Update README.md

README.md

@@ -7,22 +7,25 @@ license: mit
 
 # Fine-tuning BERT for Fill Mask task on VizWiz dataset
 
-# Fine-tuining information
-* model: bert-base-uncased
-* downstream_tasks: fill-mask
+## Fine-tuining information
+* model: **bert-base-uncased**
+* downstream_tasks: **fill-mask**
 
-# Dataset information
-* vizwiz_data_size: 22K
+## Dataset information
+* vizwiz_data_size: **22K**
+* max_token_len: 78
+### Tokens distribution
+![token_dist](tks_dist.png)
 
-# Training information
-* random_seed: 16
-* max_token_len: 78 
-* train_batch_size: 32 
-* val_batch_size: 16
-* num_epochs: 5, 
-* learning_rate: 5e-06, 
-* split_train: 0.8
-* optimizer: adamw
+## Training information
+* random_seed: **16**
+* max_token_len: **78**
+* train_batch_size: **32** 
+* val_batch_size: **16**
+* num_epochs: **5**, 
+* learning_rate: **5e-06**, 
+* split_train: **0.8**
+* optimizer: **adamw**
 
-# Learning curves
+## Learning curves
 ![learning_curves](lc.png)