Hands-On Reinforcement Learning for Games

Back to TensorBoard

With the sample from the last exercise still running, we want to return to TensorBoard and now see the output from our sample running. To do that, open a new Python/Anaconda command shell and follow the next exercise:

1. Open the shell to the same folder you are running your previous exercise code example in. Switch to your virtual environment or a special one just for TB and then run the following command to start the process:

tensorboard --logdir=runs

2. This will start TB in the current folder using that runs folder as the data dump directory. After the sample is running for a while, you may see something like the following when you visit the TB web interface now:

TensorBoard output from Chapter_10_QRDQN.py

3. Turn up ...