initial release

README.md

@@ -0,0 +1,29 @@
+---
+language:
+- "ja"
+tags:
+- "japanese"
+- "masked-lm"
+license: "cc-by-sa-4.0"
+pipeline_tag: "fill-mask"
+mask_token: "[MASK]"
+widget:
+- text: "日本に着いたら[MASK]を訪ねなさい。"
+---
+
+# deberta-large-japanese-unidic
+
+## Model Description
+
+This is a DeBERTa(V2) model pre-trained on 青空文庫 texts with BertJapaneseTokenizer. You can fine-tune `deberta-large-japanese-unidic` for downstream tasks, such as [POS-tagging](https://huggingface.co/KoichiYasuoka/deberta-large-japanese-unidic-luw-upos), dependency-parsing, and so on.
+
+## How to Use
+
+```py
+from transformers import AutoTokenizer,AutoModelForMaskedLM
+tokenizer=AutoTokenizer.from_pretrained("KoichiYasuoka/deberta-large-japanese-unidic")
+model=AutoModelForMaskedLM.from_pretrained("KoichiYasuoka/deberta-large-japanese-unidic")
+```
+
+[fugashi](https://pypi.org/project/fugashi) and [unidic-lite](https://pypi.org/project/unidic-lite) are required.
+

config.json

@@ -0,0 +1,31 @@
+{
+  "architectures": [
+    "DebertaV2ForMaskedLM"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": 0,
+  "eos_token_id": 2,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 1024,
+  "initializer_range": 0.02,
+  "intermediate_size": 4096,
+  "layer_norm_eps": 1e-07,
+  "max_position_embeddings": 512,
+  "max_relative_positions": -1,
+  "model_type": "deberta-v2",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 24,
+  "pad_token_id": 1,
+  "pooler_dropout": 0,
+  "pooler_hidden_act": "gelu",
+  "pooler_hidden_size": 1024,
+  "pos_att_type": null,
+  "position_biased_input": true,
+  "relative_attention": false,
+  "tokenizer_class": "BertJapaneseTokenizer",
+  "torch_dtype": "float32",
+  "transformers_version": "4.19.2",
+  "type_vocab_size": 0,
+  "vocab_size": 32000
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:69f1286b5c9eab5df3681988ab34b92ee6915b39984a1e8d9cba8b5aa176e6d7
+size 1346887533

special_tokens_map.json

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

tokenizer_config.json

@@ -0,0 +1 @@
+{"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]", "do_lower_case": false, "do_word_tokenize": true, "do_subword_tokenize": true, "word_tokenizer_type": "mecab", "subword_tokenizer_type": "wordpiece", "never_split": ["[CLS]", "[PAD]", "[SEP]", "[UNK]", "[MASK]"], "mecab_kwargs": {"mecab_dic": "unidic_lite"}, "model_max_length": 512, "tokenizer_class": "BertJapaneseTokenizer"}