Upload 6 files
Browse files- en_tokenizer/README.md +35 -0
- en_tokenizer/config.json +19 -0
- en_tokenizer/pytorch_model.bin +3 -0
- en_tokenizer/special_tokens_map.json +1 -0
- en_tokenizer/tokenizer_config.json +1 -0
- en_tokenizer/vocab.txt +0 -0
en_tokenizer/README.md
ADDED
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# ERNIE-2.0-large
|
2 |
+
|
3 |
+
## Introduction
|
4 |
+
|
5 |
+
ERNIE 2.0 is a continual pre-training framework proposed by Baidu in 2019,
|
6 |
+
which builds and learns incrementally pre-training tasks through constant multi-task learning.
|
7 |
+
Experimental results demonstrate that ERNIE 2.0 outperforms BERT and XLNet on 16 tasks including English tasks on GLUE benchmarks and several common tasks in Chinese.
|
8 |
+
|
9 |
+
More detail: https://arxiv.org/abs/1907.12412
|
10 |
+
|
11 |
+
## Released Model Info
|
12 |
+
|
13 |
+
This released pytorch model is converted from the officially released PaddlePaddle ERNIE model and
|
14 |
+
a series of experiments have been conducted to check the accuracy of the conversion.
|
15 |
+
|
16 |
+
- Official PaddlePaddle ERNIE repo: https://github.com/PaddlePaddle/ERNIE
|
17 |
+
- Pytorch Conversion repo: https://github.com/nghuyong/ERNIE-Pytorch
|
18 |
+
|
19 |
+
## How to use
|
20 |
+
```Python
|
21 |
+
from transformers import AutoTokenizer, AutoModel
|
22 |
+
tokenizer = AutoTokenizer.from_pretrained("nghuyong/ernie-2.0-large-en")
|
23 |
+
model = AutoModel.from_pretrained("nghuyong/ernie-2.0-large-en")
|
24 |
+
```
|
25 |
+
|
26 |
+
## Citation
|
27 |
+
|
28 |
+
```bibtex
|
29 |
+
@article{sun2019ernie20,
|
30 |
+
title={ERNIE 2.0: A Continual Pre-training Framework for Language Understanding},
|
31 |
+
author={Sun, Yu and Wang, Shuohuan and Li, Yukun and Feng, Shikun and Tian, Hao and Wu, Hua and Wang, Haifeng},
|
32 |
+
journal={arXiv preprint arXiv:1907.12412},
|
33 |
+
year={2019}
|
34 |
+
}
|
35 |
+
```
|
en_tokenizer/config.json
ADDED
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"attention_probs_dropout_prob": 0.1,
|
3 |
+
"intermediate_size": 4096,
|
4 |
+
"hidden_act": "gelu",
|
5 |
+
"hidden_dropout_prob": 0.1,
|
6 |
+
"hidden_size": 1024,
|
7 |
+
"initializer_range": 0.02,
|
8 |
+
"max_position_embeddings": 512,
|
9 |
+
"num_attention_heads": 16,
|
10 |
+
"num_hidden_layers": 24,
|
11 |
+
"type_vocab_size": 4,
|
12 |
+
"vocab_size": 30522,
|
13 |
+
"pad_token_id": 0,
|
14 |
+
"layer_norm_eps": 1e-05,
|
15 |
+
"model_type": "ernie",
|
16 |
+
"architectures": [
|
17 |
+
"ErnieModel"
|
18 |
+
]
|
19 |
+
}
|
en_tokenizer/pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:89ef99415c7bd5e585abdd19252a5e95ff42c361f4e6f8da823e989fd396395c
|
3 |
+
size 1340701491
|
en_tokenizer/special_tokens_map.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]"}
|
en_tokenizer/tokenizer_config.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"special_tokens_map_file": null, "full_tokenizer_file": null}
|
en_tokenizer/vocab.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|