first commit

README.md

@@ -8,7 +8,7 @@ The code is written in Python 3.6 and pytorch 0.3.0.
 
 ## Evaluation Results
 
-### Precion/recall curves
+### precision/recall curves
 
 Precion/recall curves of CNN+ATT_BL, CNN+ATT_BL+BAG_ATT, CNN+ATT_RA, CNN+ATT RA+BAG ATT
 
@@ -29,9 +29,9 @@ PCNN+ATT_RA | 0.403 | **0.422**
 
 ## Usage
 
-1. upzip the file `NYT_DATA.zip`
+1. upzip the file `NYT_data/NYT_data.zip`
 
-2. **Make data folder in the following structure**
+2. Make data folder in the following structure
 
 ```
 Intra-Bag-and-Inter-Bag-Attentions
@@ -69,7 +69,7 @@ cd preprocess; bash preprocess.sh; cd ..
 CUDA_VISIBLE_DEVICES=0 python train.py --pretrain --use_RA --sent_encoding pcnn --modelname PCNN_ATTRA
 ```
 
-5. plot the Precion/recall curve
+5. plot the precision/recall curve
 
 ```
 python plot.py --model_name PCNN_ATTRA_BAGATT