diff --git a/Dockerfile b/Dockerfile
new file mode 100644
index 0000000..c359250
--- /dev/null
+++ b/Dockerfile
@@ -0,0 +1,29 @@
+FROM nvidia/cuda:10.1-cudnn7-devel-ubuntu18.04
+LABEL maintainer="oeslle.lucena@gmail.com"
+
+# Set up environment
+RUN DEBIAN_FRONTEND="noninteractive" apt-get update \ 
+    && apt-get install -y \ 
+    dialog apt-utils \
+    curl \
+    htop \
+    nano \
+    tmux \
+    git \
+    python3 \
+    python3-pip \
+    protobuf-compiler
+
+# Install TaPAS 
+RUN cd /home \
+    && git clone https://github.com/google-research/tapas.git \
+    && cd tapas \
+    && pip3 install --upgrade pip \
+    && pip install --upgrade setuptools \
+    && pip install --upgrade six \
+    && pip install scipy==1.4.1 \
+    && pip install oauth2client==4.1.2 \
+    && pip install -e . \
+    && pip install tox \
+    && tox	    
+
diff --git a/README.md b/README.md
index a5ed026..7df2f48 100644
--- a/README.md
+++ b/README.md
@@ -41,6 +41,22 @@ pip install tox
 tox
 ```
 
+### Docker 
+You can also directly use TAPAS via Docker. NVIDIA driver and Docker 19.03+ must be installed beforehand. There is no need to install the CUDA toolkit on the host, but the driver needs to be installed. More details can be found on [nvidia-docker](https://github.com/NVIDIA/nvidia-docker).
+
+#### Build Image
+To build a docker image you can call the following command in the same folder the `Dockerfile` is:
+
+```
+docker build --tag tapas:latest .
+```
+
+Then, you can simply call the following command to run image:
+
+```  
+docker run -it --gpus all -v /absolute/path/to/my/data/directory:/home/data tapas:latest
+```
+
 ## Data
 
 See the section below for the pre-training data.