diff --git a/README.md b/README.md
index 31f3929..19317a6 100644
--- a/README.md
+++ b/README.md
@@ -1,8 +1,9 @@
 # Adaptive Softmax with Pytorch
 
-This is an implement of **Adaptive Softmax** with pytorch, the paper is "Efficient softmax approximation for GPUs" (http://arxiv.org/abs/1609.04309).
+This is an implementation of **Adaptive Softmax** with pytorch, the paper is "Efficient softmax approximation for GPUs"(http://arxiv.org/abs/1609.04309).
 
 ## Requirement:
+
 python2.7 or 3 (other versions may also support)
 
 pytorch0.2.0 (other versions may also support)
@@ -13,35 +14,38 @@ tqdm
 
 1. Download and prepare the data (Text8) :
 
-```
-cd data
-sh makedata-text8.sh
-```
+    ```
+    cd data
+    sh makedata-text8.sh
+    ```
 
 2. Train the language model with adaptive softmax:
 
-```
-python main.py
-```
+    ```
+    python main.py
+    ```
 
 3. Train the language model with regular softmax:
 
-```
-python main.py --with_adaptive 0
-```
+    ```
+    python main.py --with_adaptive 0
+    ```
 
 ## Current results:
 
-### Test perplexity:
-Pass |  adaptive softmax | regular softmax
- ---------------:  | ---------------:  
-1 | 227.934 | 267.424
-2 | 195.780 | 226.902
-3 | 182.551 | 211.505
-4 | 173.464 | 199.371
-5 | 168.012 | 193.128
+### **Test perplexity**:
+
+| Pass |  adaptive softmax | regular softmax |
+| ----- | ---------------  | --------------- |
+| 1 | 227.934 | 267.424 |
+| 2 | 195.780 | 226.902 |
+| 3 | 182.551 | 211.505 |
+| 4 | 173.464 | 199.371 |
+| 5 | 168.012 | 193.128 |
+
+
 
-### Time cost of one sample:
+### **Time cost of one sample**:
 
 **adaptive softmax** : 16.46it/s
 
diff --git a/utils.py b/utils.py
index 84115b8..37fa3c9 100644
--- a/utils.py
+++ b/utils.py
@@ -1,11 +1,6 @@
-import urllib
-import shutil
-import zipfile
 import os
-import pickle
 from collections import Counter
 import numpy as np
-import codecs
 
 
 def build_dict(train_file_path):