Implementation of DQN in Vernon. Can be used as substrate for other embodied AI implementations.
First draft, not fully functional (runs, but doesn't learn as well). Requires some debugging.
I will likely merge the first draft, so others can build/work on it, but comments and feedbacks are appreciated and will be incorporated in a second PR. We can also continue the discussion on github even if the PR is merged. Thanks.
Implementation of DQN in Vernon. Can be used as substrate for other embodied AI implementations.
First draft, not fully functional (runs, but doesn't learn as well). Requires some debugging.
I will likely merge the first draft, so others can build/work on it, but comments and feedbacks are appreciated and will be incorporated in a second PR. We can also continue the discussion on github even if the PR is merged. Thanks.