diff --git a/README.md b/README.md
index b851e67..8ada08c 100644
--- a/README.md
+++ b/README.md
@@ -1,2 +1,41 @@
 # MalConv-Pytorch
-Pytorch implementation of MalConv
+A Pytorch implementation of MalConv
+
+---
+## Desciprtion
+
+This is the implementation of MalConv proposed in [Malware Detection by Eating a Whole EXE](https://arxiv.org/abs/1710.09435).
+
+## Dependency
+
+Please make sure each of them is installed with the correct version
+
+- numpy
+- pytorch (0.3.0.post4)
+- pandas (0.20.3)
+
+
+## Setup
+
+#### Preparing data
+
+For the training data, please place PE files under [`data/train/`](`data/train`) and build [the label table](data/example-train-label.csv) for training set with each row being
+
+        <File Name>, <Label>
+
+where label = 1 refers to malware. Validation set should be handled in the same way.
+
+#### Training
+
+Run the following command for training progress
+
+        python3 train.py <config_file_path> <random_seed>
+        Example : python3 train.py config/example.yaml 123
+
+#### Training Log & Checkpoint
+
+Log file, prediction on validation set & Model checkpoint will be stored at the path specified in config file.
+
+## Parameters & Model Options
+
+For parameters and options availible, please refer to [`config/example.yaml`](config/example.yaml).