Releases: awjuliani/neuro-nav
Releases · awjuliani/neuro-nav
Neuro-Nav 2.5.0
Neuro-Nav 2.4.1
- Fixes 3D textures when environment class is reinitialized within a single python session.
Neuro-Nav 2.4.0
- Refactors 2D and 3D renderers for
GridEnv
. - 3D renderer is now roughly 10x faster than before.
Neuro-Nav 2.3.0
New features
- Adds support for manual reward collection in
GridEnv
using themanual_collect
flag on initialization.
Thanks to @quendera for contributing this feature.
Neuro-Nav 2.2.0
- Display reward value in
symbolic
observations ofGridEnv
.
Neuro-Nav 2.1.1
- Fixes bug where
GridEnv
would crash when run on a machine without a display.
Neuro-Nav 2.1.0
- Add
torch_obs
flag toGraphEnv
. - Fix bug in 3d observation dimension order in
GridEnv
.
Neuro-Nav 2.0.0
- Add Proximal Policy Optimization (PPO) algorithm.
- Add Soft Actor-Critic (SAC) algorithm.
- Add Deep-RL tutorial notebook.
- Fix bug in symbolic observation in
GridEnv
. - Miscellaneous documentation improvements.
Neuro-Nav 1.5.5
- Fix bug in 2D grid rendering
Neuro-Nav 1.5.4
- Use caching to significantly improve render time for visual observations in
GridEnv
. - Other minor performance improvements to
GridEnv
.