5. Run Tensorboard with PyTorch - Pommerman/pytorch-pommerman-rl GitHub Wiki

As we need to install tensorbard and related libs, please update your python environment first.

cd pytorch-pommerman-rl
conda activate pommerman
conda env update --file env.yml  --prune

After you running your training code, start your tensorboard as following:

cd pytorch-pommerman-rl
tensorboard --logdir=./runs

then check http://localhost:6006 to see graphs.