Create README.md

README.md

@@ -0,0 +1,20 @@
+---
+tags:
+- deep-reinforcement-learning
+- reinforcement-learning
+- decision-transformer
+- gym-continous-control
+
+---
+# Decision Transformer model trained on medium trajectories sampled from the Gym Walker2d environment
+This is a trained [Decision Transformer](https://arxiv.org/abs/2106.01345) model trained on medium trajectories sampled from the Gym Hopper environment
+
+<video src="https://huggingface.co/edbeeching/decision-transformer-gym-walker2d-expert/resolve/main/walker2d-medium.mp4" controls autoplay loop></video>
+
+The following normlization coeficients are required to use this model:
+
+mean = [ 1.218966, 0.14163373, -0.03704914, -0.1381431,   0.51382244, -0.0471911, -0.47288352,  0.04225416,  2.3948874,  -0.03143199,  0.04466356, -0.02390724, -0.10134014,  0.09090938, -0.00419264, -0.12120572, -0.5497064] 
+
+std = [0.12311358, 0.324188,   0.11456084, 0.26230657, 0.5640279,  0.22718786, 0.38373196, 0.7373677,  1.2387927,  0.7980206,  1.5664079,  1.8092705, 3.0256042,  4.062486,   1.4586568,  3.744569,   5.585129 ]
+See our [Blog Post](https://colab.research.google.com/drive/1K3UuajwoPY1MzRKNkONNRS3gS5DxZ-qF?usp=sharing),  [Colab notebook](https://colab.research.google.com/drive/1K3UuajwoPY1MzRKNkONNRS3gS5DxZ-qF?usp=sharing) or [Example Script](https://github.com/huggingface/transformers/tree/main/examples/research_projects/decision_transformer) for usage.
+