Skip to content

v0.3.0: Data hub, universal env converter and more!
Compare
Choose a tag to compare
@vmoens vmoens released this 31 Jan 21:40
· 426 commits to main since this release
v0.3.0
520e65f

In this release, we focused on building a Data Hub for offline RL, providing a universal 2gym conversion tool (#1795) and improving the doc.

TorchRL Data Hub

TorchRL now offers many offline datasets in robotics and control or gaming, all under a single data format (TED for TorchRL Episode Data Format). All datasets are one step away of being downloaded: dataset = <Name>ExperienceReplay(dataset_id, root="/path/to/storage", download=True) is all you need to get started.
This means that you can now download OpenX #1751 or Roboset #1743 datasets and combine them in a single replay buffer #1768 or swap one another in no time and with no extra code.
We allow many new sampling techniques, like sampling slices of trajectories with or without repetition etc.
As always you can append your favourite transform to these transforms.

TorchRL2Gym universal converter

#1795 introduces a new universal converter for simulation libraries to gym.
As RL practitioner, it's sometimes difficult to accommodate for the many different environment APIs that exist. TorchRL now provides a way of registering any env in gym(nasium). This allows users to build their dataset in torchrl and integrate them in their code base with no effort if they are already using gym as a backend. It also allows to transform DMControl or Brax envs (among others) to gym without the need for an extra library.

PPO and A2C compatibility with distributed models

Functional calls can now be turned off for PPO and A2C loss modules, allowing users to run RLHF training loops at scale! #1804

## TensorDict-free replay buffers

You can now use TorchRL's replay buffer with ANY tensor-based structure, whether it involves dict, tuples or lists. In principle, storing data contiguously on disk given any gym environment is as simple as

rb = ReplayBuffer(storage=LazyMemmapStorage(capacity))
obs_, reward, terminal, truncated, info = env.step(action)
rb.add((obs, obs_, reward, terminal, truncated, info, action))

# sampling a tuple obs, reward, terminal, truncated, info
obs, obs_, reward, terminal, truncated, info = rb.sample()

This is independent of TensorDict and it supports many components of our replay buffers as well as transforms. Check the doc here.

## Multiprocessed replay buffers

TorchRL's replay buffers can now be shared across processes. Multiprocessed RBs can not only be read from but also extended on different workers. #1724

SOTA checks

We introduce a list of scripts to check that our training scripts work ok before each release: #1822

Throughput of Gym and DMControl

We removed loads of checks in GymLikeEnv if some basic conditions are met, which improves the throughput significantly for simple envs. #1803

## Algorithms

We introduce discrete CQL #1666 , discrete IQL #1793 and Impala #1506.

What's Changed: PR description

New Contributors

Full Changelog: v0.2.1...v0.3.0

Contributors

ianbarber, DanilBaibak, and 13 other contributors
Assets 19
Loading