Update 05.container_gpu.md

Author: zyndagj

docs/05.container_gpu.md

@@ -113,23 +113,42 @@ Please note that the `--nv` flag specifically passes the GPU drivers into the co
 singularity exec caffe-latest.img caffe device_query -gpu 0
 ```
 
-For TensorFlow, you can directly pull their latest GPU image and utilize it as follows.
+For TensorFlow, you can no longer directly use the images directly from their Docker repository since they now build them against cuda 9. I have created the [tacc-maverick-ml](https://hub.docker.com/r/gzynda/tacc-maverick-ml/), which includes
+
+- tensorflow v1.6.1
+- keras
+- pytorch
+- anaconda python 3
+
+for this purpose.
 
 ```
 # Change to your $WORK directory
 cd $WORK
 #Get the software
 git clone https://github.com/tensorflow/models.git ~/models
 # Pull the image
-singularity pull docker://tensorflow/tensorflow:latest-gpu
+singularity pull docker://gzynda/tacc-maverick-ml:latest
 # Run the code
-singularity exec --nv tensorflow-latest-gpu.img python $HOME/models/tutorials/image/mnist/convolutional.py
+singularity exec --nv tacc-maverick-ml-latest.img python $HOME/models/tutorials/image/mnist/convolutional.py
 ```
 
 You probably noticed that we check out the models repository into your `$HOME` directory. This is because your `$HOME` and `$WORK` directories are only available inside the container if the root folders `/home` and `/work` exist inside the container. In the case of `tensorflow-latest-gpu.img`, the `/work` directory does _not_ exist, so any files there are inaccessible to the container.
 
 You may be thinking "what about overlayfs??". The Linux kernel on Maverick does not support overlayfs, so it had to be disabled in our singularity install.
 
+<br>
+#### Prototyping
+
+You can also prototype your application from inside your container with the `shell` command.
+
+```
+singularity shell tacc-maverick-ml-latest.img
+python --version
+exit
+python --version
+```
+
 <br>
 #### Build your APP