rl
rl copied to clipboard
A modular, primitive-first, python-first PyTorch library for Reinforcement Learning.
This PR introduces the concept of storage save hooks, which can modify a storage before serializing it TODO: - Account for more complex data structures (e.g. nested done, nested rewards)...
## Motivation There is currently no TensorSpec such that it allows sequential data. For example, if an environment was supposed to represent a sentence, it would be represented as (batch_size,...
## Describe the bug There is a bug when an EnvBase has a MultiDiscreteTensorSpec and moved to mps (I have not tested yet if the issue persists with cuda). ##...
## Describe the bug There seems to be a bug with performing a rollout when the tensordicts produced do not have the same shape. The rollout used to produce LazyStackedTensorDict,...
The "Run in Colab" button of the notebooks seems to give an error ```bash Notebook not found There was an error loading this notebook. Ensure that the file is accessible...