diff --git a/.devcontainer/cpu/devcontainer.json b/.devcontainer/cpu/devcontainer.json
index 2f6faf8..60d4737 100644
--- a/.devcontainer/cpu/devcontainer.json
+++ b/.devcontainer/cpu/devcontainer.json
@@ -1,7 +1,4 @@
 {
     "name": "muprl - CPU",
-    "build": {
-        "dockerfile": "../../Dockerfile_CPU"
-    },
-    "runArgs": ["--rm"]
+    "image": "dockercontainervm/muprl-cpu:0.1.0"
 }
\ No newline at end of file
diff --git a/.devcontainer/gpu/devcontainer.json b/.devcontainer/gpu/devcontainer.json
index 4252913..3c02e33 100644
--- a/.devcontainer/gpu/devcontainer.json
+++ b/.devcontainer/gpu/devcontainer.json
@@ -1,7 +1,4 @@
 {
     "name": "muprl - GPU",
-    "build": {
-        "dockerfile": "../../Dockerfile_GPU"
-    },
-    "runArgs": ["--rm", "--gpus", "all"]
+    "image": "dockercontainervm/muprl-gpu:0.1.0"
 }
\ No newline at end of file
diff --git a/Dockerfile_GPU b/Dockerfile_GPU
index df67729..35f13d1 100644
--- a/Dockerfile_GPU
+++ b/Dockerfile_GPU
@@ -17,6 +17,7 @@ RUN echo "Installing requirements" \
     && pip install gym==0.25.0
 
 RUN echo "Installing jax cuda" \
+    && pip install --upgrade pip \
     && pip uninstall -y jax \
     && pip uninstall -y jaxlib \
     && pip install --upgrade "jax[cuda12_pip]" -f https://storage.googleapis.com/jax-releases/jax_cuda_releases.html
diff --git a/README.md b/README.md
index 554d9b4..e7cffba 100644
--- a/README.md
+++ b/README.md
@@ -1,6 +1,28 @@
 # 1. Setting up the environment
 
-## 1.1 Build the docker container
+## 1.1 Download the docker image
+
+If you do not have an NVIDIA GPU type:
+
+```commandline
+
+docker run -it -u ${UID} --rm --mount type=bind,source="$(pwd)",target=/home/muPRL --workdir /home/muPRL --name muPRL-container dockercontainervm/muprl-cpu:0.1.0
+
+```
+
+The command will download the container image `dockercontainervm/muprl-cpu:0.1.0` that should be around 2.6 GB.
+
+If you have an NVIDIA GPU, make sure to install the [Nvidia Container Toolkit](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html) and type:
+
+```commandline
+
+docker run --gpus all -it -u ${UID} --rm --mount type=bind,source="$(pwd)",target=/home/muPRL --workdir /home/muPRL --name muPRL-container dockercontainervm/muprl-gpu:0.1.0
+
+```
+
+The command will download the container image `dockercontainervm/muprl-gpu:0.1.0` that should be around * GB.
+
+## 1.2 (Optional): Build the docker container instead of step 1.1
 
 If you do not have an NVIDIA GPU type:
 
@@ -24,7 +46,7 @@ docker run --gpus all -it -u ${UID} --rm --mount type=bind,source="$(pwd)",targe
 
 The image size should be around 8 GB.
 
-## 1.2 (Optional): Use VSCode Devcontainer instead of step 1.1
+## 1.3 (Optional): Use VSCode Devcontainer instead of step 1.1
 
 - Download [VSCode](https://code.visualstudio.com/Download) for your platform;
 - Install DevContainer Extension;