chrisfel/tensorflow repository overview

⁠Steps

1. Create a docker-compose.yml file in an empty directory and paste the following in it

   version: '3'
   
   services:
     tensorflow:
       image: chrisfel/tensorflow:latest
       ports:
         - 8888:8888
       volumes: 
         - ${PWD}/notebooks:/notebooks
     tensorboard:
       image: chrisfel/tensorflow:latest
       command: tensorboard --logdir=/notebooks/logs
       ports:
         - 6006:6006
       volumes: 
         - ${PWD}/notebooks:/notebooks
   

   Copy

2. In the same directory create a directory called notebooks. This is where your Jupyter notebooks will be mounted from.

3. In the notebooks directory, create an empty directory called logs. This is where logs from training will be placed, and tensorboard will read from that location. NOTE: You should add a rm -rf ./logs to your script before training so that you do not chart old data.

4. Use docker-compose up to bring up the environment. In the output you will see a line to launch your Jupyter notebook environment. e.g. http://localhost:8888/?token=sometoken⁠

5. You can view tensorboard at http://localhost:6006⁠