Bharathan Balaji

MultiDiscrete Gym Environments

8

Gym API supports MultiDiscrete action spaces: https://github.com/openai/gym/blob/master/gym/spaces/multi_discrete.py This is useful when you want to discretize a continuous control problem, a technique common in literature: https://arxiv.org/abs/1808.00177 But MultiDiscrete action spaces are...

Are multi-agent environments supported?

1

This comment in code indicates that they are supported: https://github.com/NervanaSystems/coach/blob/master/rl_coach/level_manager.py#L50 But don't see any documentation or examples.

Checkpointing Features

4

Right now Coach saves checkpoints every X seconds. It would be great if I can save checkpoints every X iterations or save checkpoints if it reaches an evaluation reward threshold....

waf build error: ‘zmq::recv_flags’ has not been declared

18

I hit the following error when I tried to install as per instructions in Readme. ``` ../src/opengym/model/opengym_interface.cc: In member function ‘void ns3::OpenGymInterface::Init()’: ../src/opengym/model/opengym_interface.cc:192:36: error: ‘zmq::send_flags’ has not been declared m_zmq_socket.send...

The directory where I run the gym environment is hard encoded into the waf command. Is it possible to wrap the environment in a way that it can be invoked...

CNN Policy

1

Can you please add an example of a CNN policy? All the code is oriented towards MLP policies.

Error in Example: Please provide the bounding box and words or pass the argument "use_ocr" = True

1

Ran into this error while running the example notebook. ``` --------------------------------------------------------------------------- Exception Traceback (most recent call last) /tmp/ipykernel_33283/863471981.py in () 1 ## Using a single batch for the forward propagation...

It'd be nice to get a Python implementation :) Found a good integration with Python here though: https://github.com/UKPLab/arxiv2018-bayesian-ensembles

Error importing torchaudio

2

I'm trying to install torchaudio inside the PyTorch container and run into this error. Looking at online forums indicate that multiple torch versions or CUDA issues lead to this error....