Read the blog post first here (link to be added). In each directory, you'll find the logger code with the run example. Remember to install each framework following up-to-date instructions on its website. Moreover, remember to set the environment variable NEPTUNE_API_TOKEN to a valid API token, read more here. Below are extra instructions concerning the example for each framework.
Run python run_d4pg.py --help to show available (and required) parameters.
Note that this RLlib example will run a grid of three experiments with different learning rates. This is very useful for the hyper-parameter search.
Replace <namespace/project_name> with a valid qualified name of your project (this is most probably: your user name / a project name e.g. jankowalski/sandbox).
Remember that in the SpinUp case you have to replace the logx.py in the Spinning Up code with the one supplied in this repository.
Replace <namespace/project_name> with a valid qualified name of your project (this is most probably: your user name / a project name e.g. jankowalski/sandbox).
Replace <namespace/project_name> with a valid qualified name of your project (this is most probably: your user name / a project name e.g. jankowalski/sandbox).