Update README.md

README.md

@@ -6,38 +6,38 @@ For some experiments, you might wanna pass only the `background` of imagenet ima
 # Restarting from the scratch
 Downloading:
 First download the data from [https://image-net.org/data/bboxes_annotations.tar.gz](here):
-```
+```bash
 wget "https://image-net.org/data/bboxes_annotations.tar.gz"
 ```
 
 Extract the File:
-```
+```bash
 tar -xvf bboxes_annotations.tar.gz 
 ```
 
 Extract every subfolder:
-```
+```bash
 cd bboxes_annotations
 ls | grep .tar.gz | while read f ; do tar -xvf "${f}" ; done
 ```
 
 Convert dataset to JS:
-```
+```bash
 python read_xml.py 
 ```
 
 Clean the extra 50GB extracted files:
-```
+```bash
 rm *.tar.gz
 ls | grep "n.*" | while read f ; do rm -rf "${f}"  ; done 
 ```
 
 Get Indices that have bounding boxes:
-```
+```bash
 python get_indices.py 
 ```
 
 Then simply use pass the files `boxes.pt` and `indices.pt` to your `BackgroundForegroundImageNet` constructor
-```
+```python 
 dataset = BackgroundForegroundImageNet(root='.', download=False, boxes='boxes.pt', indices='indices.pt')
 ```