wassname/ray
1
0
Fork 0
mirror of https://github.com/wassname/ray.git synced 2026-07-24 13:20:22 +08:00
Code Issues Packages Projects Releases Wiki Activity

[RLlib] Add rainbow config hint to algo-documentation. (#7052)

Browse Source
This commit is contained in:
Sven Mika
2020-02-05 12:01:43 -08:00
committed by GitHub
parent 580314bf81
commit 0e3960893a
1 changed files with 10 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
doc/source/rllib-algorithms.rst
+10
View File
@@ -191,6 +191,16 @@ Tuned examples: `PongDeterministic-v4 <https://github.com/ray-project/ray/blob/m
.. tip::
Consider using `Ape-X <#distributed-prioritized-experience-replay-ape-x>`__ for faster training with similar timestep efficiency.
.. hint::
For a complete `rainbow <https://arxiv.org/pdf/1710.02298.pdf>`__ setup,
make the following changes to the default DQN config:
``"n_step": [between 1 and 10],
"noisy": True,
"num_atoms": [more than 1],
"v_min": -10.0,
"v_max": 10.0``
(set ``v_min`` and ``v_max`` according to your expected range of returns).
**Atari results @10M steps**: `more details <https://github.com/ray-project/rl-experiments>`__
============= ======================== ============================= ============================== ===============================