Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[Unreleased]

Added

Changed

Fixed

v0.2.0

Added

• Introduce cherry.nn.Policy, cherry.nn.ActionValue, and cherry.nn.StateValue.
• Algorithm class utilities for: A2C, PPO, TRPO, DDPG, TD3, SAC, and DrQ/DrQv2.
• DMC examples for SAC, DrQ, and DrQv2.
• N-steps returns sampling in ExperienceReplay.

Changed

• Discontinue most of cherry.wrappers.

Fixed

• Fixes return value of StateNormalizer and RewardNormalizer wrappers.
• Requirements to generate docs.

v0.1.4

Fixed

• Support for torch 1.5 and new _parse_to behavior in ExperienceReplay. (thanks @ManifoldFR)

v0.1.3

Added

• A CHANGELOG.md file.

Changed

• Travis testing with different versions of Python (3.6, 3.7), torch (1.1, 1.2, 1.3, 1.4), and torchvision (0.3, 0.4, 0.5).

Fixed

• fix bug in torch_wrapper when use GPU by callling Tensor.cpu().detach().numpy() to convert CUDA tensor to numpy.(@walkacross)
• Bugfix when using td.discount with replays coming from vectorized environments (@galatolofederico)
• env.action_size and env.state_size when the number of vectorized environments is 1. (thanks @galatolofederico)
• Actor-critic integration test being to finicky.
• cherry.onehot support for numpy's float and integer types. (thanks @ngoby)