dodox/itpal_jax
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

README <br>

Browse Source
This commit is contained in:
Dominik Moritz Roth 2024-12-21 17:50:26 +01:00
parent c59d8dd635
commit 44eb3335ff
1 changed files with 3 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
README.md
View File

@ -1,8 +1,9 @@
<h1 align="center"> <h1 align="center">
<br>
<img src='./itpal_jax.svg' width="250px"> <img src='./itpal_jax.svg' width="250px">
<br> <br><br>
<b>ITPAL JAX</b> <b>ITPAL JAX</b>
<br> <br><br>
</h1> </h1>
JAX bindings and native implementations of differentiable trust region projections for Gaussian policies. The KL projection is handled by [ITPAL](https://github.com/ALRhub/ITPAL)'s C++ implementation, while Wasserstein and Frobenius projections are implemented in JAX. These projections provide exact solutions for trust region constraints, unlike approximate methods like PPO. JAX bindings and native implementations of differentiable trust region projections for Gaussian policies. The KL projection is handled by [ITPAL](https://github.com/ALRhub/ITPAL)'s C++ implementation, while Wasserstein and Frobenius projections are implemented in JAX. These projections provide exact solutions for trust region constraints, unlike approximate methods like PPO.