Also updated environment.yaml to support the tensorboard, which entailed using pytorch 1.2.0
Default path for logs directory is in models/ or in models/policy_networks.
To run tensorboard, just run tensorboard --log-dir models or tensorboard --log-dir policy_networks, depending on what results you want to see.
That script runs a server for the web-page which can be accessed at localhost:6006. If you run it on a remote machine, you can use ssh port tunneling to access the page.
Atcold
commented
5 years ago
Added tensorboard support to:
train_fm.pytrain_MPER.pytrain_MPUR.pytrain_IL.pyeval_policy.pyAlso updated
environment.yamlto support the tensorboard, which entailed using pytorch 1.2.0Default path for logs directory is in
models/or inmodels/policy_networks. To run tensorboard, just runtensorboard --log-dir modelsortensorboard --log-dir policy_networks, depending on what results you want to see.That script runs a server for the web-page which can be accessed at
localhost:6006. If you run it on a remote machine, you can use ssh port tunneling to access the page.