ken11
commited on
Commit
•
098df7a
1
Parent(s):
ed4659c
init
Browse files- README.md +54 -0
- config.json +29 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +1 -0
- tokenizer_config.json +1 -0
- vocab.txt +0 -0
README.md
CHANGED
@@ -1,3 +1,57 @@
|
|
1 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2 |
license: mit
|
|
|
|
|
|
|
|
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
+
tags:
|
3 |
+
- fill-mask
|
4 |
+
- japanese
|
5 |
+
- albert
|
6 |
+
|
7 |
+
language:
|
8 |
+
- ja
|
9 |
+
|
10 |
license: mit
|
11 |
+
|
12 |
+
widget:
|
13 |
+
- text: "明日は明日の[MASK]が吹く"
|
14 |
+
|
15 |
---
|
16 |
+
## albert-base-japanese-v1-with-japanese
|
17 |
+
日本語事前学習済みALBERTモデルです
|
18 |
+
このモデルではTokenizerに[BertJapaneseTokenizerクラス](https://huggingface.co/docs/transformers/main/en/model_doc/bert-japanese#transformers.BertJapaneseTokenizer)を利用しています
|
19 |
+
[albert-base-japanese-v1](https://huggingface.co/ken11/albert-base-japanese-v1)よりトークナイズ処理が楽になっています
|
20 |
+
|
21 |
+
## How to use
|
22 |
+
### ファインチューニング
|
23 |
+
このモデルはPreTrainedモデルです
|
24 |
+
基本的には各種タスク用にファインチューニングして使用されることを想定しています
|
25 |
+
|
26 |
+
### Fill-Mask
|
27 |
+
#### for PyTorch
|
28 |
+
```py
|
29 |
+
from transformers import (
|
30 |
+
AutoModelForMaskedLM, AutoTokenizer
|
31 |
+
)
|
32 |
+
|
33 |
+
|
34 |
+
tokenizer = AutoTokenizer.from_pretrained("ken11/albert-base-japanese-v1-with-japanese-tokenizer")
|
35 |
+
model = AutoModelForMaskedLM.from_pretrained("ken11/albert-base-japanese-v1-with-japanese-tokenizer")
|
36 |
+
|
37 |
+
text = "明日は明日の[MASK]が吹く"
|
38 |
+
tokens = tokenizer(text, return_tensors="pt")
|
39 |
+
mask_index = tokens["input_ids"][0].tolist().index(tokenizer.mask_token_id)
|
40 |
+
predict = model(**tokens)[0]
|
41 |
+
_, result = predict[0, mask_index].topk(5)
|
42 |
+
|
43 |
+
print(tokenizer.convert_ids_to_tokens(result.tolist()))
|
44 |
+
```
|
45 |
+
|
46 |
+
## Training Data
|
47 |
+
学習には
|
48 |
+
- [日本語Wikipediaの全文](https://ja.wikipedia.org/wiki/Wikipedia:%E3%83%87%E3%83%BC%E3%82%BF%E3%83%99%E3%83%BC%E3%82%B9%E3%83%80%E3%82%A6%E3%83%B3%E3%83%AD%E3%83%BC%E3%83%89)
|
49 |
+
|
50 |
+
を利用しています
|
51 |
+
|
52 |
+
## Tokenizer
|
53 |
+
トークナイザーは[BertJapaneseTokenizerクラス](https://huggingface.co/docs/transformers/main/en/model_doc/bert-japanese#transformers.BertJapaneseTokenizer)を利用しています
|
54 |
+
こちらも学習データは同様です
|
55 |
+
|
56 |
+
## Licenese
|
57 |
+
[The MIT license](https://opensource.org/licenses/MIT)
|
config.json
ADDED
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"architectures": [
|
3 |
+
"AlbertForMaskedLM"
|
4 |
+
],
|
5 |
+
"attention_probs_dropout_prob": 0.1,
|
6 |
+
"bos_token_id": 2,
|
7 |
+
"classifier_dropout_prob": 0.1,
|
8 |
+
"embedding_size": 128,
|
9 |
+
"eos_token_id": 3,
|
10 |
+
"hidden_act": "gelu",
|
11 |
+
"hidden_dropout_prob": 0,
|
12 |
+
"hidden_size": 768,
|
13 |
+
"initializer_range": 0.02,
|
14 |
+
"inner_group_num": 1,
|
15 |
+
"intermediate_size": 3072,
|
16 |
+
"layer_norm_eps": 1e-12,
|
17 |
+
"max_position_embeddings": 512,
|
18 |
+
"model_type": "albert",
|
19 |
+
"num_attention_heads": 12,
|
20 |
+
"num_hidden_groups": 1,
|
21 |
+
"num_hidden_layers": 12,
|
22 |
+
"pad_token_id": 0,
|
23 |
+
"position_embedding_type": "absolute",
|
24 |
+
"tokenizer_class": "BertJapaneseTokenizer",
|
25 |
+
"torch_dtype": "float32",
|
26 |
+
"transformers_version": "4.17.0",
|
27 |
+
"type_vocab_size": 2,
|
28 |
+
"vocab_size": 32000
|
29 |
+
}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a33d484b14c95757687cc6d2959820e70e0e8fe7415f74503732aa6f826298e4
|
3 |
+
size 45933966
|
special_tokens_map.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]"}
|
tokenizer_config.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
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": null, "mecab_kwargs": null, "tokenizer_class": "BertJapaneseTokenizer"}
|
vocab.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|