christianhidber / easyagents

Reinforcement Learning for Practitioners.

christianhidber/easyagents Issues

Is it possible to resume training with a saved model?
Updated 3 years ago3
Is there currently an elegant way of adding an RNN or LSTM to an agent?
Updated 4 years ago3
Running on a GPU
Closed 4 years ago3
Add support for RLLib
Updated 4 years ago5
Pass args and kwargs to environment upon registration
Closed 4 years ago5
Spec_shape error
Closed 5 years ago1
Support for persistence
Closed 5 years ago1
lineworld as test env
Closed 5 years ago1
AttributeError: module 'tensorflow_probability.python.distributions' has no attribute 'SeedStream'
Closed 5 years ago1
Access to the trained policy.
Updated 5 years ago
flexible loss metrics / plot.Loss
Updated 5 years ago
Add a score method to an agent
Closed 5 years ago12
log output is cleared after train even if no plot callback is registered
Closed 5 years ago1
add num_iterations_between_plots argument
Closed 5 years ago
CartPole ist not rendered in tensorforce PpoAgent
Closed 5 years ago
investigate arbitrary Keras models.
Updated 5 years ago
Custom networks for ppo and dqn
Closed 5 years ago1
Support for OpenAI baselines & huskarl
Closed 5 years ago1
Simplified box boundaries do not broadcast in Agent
Closed 5 years ago
max value of the episode rewards in the orso book training visualization always >= 0
Closed 5 years ago6
Orso notebook broken
Closed 5 years ago
How can we test a jupyter notebook?
Updated 5 years ago
"State image" is missing in lineworld notebook
Closed 5 years ago2
Working with different envs not possible
Closed 5 years ago1
can dev branch be deleted ?
Closed 5 years ago1
ReinforceAgent missing from pypi deployment
Closed 5 years ago1