Release 0.5.2

Minor release

Release 0.5.2 fixes a bug in the SAC implementation provided in version 0.5.1, which had been inadvertently pushed from a development branch.

Note: this version introduces a new way of recording replays in TrackMania, but this is not officially supported yet and won't work without the corresponding OpenPlanet script. If you want to use this feature before the next version is out, please contact us.

Release 0.5.1

Release 0.5.1

---

This release complies with rtgym>=0.9, which in turns complies with the gymnasium signature of the reset function.

In case you are using custom rtgym interfaces in tmrl, you will want to update your reset implementations. This is straightforward, you can just replace:

def reset(self):

with:

def reset(self, seed=None, options=None):

Release 0.5.0

Major release 0.5.0

---

Release 0.5.0 upgrades the tmrl framework to the the maintained gymnasium framework, instead of the old and abandoned gym framework.

This release follows the changes made in rtgym.

Version 0.5.0 does not introduce any breaking change, but you still need to delete your checkpoints from previous versions in TmrlData/checkpoints (or you can perform a clean install by deleting the TmrlData folder prior to pip-installing version 0.5.0), since older versions of the framework use to import gym where it now imports gymnasium.

Release 0.4.2

Minor release 0.4.2

---

This update improves TrackMania support (thanks @LaurensNeinders !)

Perform a clean install by deleting the TmrlData folder AND Plugin_GrabData_0_1.as from OpenPlanetNext\Scripts

Release 0.4.2 ships with an optimized version of the OpenPlanet script for the default TrackMania environments. In particular, it should alleviate a common issue where the environment randomly times-out for a roughly constant amount of time even on high-end PCs.

Release 0.4.1

Minor release 0.4.1

This release uses a new feature of tlspyo 0.2.5 to stop randomly deserializing objects in the background during episode collection.

Release 0.4.0

Major release 0.4.0

---

⚠️ This release is backward-incompatible!

tmrl 0.4.0 introduces network security and many changes to the API.

In particular:

• network security is now handled by tlspyo
• the API is now compatible with any framework (i.e., we do not force the use of PyTorch anymore)
• the framework now supports color images
• the default config.json file is tuned for fast training of CNN (grayscale) policies

tmrl 0.4.0 also "officially" launches the first Beta of the TrackMania Roborace League competition.
The competition tutorial script is particularly useful for ML developers who desire to code advanced training pipelines in TrackMania.

Release 0.3.1

• Corrected warning from 0.3.0
• Removed unsupported old code

Release 0.3.0

Major release 0.3.0.

This release is backward incompatible!

Notable changes:

• The API underwent major changes following the new OpenAI API. See https://github.com/openai/gym/releases/tag/0.26.0
• This release introduces long-awaiten support for full screenshot environments in TrackMania 2020 🚀

Please perform a clean install by deleting the TmrlData folder in your home directory.

Release 0.2.0

This release introduces new flavors of the TrackMania Gym environments: one featuring raw screenshots, and one featuring a track-progress observation along the LIDAR (you can use this by ending the "ENV" entry with "LIDARPROGRESS" in config.json).

This release also fixes the latest breaking changes that the new Gym maintainers team keeps introducing...

This version is backward incompatible!

Please perform a clean install by deleting your TmrlData folder before pip-installing.

Release 0.1.6

On Windows, tmrl now attempts to fix pywin32 in case it fails to import (this happens when installing pywin32 from pip )