RL Toolkit v5.0.0

update setup

RL Toolkit v4.1.1

Release v4.1.1

Changelog

• update default config.yaml

RL Toolkit v4.1.0

Release v4.1.0

Changelog

Features 🔊

• .fit()
• AgentCallback

RL Toolkit v4.0.0

Release v4.0.0

Changelog

Features 🔊

• Render environments to WanDB
• Grouping of runs in WanDB
• SampleToInsertRatio rate limiter
• Global Gradient Clipping to avoid exploding gradients
• Softplus for numerical stability
• YAML configuration file
• LogCosh instead of Huber loss
• Critic network with Add layer applied on state & action branches
• Custom uniform initializer
• XLA (Accelerated Linear Algebra) compiler
• Optimized Replay Buffer (google-deepmind/reverb#90)
• split into Agent, Learner, Tester and Server

Bug fixes 🛠️

• Fixed creating of saving path for models
• Fixed model's summary()

RL Toolkit v3.2.5

Release v3.2.5

Changelog

• Fix out of memory

RL Toolkit v3.2.4

Release v3.2.4

Changelog

• Reverb
• setup.py (package is available on PyPI)
• Split into agent, learner and tester roles
• Use custom model and layer for defining Actor-Critic
• MultiCritic - concatenating multiple critic networks into one network
• Truncated Quantile Critics

RL Toolkit v2.0.2

Release v2.0.2

Changelog

• + update Dockerfile
• + update README.md
• + formatted code by Black & Flake8

RL-Toolkit v2.0.1

Release v2.0.1

Changelog

• fix Critic model

RL-Toolkit v2.0

Release v2.0

Changelog

• + Huber loss,
• + Rendering to the video file (test mode),
• + Normalized observation by Min-max method,
• + removed TD3 support,
• ± instead of Concatenate layer is used Add layer (Critic network)