dataset
mujoco-sota-benchmark
ParamTatva
or hover any field below to flag it
Overview
Name
mujoco-sota-benchmark
Source
ParamTatva
Episodes
0
Robot count
0
Format
other
Description
MuJoCo SOTA Benchmark
Standard MuJoCo continuous control benchmarks from Gymnasium used to evaluate reinforcement learning algorithms.
Benchmark Environments
Environment
Obs Dim
Act Dim
CleanRL SOTA
ParamTatva Best
Hopper-v5
11
3
2,382 +/- 271
3,183.2 (134%)
Walker2d-v5
17
6
~4,000
4,918.5 (123%)
HalfCheetah-v5
17
6
~6,000
5,803.9 (97%)
Reacher-v5
8
2
~-4
-4.2 (~100%)
Ant-v5
27
8
~5,000
886.6 (training)
Humanoid-v5
348
17
~5,000
573.8 (training)… See the full description on the dataset page: https://huggingface.co/datasets/ParamTatva/mujoco-sota-benchmark.
Robots used
null
Links
HuggingFace dataset