK]"
+---
+
+# roberta-base-korean-hanja
+
+## Model Description
+
+This is a RoBERTa model pre-trained on Korean texts, derived from [klue/roberta-base](https://huggingface.co/klue/roberta-base). Token-embeddings are enhanced to include all 인명용 한자 characters. You can fine-tune `roberta-base-korean-hanja` for downstream tasks, such as [POS-tagging](https://huggingface.co/KoichiYasuoka/roberta-base-korean-upos), [dependency-parsing](https://huggingface.co/KoichiYasuoka/roberta-base-korean-ud-goeswith), and so on.
+
+## How to Use
+
+```py
+from transformers import AutoTokenizer,AutoModelForMaskedLM
+tokenizer=AutoTokenizer.from_pretrained("KoichiYasuoka/roberta-base-korean-hanja")
+model=AutoModelForMaskedLM.from_pretrained("KoichiYasuoka/roberta-base-korean-hanja")
+```
+

config.json

@@ -0,0 +1,28 @@
+{
+  "architectures": [
+    "RobertaForMaskedLM"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": 0,
+  "classifier_dropout": null,
+  "eos_token_id": 2,
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 514,
+  "model_type": "roberta",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "pad_token_id": 1,
+  "position_embedding_type": "absolute",
+  "tokenizer_class": "BertTokenizer",
+  "torch_dtype": "float32",
+  "transformers_version": "4.22.1",
+  "type_vocab_size": 1,
+  "use_cache": true,
+  "vocab_size": 39255
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:53f2e80a033d9e3e937c9cc27219e1c66219714ab4e525f6640b997a60c631ca
+size 464992107

special_tokens_map.json

@@ -0,0 +1,9 @@
+{
+  "bos_token": "[CLS]",
+  "cls_token": "[CLS]",
+  "eos_token": "[SEP]",
+  "mask_token": "[MASK]",
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "unk_token": "[UNK]"
+}

tokenizer_config.json

@@ -0,0 +1,16 @@
+{
+  "bos_token": "[CLS]",
+  "cls_token": "[CLS]",
+  "do_basic_tokenize": true,
+  "do_lower_case": false,
+  "eos_token": "[SEP]",
+  "mask_token": "[MASK]",
+  "model_max_length": 512,
+  "never_split": null,
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "BertTokenizer",
+  "unk_token": "[UNK]"
+}