Update README.md
Browse files
README.md
CHANGED
|
@@ -6,11 +6,15 @@ license: apache-2.0
|
|
| 6 |
|
| 7 |
# G2PTL-1
|
| 8 |
|
|
|
|
| 9 |
## Introduction
|
| 10 |
|
| 11 |
G2PTL-1: A Geography-Graph Pre-trained model for address.
|
| 12 |
This work is the first version of G2PTL (v1.0)
|
| 13 |
|
|
|
|
|
|
|
|
|
|
| 14 |
|
| 15 |
## Model description
|
| 16 |
G2PTL is a Transformer model that is pretrained on a large corpus of Chinese addresses in a self-supervised manner. It has three pretraining objectives:
|
|
|
|
| 6 |
|
| 7 |
# G2PTL-1
|
| 8 |
|
| 9 |
+
|
| 10 |
## Introduction
|
| 11 |
|
| 12 |
G2PTL-1: A Geography-Graph Pre-trained model for address.
|
| 13 |
This work is the first version of G2PTL (v1.0)
|
| 14 |
|
| 15 |
+
## Other work
|
| 16 |
+
We also provide an integrated system for text-based address analysis in logistics field, named **TAAS**, which supports several address perception tasks such as Address Standardization, Address Completion, as well as other logistics related tasks such as Geo-locating From Text to Geospatial and so on. TAAS is available at https://huggingface.co/Cainiao-AI/TAAS.
|
| 17 |
+
|
| 18 |
|
| 19 |
## Model description
|
| 20 |
G2PTL is a Transformer model that is pretrained on a large corpus of Chinese addresses in a self-supervised manner. It has three pretraining objectives:
|