Skip to content

Latest commit

 

History

History
93 lines (73 loc) · 3.28 KB

CHANGELOG.md

File metadata and controls

93 lines (73 loc) · 3.28 KB
Raw

Changelog

Release 2.7.1 (2024-04-19)

  • Add Continuous Mountain Car environment
  • A2C algorithm supports conditional down-sampling for bad trajectories

Release 2.7 (2024-02-17)

  • Support continuous actions
  • Add Pendulum environment that can run up to 100K concurrent replicates
  • Add DDPG algorithms for training continuous action policies

Release 2.6.2 (2023-12-12)

  • Add Acrobot environment that can run up to 100K concurrent replicates.
  • Add Mountain Car environment that can run up to 100K concurrent replicates.

Release 2.6.1 (2023-11-05)

  • Support single agent framework and start to add gym.classic_control
  • Add Cartpole environment that can run up to 100K concurrent replicates.

Release 2.5 (2023-07-27)

  • Introduce environment reset pool, so concurrent enviornment replicas can randomly reset themselves from the pool.

Release 2.4 (2023-06-16)

  • Introduce new device context management and autoinit_pycuda
  • Therefore, Torch (any version) will not conflict with PyCUDA in the GPU context

Release 2.3 (2023-03-22)

  • Add ModelFactory class to manage custom models
  • Add Xavier initialization for the model
  • Improve trainer.fetch_episode_states() so it can fetch (s, a, r) and can replay with argmax.

Release 2.2 (2022-12-20)

  • Factorize the data loading for placeholders and batches (obs, actions and rewards) for the trainer.

Release 2.1 (2022-10-26)

  • v2 trainer integration with Pytorch Lightning

Release 2.0 (2022-09-20)

Big release:

  • WarpDrive:
    • Added data and function managers for both CUDA C and Numba.
    • Added core library (sampler and reset) for Numba.
    • Dual environment backends, supporting both CUDA C and Numba.
    • Training pipeline compatible with both CUDA C and Numba.
    • Full backward compatibility with version 1.
  • Environments
    • tag (continuous version) implemented in Numba.
    • tag (gridworld version) implemented in Numba.

Release 1.7 (2022-09-08)

  • Update PyCUDA version to 2022.1

Release 1.6 (2022-04-05)

  • Allow for envs to span multiple blocks, adding the capability to train simulations with thousands of agents.

Release 1.5 (2022-03-01)

Release 1.4 (2022-01-29)

  • Added multi-GPU support.

Release 1.3 (2022-01-10)

  • Auto-scaling to maximize the number of environment replicas and training batch size (on a single GPU).
  • Added Python logging.

Release 1.2.2 (2021-12-16)

  • Added a trainer module to fetch environment states for an episode.

Release 1.2.1 (2021-12-07)

  • Add policy-specific training parameters.

Release 1.2 (2021-12-02)

  • Added a parameter scheduler.
  • Option to push a list of data arrays to the GPU at once.
  • Option to pass multiple arguments to the CUDA step function as a list.
  • CUDA utility to help index multi-dimensional arrays.
  • Log the episodic rewards.
  • Save metrics during training.

Release 1.1 (2021-09-27)

  • Support to register custom environments.
  • Support for 'Dict' observation spaces.

Release 1.0 (2021-09-01)

  • WarpDrive
    • data and function managers.
    • CUDA C core library.
    • environment wrapper.
    • Python (CPU) vs. CUDA C (GPU) simulation implementation consistency checker
    • training pipeline (with FC network, and A2C, PPO agents).
  • Environments
    • tag (grid-world version).
    • tag (continuous version).