Added demo data.

Author: Hsiu-Wei Yang

README.md

@@ -5,6 +5,13 @@ The code for the research paper [Aligning Cross-lingual Entities with Multi-Aspe
 
 *Our code is built on top of [GCN-Align](https://github.com/1049451037/GCN-Align).
 
+## Quick Demo
+We store the embeddings for demo in the directory `demo_embd`, and we can evaluate `ZH-EN` as follows:
+```
+python weighted_concat.py -d demo_embd/pairwise_dump.json -g demo_embd/zh_en_graph_embd.pkl -i data/zh_en/test
+```
+
+
 ## Run Graph-based Embeddings (HMAN/MAN)
 
 #### HMAN
@@ -21,7 +28,7 @@ python run.py --lang zh_en --gpu 0 --hybrid 0
 ```
 bash train_bert.sh 0 1
 ```
-Note that you need to stop the training manually.
+(Note that you need to stop the training manually.)
 
 #### Evaluation
 ```
@@ -30,9 +37,7 @@ bash eval_bert.sh 0 1
 
 ## Integration
 ```
-python weighted_concat.py --desc relogic/saves/pair_matching/1/pairwise_dump.json \ 
-                          --graph graph_ckpt/zh_en_graph_embd.pkl \
-                          --ill data/zh_en/test
+python weighted_concat.py --desc relogic/saves/pair_matching/1/pairwise_dump.json --graph graph_ckpt/zh_en_graph_embd.pkl --ill data/zh_en/test
 ```