update readme
Browse files
README.md
CHANGED
|
@@ -10,10 +10,9 @@ widget:
|
|
| 10 |
This a albert_chinese_base model from [Google's github](https://github.com/google-research/ALBERT)
|
| 11 |
converted by huggingface's [script](https://github.com/huggingface/transformers/blob/master/src/transformers/convert_albert_original_tf_checkpoint_to_pytorch.py)
|
| 12 |
|
| 13 |
-
## Update
|
| 14 |
-
Support AutoTokenizer
|
| 15 |
|
| 16 |
-
##
|
|
|
|
| 17 |
|
| 18 |
Since sentencepiece is not used in albert_chinese_base model
|
| 19 |
you have to call BertTokenizer instead of AlbertTokenizer !!!
|
|
@@ -24,7 +23,6 @@ we can eval it using an example on MaskedLM
|
|
| 24 |
我們可以跑MaskedLM預測來驗證這個做法是否正確
|
| 25 |
|
| 26 |
## Justify (驗證有效性)
|
| 27 |
-
[colab trial](https://colab.research.google.com/drive/1Wjz48Uws6-VuSHv_-DcWLilv77-AaYgj)
|
| 28 |
```python
|
| 29 |
from transformers import AutoTokenizer, AlbertForMaskedLM
|
| 30 |
import torch
|
|
|
|
| 10 |
This a albert_chinese_base model from [Google's github](https://github.com/google-research/ALBERT)
|
| 11 |
converted by huggingface's [script](https://github.com/huggingface/transformers/blob/master/src/transformers/convert_albert_original_tf_checkpoint_to_pytorch.py)
|
| 12 |
|
|
|
|
|
|
|
| 13 |
|
| 14 |
+
## Notice
|
| 15 |
+
*Support AutoTokenizer*
|
| 16 |
|
| 17 |
Since sentencepiece is not used in albert_chinese_base model
|
| 18 |
you have to call BertTokenizer instead of AlbertTokenizer !!!
|
|
|
|
| 23 |
我們可以跑MaskedLM預測來驗證這個做法是否正確
|
| 24 |
|
| 25 |
## Justify (驗證有效性)
|
|
|
|
| 26 |
```python
|
| 27 |
from transformers import AutoTokenizer, AlbertForMaskedLM
|
| 28 |
import torch
|