Update barkour notebook.
PiperOrigin-RevId: 610906071 Change-Id: Ie67b86888e195bbd8a8e879547e65fbc2c864534
This commit is contained in:
committed by
Copybara-Service
parent
2067e2083a
commit
2e43f18bee
@@ -1,14 +1,4 @@
|
|||||||
# Google Barkour v0 Joystick Policy
|
# Google Barkour v0
|
||||||
|
|
||||||
## Overview
|
|
||||||
|
|
||||||
This folder contains a training script for a flat-terrain joystick policy for the [Barkour v0 Quadruped](https://ai.googleblog.com/2023/05/barkour-benchmarking-animal-level.html) which demonstrates sim2real transfer.
|
|
||||||
|
|
||||||
`barkour_joystick.py` contains the environment definition, while the [colab](https://colab.research.google.com/github/google/brax/blob/main/experimental/barkour_v0/barkour_v0_joystick.ipynb) shows how to train the policy.
|
|
||||||
|
|
||||||
<p float="left">
|
|
||||||
<img src="assets/joystick.gif" width="400">
|
|
||||||
</p>
|
|
||||||
|
|
||||||
## MJCF Instructions
|
## MJCF Instructions
|
||||||
|
|
||||||
|
|||||||
Binary file not shown.
|
Before Width: | Height: | Size: 4.0 MiB |
+3
-3
@@ -606,7 +606,7 @@
|
|||||||
"source": [
|
"source": [
|
||||||
"## Train Humanoid Policy\n",
|
"## Train Humanoid Policy\n",
|
||||||
"\n",
|
"\n",
|
||||||
"Let's now train a policy with PPO to make the Humanoid run forwards. Training takes about 8-9 minutes on a Tesla A100 GPU."
|
"Let's now train a policy with PPO to make the Humanoid run forwards. Training takes about 6 minutes on a Tesla A100 GPU."
|
||||||
]
|
]
|
||||||
},
|
},
|
||||||
{
|
{
|
||||||
@@ -1312,7 +1312,7 @@
|
|||||||
"source": [
|
"source": [
|
||||||
"## Train Policy\n",
|
"## Train Policy\n",
|
||||||
"\n",
|
"\n",
|
||||||
"To train a policy with domain randomization, we pass in the domain randomization function into the brax train function; brax will call the domain randomization function when rolling out episodes. Training the quadruped takes 8-9 minutes on a Tesla A100 GPU."
|
"To train a policy with domain randomization, we pass in the domain randomization function into the brax train function; brax will call the domain randomization function when rolling out episodes. Training the quadruped takes 6 minutes on a Tesla A100 GPU."
|
||||||
]
|
]
|
||||||
},
|
},
|
||||||
{
|
{
|
||||||
@@ -1448,7 +1448,7 @@
|
|||||||
},
|
},
|
||||||
"outputs": [],
|
"outputs": [],
|
||||||
"source": [
|
"source": [
|
||||||
"HTML(html.render(eval_env.sys, rollout))"
|
"HTML(html.render(eval_env.sys.replace(dt=eval_env.dt), rollout))"
|
||||||
]
|
]
|
||||||
}
|
}
|
||||||
],
|
],
|
||||||
|
|||||||
Reference in New Issue
Block a user