Merge pull request #100 from D-o-d-o-x/bruce_port_envs
Port Additions by Bruce to current version
This commit is contained in:
Onur
GitHub
commit
758aa534ed
BIN
docs/build/doctrees/environment.pickle
vendored
BIN
docs/build/doctrees/environment.pickle
vendored
Binary file not shown.
BIN
docs/build/doctrees/envs/fancy/mujoco.doctree
vendored
BIN
docs/build/doctrees/envs/fancy/mujoco.doctree
vendored
Binary file not shown.
BIN
docs/build/doctrees/examples/dmc.doctree
vendored
BIN
docs/build/doctrees/examples/dmc.doctree
vendored
Binary file not shown.
BIN
docs/build/doctrees/examples/general.doctree
vendored
BIN
docs/build/doctrees/examples/general.doctree
vendored
Binary file not shown.
BIN
docs/build/doctrees/examples/metaworld.doctree
vendored
BIN
docs/build/doctrees/examples/metaworld.doctree
vendored
Binary file not shown.
Binary file not shown.
BIN
docs/build/doctrees/examples/open_ai.doctree
vendored
BIN
docs/build/doctrees/examples/open_ai.doctree
vendored
Binary file not shown.
BIN
docs/build/doctrees/examples/replanning_envs.doctree
vendored
BIN
docs/build/doctrees/examples/replanning_envs.doctree
vendored
Binary file not shown.
BIN
docs/build/doctrees/guide/installation.doctree
vendored
BIN
docs/build/doctrees/guide/installation.doctree
vendored
Binary file not shown.
2
docs/build/html/.buildinfo
vendored
2
docs/build/html/.buildinfo
vendored
@ -1,4 +1,4 @@
|
||||
# Sphinx build info version 1
|
||||
# This file hashes the configuration used when building these files. When it is not found, a full rebuild will be done.
|
||||
config: 28ec069496fc0ad05c8b9641549626a6
|
||||
config: 36919d67c12a677d3f16f60d980b0313
|
||||
tags: 645f666f9bcd5a90fca523b33c5a78b7
|
||||
|
||||
@ -3,7 +3,7 @@
|
||||
<head>
|
||||
<meta charset="utf-8" />
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>fancy_gym.envs.registry — Fancy Gym 0.2 documentation</title>
|
||||
<title>fancy_gym.envs.registry — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="../../../_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../../../_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../../../_static/style.css" type="text/css" />
|
||||
@ -38,7 +38,7 @@
|
||||
<img src="../../../_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="../../../search.html" method="get">
|
||||
|
||||
4
docs/build/html/_modules/index.html
vendored
4
docs/build/html/_modules/index.html
vendored
@ -3,7 +3,7 @@
|
||||
<head>
|
||||
<meta charset="utf-8" />
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>Overview: module code — Fancy Gym 0.2 documentation</title>
|
||||
<title>Overview: module code — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/style.css" type="text/css" />
|
||||
@ -38,7 +38,7 @@
|
||||
<img src="../_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
|
||||
|
||||
@ -18,6 +18,12 @@ A composite reward function serves as the performance metric for the RL system.
|
||||
|
||||
Variations of this environment are available, differing in reward structures and the optionality of randomizing the box's initial position. These variations are purposefully designed to challenge RL algorithms, enhancing their generalization and adaptation capabilities. Temporally sparse environments only provide a reward at the last timestep. Spatially sparse environments only provide a reward, if the goal is almost reached, the box is close enought to the goal and somewhat correctly aligned.
|
||||
|
||||
These environments all provide smoothness metrics as part of the return infos:
|
||||
|
||||
- mean_squared_jerk: Averages the square of jerk (rate of acceleration change) across the motion. Lower values indicate smoother movement.
|
||||
- maximum_jerk: Identifies the highest jerk value encountered.
|
||||
- dimensionless_jerk: Normalizes the summed squared jerk over the motion's duration and peak velocity, offering a scale-independent metric of smoothness
|
||||
|
||||
| Name | Description | Horizon | Action Dimension | Observation Dimension |
|
||||
| ------------------------------------------ | -------------------------------------------------------------------- | ------- | ---------------- | --------------------- |
|
||||
| `fancy/BoxPushingDense-v0` | Custom Box-pushing task with dense rewards | 100 | 3 | 13 |
|
||||
@ -49,6 +55,9 @@ Variations of the table tennis environment are available to cater to different r
|
||||
| `fancy/TableTennisWind-v0` | Table Tennis task with wind effects, based on a custom environment for table tennis | 350 | 7 | 19 |
|
||||
| `fancy/TableTennisGoalSwitching-v0` | Table Tennis task with goal switching, based on a custom environment for table tennis | 350 | 7 | 19 |
|
||||
| `fancy/TableTennisWindReplan-v0` | Table Tennis task with wind effects and replanning, based on a custom environment for table tennis | 350 | 7 | 19 |
|
||||
| `fancy/TableTennisRndRobot-v0` | Table Tennis task with random initial robot joint positions \* | 350 | 7 | 19 |
|
||||
|
||||
\* Random initialization of robot joint position and speed can be enabled by providing `random_pos_scale` / `random_vel_scale` to make. `TableTennisRndRobot` is equivalent to `TableTennis4D` except, that `random_pos_scale` is set to 0.1 instead of 0 per default.
|
||||
|
||||
---
|
||||
|
||||
@ -89,8 +98,9 @@ A successful throw in this task is determined by the ball landing in the cup at
|
||||
| `fancy/Reacher5dSparse-v0` | Sparse Reacher task with 5 links, based on Gymnasium's `gym.envs.mujoco.ReacherEnv` | 200 | 5 | 20 |
|
||||
| `fancy/Reacher7d-v0` | Reacher task with 7 links, based on Gymnasium's `gym.envs.mujoco.ReacherEnv` | 200 | 7 | 22 |
|
||||
| `fancy/Reacher7dSparse-v0` | Sparse Reacher task with 7 links, based on Gymnasium's `gym.envs.mujoco.ReacherEnv` | 200 | 7 | 22 |
|
||||
| `fancy/HopperJumpSparse-v0` | Hopper Jump task with sparse rewards, based on Gymnasium's `gym.envs.mujoco.Hopper` | 250 | 3 | 15 / 16\* |
|
||||
| `fancy/HopperJump-v0` | Hopper Jump task with continuous rewards, based on Gymnasium's `gym.envs.mujoco.Hopper` | 250 | 3 | 15 / 16\* |
|
||||
| `fancy/HopperJumpMarkov-v0` | `fancy/HopperJump-v0`, but with an alternative reward that is markovian. | 250 | 3 | 15 / 16\* |
|
||||
| `fancy/HopperJumpSparse-v0` | Hopper Jump task with sparse rewards, based on Gymnasium's `gym.envs.mujoco.Hopper` | 250 | 3 | 15 / 16\* |
|
||||
| `fancy/AntJump-v0` | Ant Jump task, based on Gymnasium's `gym.envs.mujoco.Ant` | 200 | 8 | 119 |
|
||||
| `fancy/HalfCheetahJump-v0` | HalfCheetah Jump task, based on Gymnasium's `gym.envs.mujoco.HalfCheetah` | 100 | 6 | 112 |
|
||||
| `fancy/HopperJumpOnBox-v0` | Hopper Jump on Box task, based on Gymnasium's `gym.envs.mujoco.Hopper` | 250 | 4 | 16 / 100\* |
|
||||
|
||||
@ -32,7 +32,7 @@ since they are not avaible on PyPI yet. Install metaworld via
|
||||
|
||||
.. code:: bash
|
||||
|
||||
pip install metaworld@git+https://github.com/Farama-Foundation/Metaworld.git@d155d0051630bb365ea6a824e02c66c068947439#egg=metaworld
|
||||
pip install metaworld@git+https://github.com/Farama-Foundation/Metaworld.git@c822f28f582ba1ad49eb5dcf61016566f28003ba#egg=metaworld
|
||||
|
||||
Installation from master
|
||||
~~~~~~~~~~~~~~~~~~~~~~~~
|
||||
@ -70,4 +70,4 @@ Metaworld has to be installed manually with
|
||||
|
||||
.. code:: bash
|
||||
|
||||
pip install metaworld@git+https://github.com/Farama-Foundation/Metaworld.git@d155d0051630bb365ea6a824e02c66c068947439#egg=metaworld
|
||||
pip install metaworld@git+https://github.com/Farama-Foundation/Metaworld.git@c822f28f582ba1ad49eb5dcf61016566f28003ba#egg=metaworld
|
||||
|
||||
@ -1,6 +1,6 @@
|
||||
var DOCUMENTATION_OPTIONS = {
|
||||
URL_ROOT: document.getElementById("documentation_options").getAttribute('data-url_root'),
|
||||
VERSION: '0.2',
|
||||
VERSION: '0.3.0',
|
||||
LANGUAGE: 'en',
|
||||
COLLAPSE_INDEX: false,
|
||||
BUILDER: 'html',
|
||||
|
||||
4
docs/build/html/api.html
vendored
4
docs/build/html/api.html
vendored
@ -4,7 +4,7 @@
|
||||
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>API — Fancy Gym 0.2 documentation</title>
|
||||
<title>API — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="_static/style.css" type="text/css" />
|
||||
@ -41,7 +41,7 @@
|
||||
<img src="_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="search.html" method="get">
|
||||
|
||||
4
docs/build/html/envs/dmc.html
vendored
4
docs/build/html/envs/dmc.html
vendored
@ -4,7 +4,7 @@
|
||||
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>DeepMind Control (DMC) — Fancy Gym 0.2 documentation</title>
|
||||
<title>DeepMind Control (DMC) — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/style.css" type="text/css" />
|
||||
@ -41,7 +41,7 @@
|
||||
<img src="../_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
|
||||
|
||||
4
docs/build/html/envs/fancy/airhockey.html
vendored
4
docs/build/html/envs/fancy/airhockey.html
vendored
@ -4,7 +4,7 @@
|
||||
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>AirHockey — Fancy Gym 0.2 documentation</title>
|
||||
<title>AirHockey — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="../../_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../../_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../../_static/style.css" type="text/css" />
|
||||
@ -41,7 +41,7 @@
|
||||
<img src="../../_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="../../search.html" method="get">
|
||||
|
||||
@ -4,7 +4,7 @@
|
||||
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>Classic Control — Fancy Gym 0.2 documentation</title>
|
||||
<title>Classic Control — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="../../_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../../_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../../_static/style.css" type="text/css" />
|
||||
@ -41,7 +41,7 @@
|
||||
<img src="../../_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="../../search.html" method="get">
|
||||
|
||||
4
docs/build/html/envs/fancy/index.html
vendored
4
docs/build/html/envs/fancy/index.html
vendored
@ -4,7 +4,7 @@
|
||||
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>Fancy — Fancy Gym 0.2 documentation</title>
|
||||
<title>Fancy — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="../../_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../../_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../../_static/style.css" type="text/css" />
|
||||
@ -41,7 +41,7 @@
|
||||
<img src="../../_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="../../search.html" method="get">
|
||||
|
||||
47
docs/build/html/envs/fancy/mujoco.html
vendored
47
docs/build/html/envs/fancy/mujoco.html
vendored
@ -4,7 +4,7 @@
|
||||
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>Mujoco — Fancy Gym 0.2 documentation</title>
|
||||
<title>Mujoco — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="../../_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../../_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../../_static/style.css" type="text/css" />
|
||||
@ -41,7 +41,7 @@
|
||||
<img src="../../_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="../../search.html" method="get">
|
||||
@ -135,6 +135,12 @@
|
||||
<p>The observation space includes the sine and cosine values of the robotic joint angles, their velocities, and quaternion orientations for the end-effector and the box. The action space describes the applied torques for each joint.</p>
|
||||
<p>A composite reward function serves as the performance metric for the RL system. It accounts for the distance to the goal, the box’s orientation, maintaining a rod within the box, achieving the rod’s desired orientation, and includes penalties for joint position and velocity limit violations, as well as an action cost for energy expenditure.</p>
|
||||
<p>Variations of this environment are available, differing in reward structures and the optionality of randomizing the box’s initial position. These variations are purposefully designed to challenge RL algorithms, enhancing their generalization and adaptation capabilities. Temporally sparse environments only provide a reward at the last timestep. Spatially sparse environments only provide a reward, if the goal is almost reached, the box is close enought to the goal and somewhat correctly aligned.</p>
|
||||
<p>These environments all provide smoothness metrics as part of the return infos:</p>
|
||||
<ul class="simple">
|
||||
<li><p>mean_squared_jerk: Averages the square of jerk (rate of acceleration change) across the motion. Lower values indicate smoother movement.</p></li>
|
||||
<li><p>maximum_jerk: Identifies the highest jerk value encountered.</p></li>
|
||||
<li><p>dimensionless_jerk: Normalizes the summed squared jerk over the motion’s duration and peak velocity, offering a scale-independent metric of smoothness</p></li>
|
||||
</ul>
|
||||
<table class="docutils align-default">
|
||||
<thead>
|
||||
<tr class="row-odd"><th class="head"><p>Name</p></th>
|
||||
@ -228,8 +234,15 @@
|
||||
<td><p>7</p></td>
|
||||
<td><p>19</p></td>
|
||||
</tr>
|
||||
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">fancy/TableTennisRndRobot-v0</span></code></p></td>
|
||||
<td><p>Table Tennis task with random initial robot joint positions *</p></td>
|
||||
<td><p>350</p></td>
|
||||
<td><p>7</p></td>
|
||||
<td><p>19</p></td>
|
||||
</tr>
|
||||
</tbody>
|
||||
</table>
|
||||
<p>* Random initialization of robot joint position and speed can be enabled by providing <code class="docutils literal notranslate"><span class="pre">random_pos_scale</span></code> / <code class="docutils literal notranslate"><span class="pre">random_vel_scale</span></code> to make. <code class="docutils literal notranslate"><span class="pre">TableTennisRndRobot</span></code> is equivalent to <code class="docutils literal notranslate"><span class="pre">TableTennis4D</span></code> except, that <code class="docutils literal notranslate"><span class="pre">random_pos_scale</span></code> is set to 0.1 instead of 0 per default.</p>
|
||||
</section>
|
||||
<hr class="docutils" />
|
||||
<section id="beer-pong">
|
||||
@ -335,49 +348,55 @@
|
||||
<td><p>7</p></td>
|
||||
<td><p>22</p></td>
|
||||
</tr>
|
||||
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">fancy/HopperJump-v0</span></code></p></td>
|
||||
<td><p>Hopper Jump task with continuous rewards, based on Gymnasium’s <code class="docutils literal notranslate"><span class="pre">gym.envs.mujoco.Hopper</span></code></p></td>
|
||||
<td><p>250</p></td>
|
||||
<td><p>3</p></td>
|
||||
<td><p>15 / 16*</p></td>
|
||||
</tr>
|
||||
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">fancy/HopperJumpMarkov-v0</span></code></p></td>
|
||||
<td><p><code class="docutils literal notranslate"><span class="pre">fancy/HopperJump-v0</span></code>, but with an alternative reward that is markovian.</p></td>
|
||||
<td><p>250</p></td>
|
||||
<td><p>3</p></td>
|
||||
<td><p>15 / 16*</p></td>
|
||||
</tr>
|
||||
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">fancy/HopperJumpSparse-v0</span></code></p></td>
|
||||
<td><p>Hopper Jump task with sparse rewards, based on Gymnasium’s <code class="docutils literal notranslate"><span class="pre">gym.envs.mujoco.Hopper</span></code></p></td>
|
||||
<td><p>250</p></td>
|
||||
<td><p>3</p></td>
|
||||
<td><p>15 / 16*</p></td>
|
||||
</tr>
|
||||
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">fancy/HopperJump-v0</span></code></p></td>
|
||||
<td><p>Hopper Jump task with continuous rewards, based on Gymnasium’s <code class="docutils literal notranslate"><span class="pre">gym.envs.mujoco.Hopper</span></code></p></td>
|
||||
<td><p>250</p></td>
|
||||
<td><p>3</p></td>
|
||||
<td><p>15 / 16*</p></td>
|
||||
</tr>
|
||||
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">fancy/AntJump-v0</span></code></p></td>
|
||||
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">fancy/AntJump-v0</span></code></p></td>
|
||||
<td><p>Ant Jump task, based on Gymnasium’s <code class="docutils literal notranslate"><span class="pre">gym.envs.mujoco.Ant</span></code></p></td>
|
||||
<td><p>200</p></td>
|
||||
<td><p>8</p></td>
|
||||
<td><p>119</p></td>
|
||||
</tr>
|
||||
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">fancy/HalfCheetahJump-v0</span></code></p></td>
|
||||
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">fancy/HalfCheetahJump-v0</span></code></p></td>
|
||||
<td><p>HalfCheetah Jump task, based on Gymnasium’s <code class="docutils literal notranslate"><span class="pre">gym.envs.mujoco.HalfCheetah</span></code></p></td>
|
||||
<td><p>100</p></td>
|
||||
<td><p>6</p></td>
|
||||
<td><p>112</p></td>
|
||||
</tr>
|
||||
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">fancy/HopperJumpOnBox-v0</span></code></p></td>
|
||||
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">fancy/HopperJumpOnBox-v0</span></code></p></td>
|
||||
<td><p>Hopper Jump on Box task, based on Gymnasium’s <code class="docutils literal notranslate"><span class="pre">gym.envs.mujoco.Hopper</span></code></p></td>
|
||||
<td><p>250</p></td>
|
||||
<td><p>4</p></td>
|
||||
<td><p>16 / 100*</p></td>
|
||||
</tr>
|
||||
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">fancy/HopperThrow-v0</span></code></p></td>
|
||||
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">fancy/HopperThrow-v0</span></code></p></td>
|
||||
<td><p>Hopper Throw task, based on Gymnasium’s <code class="docutils literal notranslate"><span class="pre">gym.envs.mujoco.Hopper</span></code></p></td>
|
||||
<td><p>250</p></td>
|
||||
<td><p>3</p></td>
|
||||
<td><p>18 / 100*</p></td>
|
||||
</tr>
|
||||
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">fancy/HopperThrowInBasket-v0</span></code></p></td>
|
||||
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">fancy/HopperThrowInBasket-v0</span></code></p></td>
|
||||
<td><p>Hopper Throw in Basket task, based on Gymnasium’s <code class="docutils literal notranslate"><span class="pre">gym.envs.mujoco.Hopper</span></code></p></td>
|
||||
<td><p>250</p></td>
|
||||
<td><p>3</p></td>
|
||||
<td><p>18 / 100*</p></td>
|
||||
</tr>
|
||||
<tr class="row-odd"><td><p><code class="docutils literal notranslate"><span class="pre">fancy/Walker2DJump-v0</span></code></p></td>
|
||||
<tr class="row-even"><td><p><code class="docutils literal notranslate"><span class="pre">fancy/Walker2DJump-v0</span></code></p></td>
|
||||
<td><p>Walker 2D Jump task, based on Gymnasium’s <code class="docutils literal notranslate"><span class="pre">gym.envs.mujoco.Walker2d</span></code></p></td>
|
||||
<td><p>300</p></td>
|
||||
<td><p>6</p></td>
|
||||
|
||||
4
docs/build/html/envs/meta.html
vendored
4
docs/build/html/envs/meta.html
vendored
@ -4,7 +4,7 @@
|
||||
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>Metaworld — Fancy Gym 0.2 documentation</title>
|
||||
<title>Metaworld — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/style.css" type="text/css" />
|
||||
@ -41,7 +41,7 @@
|
||||
<img src="../_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
|
||||
|
||||
4
docs/build/html/envs/open_ai.html
vendored
4
docs/build/html/envs/open_ai.html
vendored
@ -4,7 +4,7 @@
|
||||
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>Gymnasium — Fancy Gym 0.2 documentation</title>
|
||||
<title>Gymnasium — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/style.css" type="text/css" />
|
||||
@ -41,7 +41,7 @@
|
||||
<img src="../_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
|
||||
|
||||
120
docs/build/html/examples/dmc.html
vendored
120
docs/build/html/examples/dmc.html
vendored
@ -4,7 +4,7 @@
|
||||
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>DeepMind Control Examples — Fancy Gym 0.2 documentation</title>
|
||||
<title>DeepMind Control Examples — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/style.css" type="text/css" />
|
||||
@ -41,7 +41,7 @@
|
||||
<img src="../_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
|
||||
@ -126,7 +126,7 @@
|
||||
<span class="linenos"> 17</span><span class="sd"> Returns:</span>
|
||||
<span class="linenos"> 18</span>
|
||||
<span class="linenos"> 19</span><span class="sd"> """</span>
|
||||
<span class="linenos"> 20</span> <span class="n">env</span> <span class="o">=</span> <span class="n">gym</span><span class="o">.</span><span class="n">make</span><span class="p">(</span><span class="n">env_id</span><span class="p">)</span>
|
||||
<span class="linenos"> 20</span> <span class="n">env</span> <span class="o">=</span> <span class="n">gym</span><span class="o">.</span><span class="n">make</span><span class="p">(</span><span class="n">env_id</span><span class="p">,</span> <span class="n">render_mode</span><span class="o">=</span><span class="s1">'human'</span> <span class="k">if</span> <span class="n">render</span> <span class="k">else</span> <span class="kc">None</span><span class="p">)</span>
|
||||
<span class="linenos"> 21</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos"> 22</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="n">seed</span><span class="p">)</span>
|
||||
<span class="linenos"> 23</span> <span class="nb">print</span><span class="p">(</span><span class="s2">"observation shape:"</span><span class="p">,</span> <span class="n">env</span><span class="o">.</span><span class="n">observation_space</span><span class="o">.</span><span class="n">shape</span><span class="p">)</span>
|
||||
@ -135,7 +135,7 @@
|
||||
<span class="linenos"> 26</span> <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">iterations</span><span class="p">):</span>
|
||||
<span class="linenos"> 27</span> <span class="n">ac</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">action_space</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
|
||||
<span class="linenos"> 28</span> <span class="k">if</span> <span class="n">render</span><span class="p">:</span>
|
||||
<span class="linenos"> 29</span> <span class="n">env</span><span class="o">.</span><span class="n">render</span><span class="p">(</span><span class="n">mode</span><span class="o">=</span><span class="s2">"human"</span><span class="p">)</span>
|
||||
<span class="linenos"> 29</span> <span class="n">env</span><span class="o">.</span><span class="n">render</span><span class="p">()</span>
|
||||
<span class="linenos"> 30</span> <span class="n">obs</span><span class="p">,</span> <span class="n">reward</span><span class="p">,</span> <span class="n">terminated</span><span class="p">,</span> <span class="n">truncated</span><span class="p">,</span> <span class="n">info</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">step</span><span class="p">(</span><span class="n">ac</span><span class="p">)</span>
|
||||
<span class="linenos"> 31</span> <span class="n">rewards</span> <span class="o">+=</span> <span class="n">reward</span>
|
||||
<span class="linenos"> 32</span>
|
||||
@ -193,58 +193,68 @@
|
||||
<span class="linenos"> 84</span> <span class="c1"># basis_generator_kwargs = {'basis_generator_type': 'rbf',</span>
|
||||
<span class="linenos"> 85</span> <span class="c1"># 'num_basis': 5</span>
|
||||
<span class="linenos"> 86</span> <span class="c1"># }</span>
|
||||
<span class="linenos"> 87</span> <span class="n">env</span> <span class="o">=</span> <span class="n">fancy_gym</span><span class="o">.</span><span class="n">make_bb</span><span class="p">(</span><span class="n">env_id</span><span class="o">=</span><span class="n">base_env_id</span><span class="p">,</span> <span class="n">wrappers</span><span class="o">=</span><span class="n">wrappers</span><span class="p">,</span> <span class="n">black_box_kwargs</span><span class="o">=</span><span class="p">{},</span>
|
||||
<span class="linenos"> 88</span> <span class="n">traj_gen_kwargs</span><span class="o">=</span><span class="n">trajectory_generator_kwargs</span><span class="p">,</span> <span class="n">controller_kwargs</span><span class="o">=</span><span class="n">controller_kwargs</span><span class="p">,</span>
|
||||
<span class="linenos"> 89</span> <span class="n">phase_kwargs</span><span class="o">=</span><span class="n">phase_generator_kwargs</span><span class="p">,</span> <span class="n">basis_kwargs</span><span class="o">=</span><span class="n">basis_generator_kwargs</span><span class="p">,</span>
|
||||
<span class="linenos"> 90</span> <span class="n">seed</span><span class="o">=</span><span class="n">seed</span><span class="p">)</span>
|
||||
<span class="linenos"> 91</span>
|
||||
<span class="linenos"> 92</span> <span class="c1"># This renders the full MP trajectory</span>
|
||||
<span class="linenos"> 93</span> <span class="c1"># It is only required to call render() once in the beginning, which renders every consecutive trajectory.</span>
|
||||
<span class="linenos"> 94</span> <span class="c1"># Resetting to no rendering, can be achieved by render(mode=None).</span>
|
||||
<span class="linenos"> 95</span> <span class="c1"># It is also possible to change them mode multiple times when</span>
|
||||
<span class="linenos"> 96</span> <span class="c1"># e.g. only every nth trajectory should be displayed.</span>
|
||||
<span class="linenos"> 97</span> <span class="k">if</span> <span class="n">render</span><span class="p">:</span>
|
||||
<span class="linenos"> 98</span> <span class="n">env</span><span class="o">.</span><span class="n">render</span><span class="p">(</span><span class="n">mode</span><span class="o">=</span><span class="s2">"human"</span><span class="p">)</span>
|
||||
<span class="linenos"> 99</span>
|
||||
<span class="linenos">100</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos">101</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">102</span>
|
||||
<span class="linenos">103</span> <span class="c1"># number of samples/full trajectories (multiple environment steps)</span>
|
||||
<span class="linenos">104</span> <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">iterations</span><span class="p">):</span>
|
||||
<span class="linenos">105</span> <span class="n">ac</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">action_space</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
|
||||
<span class="linenos">106</span> <span class="n">obs</span><span class="p">,</span> <span class="n">reward</span><span class="p">,</span> <span class="n">terminated</span><span class="p">,</span> <span class="n">truncated</span><span class="p">,</span> <span class="n">info</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">step</span><span class="p">(</span><span class="n">ac</span><span class="p">)</span>
|
||||
<span class="linenos">107</span> <span class="n">rewards</span> <span class="o">+=</span> <span class="n">reward</span>
|
||||
<span class="linenos">108</span>
|
||||
<span class="linenos">109</span> <span class="k">if</span> <span class="n">terminated</span> <span class="ow">or</span> <span class="n">truncated</span><span class="p">:</span>
|
||||
<span class="linenos">110</span> <span class="nb">print</span><span class="p">(</span><span class="n">base_env_id</span><span class="p">,</span> <span class="n">rewards</span><span class="p">)</span>
|
||||
<span class="linenos">111</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos">112</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">113</span>
|
||||
<span class="linenos">114</span> <span class="n">env</span><span class="o">.</span><span class="n">close</span><span class="p">()</span>
|
||||
<span class="linenos">115</span> <span class="k">del</span> <span class="n">env</span>
|
||||
<span class="linenos">116</span>
|
||||
<span class="linenos"> 87</span> <span class="n">base_env</span> <span class="o">=</span> <span class="n">gym</span><span class="o">.</span><span class="n">make</span><span class="p">(</span><span class="n">base_env_id</span><span class="p">,</span> <span class="n">render_mode</span><span class="o">=</span><span class="s1">'human'</span> <span class="k">if</span> <span class="n">render</span> <span class="k">else</span> <span class="kc">None</span><span class="p">)</span>
|
||||
<span class="linenos"> 88</span> <span class="n">env</span> <span class="o">=</span> <span class="n">fancy_gym</span><span class="o">.</span><span class="n">make_bb</span><span class="p">(</span><span class="n">env</span><span class="o">=</span><span class="n">base_env</span><span class="p">,</span> <span class="n">wrappers</span><span class="o">=</span><span class="n">wrappers</span><span class="p">,</span> <span class="n">black_box_kwargs</span><span class="o">=</span><span class="p">{},</span>
|
||||
<span class="linenos"> 89</span> <span class="n">traj_gen_kwargs</span><span class="o">=</span><span class="n">trajectory_generator_kwargs</span><span class="p">,</span> <span class="n">controller_kwargs</span><span class="o">=</span><span class="n">controller_kwargs</span><span class="p">,</span>
|
||||
<span class="linenos"> 90</span> <span class="n">phase_kwargs</span><span class="o">=</span><span class="n">phase_generator_kwargs</span><span class="p">,</span> <span class="n">basis_kwargs</span><span class="o">=</span><span class="n">basis_generator_kwargs</span><span class="p">,</span>
|
||||
<span class="linenos"> 91</span> <span class="n">seed</span><span class="o">=</span><span class="n">seed</span><span class="p">)</span>
|
||||
<span class="linenos"> 92</span>
|
||||
<span class="linenos"> 93</span> <span class="c1"># This renders the full MP trajectory</span>
|
||||
<span class="linenos"> 94</span> <span class="c1"># It is only required to call render() once in the beginning, which renders every consecutive trajectory.</span>
|
||||
<span class="linenos"> 95</span> <span class="c1"># Resetting to no rendering, can be achieved by render(mode=None).</span>
|
||||
<span class="linenos"> 96</span> <span class="c1"># It is also possible to change them mode multiple times when</span>
|
||||
<span class="linenos"> 97</span> <span class="c1"># e.g. only every nth trajectory should be displayed.</span>
|
||||
<span class="linenos"> 98</span> <span class="k">if</span> <span class="n">render</span><span class="p">:</span>
|
||||
<span class="linenos"> 99</span> <span class="n">env</span><span class="o">.</span><span class="n">render</span><span class="p">()</span>
|
||||
<span class="linenos">100</span>
|
||||
<span class="linenos">101</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos">102</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">103</span>
|
||||
<span class="linenos">104</span> <span class="c1"># number of samples/full trajectories (multiple environment steps)</span>
|
||||
<span class="linenos">105</span> <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">iterations</span><span class="p">):</span>
|
||||
<span class="linenos">106</span> <span class="n">ac</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">action_space</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
|
||||
<span class="linenos">107</span> <span class="n">obs</span><span class="p">,</span> <span class="n">reward</span><span class="p">,</span> <span class="n">terminated</span><span class="p">,</span> <span class="n">truncated</span><span class="p">,</span> <span class="n">info</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">step</span><span class="p">(</span><span class="n">ac</span><span class="p">)</span>
|
||||
<span class="linenos">108</span> <span class="n">rewards</span> <span class="o">+=</span> <span class="n">reward</span>
|
||||
<span class="linenos">109</span>
|
||||
<span class="linenos">110</span> <span class="k">if</span> <span class="n">terminated</span> <span class="ow">or</span> <span class="n">truncated</span><span class="p">:</span>
|
||||
<span class="linenos">111</span> <span class="nb">print</span><span class="p">(</span><span class="n">base_env_id</span><span class="p">,</span> <span class="n">rewards</span><span class="p">)</span>
|
||||
<span class="linenos">112</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos">113</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">114</span>
|
||||
<span class="linenos">115</span> <span class="n">env</span><span class="o">.</span><span class="n">close</span><span class="p">()</span>
|
||||
<span class="linenos">116</span> <span class="k">del</span> <span class="n">env</span>
|
||||
<span class="linenos">117</span>
|
||||
<span class="linenos">118</span><span class="k">if</span> <span class="vm">__name__</span> <span class="o">==</span> <span class="s1">'__main__'</span><span class="p">:</span>
|
||||
<span class="linenos">119</span> <span class="c1"># Disclaimer: DMC environments require the seed to be specified in the beginning.</span>
|
||||
<span class="linenos">120</span> <span class="c1"># Adjusting it afterwards with env.seed() is not recommended as it does not affect the underlying physics.</span>
|
||||
<span class="linenos">121</span>
|
||||
<span class="linenos">122</span> <span class="c1"># For rendering DMC</span>
|
||||
<span class="linenos">123</span> <span class="c1"># export MUJOCO_GL="osmesa"</span>
|
||||
<span class="linenos">124</span> <span class="n">render</span> <span class="o">=</span> <span class="kc">True</span>
|
||||
<span class="linenos">125</span>
|
||||
<span class="linenos">126</span> <span class="c1"># # Standard DMC Suite tasks</span>
|
||||
<span class="linenos">127</span> <span class="n">example_dmc</span><span class="p">(</span><span class="s2">"dm_control/fish-swim"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1000</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">128</span> <span class="c1">#</span>
|
||||
<span class="linenos">129</span> <span class="c1"># # Manipulation tasks</span>
|
||||
<span class="linenos">130</span> <span class="c1"># # Disclaimer: The vision versions are currently not integrated and yield an error</span>
|
||||
<span class="linenos">131</span> <span class="n">example_dmc</span><span class="p">(</span><span class="s2">"dm_control/manipulation-reach_site_features"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">250</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">132</span> <span class="c1">#</span>
|
||||
<span class="linenos">133</span> <span class="c1"># # Gym + DMC hybrid task provided in the MP framework</span>
|
||||
<span class="linenos">134</span> <span class="n">example_dmc</span><span class="p">(</span><span class="s2">"dm_control_ProMP/ball_in_cup-catch-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">135</span>
|
||||
<span class="linenos">136</span> <span class="c1"># Custom DMC task # Different seed, because the episode is longer for this example and the name+seed combo is</span>
|
||||
<span class="linenos">137</span> <span class="c1"># already registered above</span>
|
||||
<span class="linenos">138</span> <span class="n">example_custom_dmc_and_mp</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="mi">11</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">118</span><span class="k">def</span> <span class="nf">main</span><span class="p">(</span><span class="n">render</span> <span class="o">=</span> <span class="kc">False</span><span class="p">):</span>
|
||||
<span class="linenos">119</span> <span class="c1"># # Standard DMC Suite tasks</span>
|
||||
<span class="linenos">120</span> <span class="n">example_dmc</span><span class="p">(</span><span class="s2">"dm_control/fish-swim"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1000</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">121</span> <span class="c1">#</span>
|
||||
<span class="linenos">122</span> <span class="c1"># # Manipulation tasks</span>
|
||||
<span class="linenos">123</span> <span class="c1"># # Disclaimer: The vision versions are currently not integrated and yield an error</span>
|
||||
<span class="linenos">124</span> <span class="n">example_dmc</span><span class="p">(</span><span class="s2">"dm_control/reach_site_features"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">250</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">125</span> <span class="c1">#</span>
|
||||
<span class="linenos">126</span> <span class="c1"># # Gym + DMC hybrid task provided in the MP framework</span>
|
||||
<span class="linenos">127</span> <span class="n">example_dmc</span><span class="p">(</span><span class="s2">"dm_control_ProMP/ball_in_cup-catch-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">128</span>
|
||||
<span class="linenos">129</span> <span class="c1"># Custom DMC task # Different seed, because the episode is longer for this example and the name+seed combo is</span>
|
||||
<span class="linenos">130</span> <span class="c1"># already registered above</span>
|
||||
<span class="linenos">131</span> <span class="n">example_custom_dmc_and_mp</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="mi">11</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">132</span>
|
||||
<span class="linenos">133</span> <span class="c1"># # Standard DMC Suite tasks</span>
|
||||
<span class="linenos">134</span> <span class="n">example_dmc</span><span class="p">(</span><span class="s2">"dm_control/fish-swim"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1000</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">135</span> <span class="c1">#</span>
|
||||
<span class="linenos">136</span> <span class="c1"># # Manipulation tasks</span>
|
||||
<span class="linenos">137</span> <span class="c1"># # Disclaimer: The vision versions are currently not integrated and yield an error</span>
|
||||
<span class="linenos">138</span> <span class="n">example_dmc</span><span class="p">(</span><span class="s2">"dm_control/reach_site_features"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">250</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">139</span> <span class="c1">#</span>
|
||||
<span class="linenos">140</span> <span class="c1"># # Gym + DMC hybrid task provided in the MP framework</span>
|
||||
<span class="linenos">141</span> <span class="n">example_dmc</span><span class="p">(</span><span class="s2">"dm_control_ProMP/ball_in_cup-catch-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">142</span>
|
||||
<span class="linenos">143</span> <span class="c1"># Custom DMC task # Different seed, because the episode is longer for this example and the name+seed combo is</span>
|
||||
<span class="linenos">144</span> <span class="c1"># already registered above</span>
|
||||
<span class="linenos">145</span> <span class="n">example_custom_dmc_and_mp</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="mi">11</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">146</span>
|
||||
<span class="linenos">147</span><span class="k">if</span> <span class="vm">__name__</span> <span class="o">==</span> <span class="s1">'__main__'</span><span class="p">:</span>
|
||||
<span class="linenos">148</span> <span class="n">main</span><span class="p">()</span>
|
||||
</pre></div>
|
||||
</div>
|
||||
</section>
|
||||
|
||||
28
docs/build/html/examples/general.html
vendored
28
docs/build/html/examples/general.html
vendored
@ -4,7 +4,7 @@
|
||||
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>General Usage Examples — Fancy Gym 0.2 documentation</title>
|
||||
<title>General Usage Examples — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/style.css" type="text/css" />
|
||||
@ -41,7 +41,7 @@
|
||||
<img src="../_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
|
||||
@ -130,7 +130,7 @@
|
||||
<span class="linenos"> 21</span>
|
||||
<span class="linenos"> 22</span><span class="sd"> """</span>
|
||||
<span class="linenos"> 23</span>
|
||||
<span class="linenos"> 24</span> <span class="n">env</span> <span class="o">=</span> <span class="n">gym</span><span class="o">.</span><span class="n">make</span><span class="p">(</span><span class="n">env_id</span><span class="p">)</span>
|
||||
<span class="linenos"> 24</span> <span class="n">env</span> <span class="o">=</span> <span class="n">gym</span><span class="o">.</span><span class="n">make</span><span class="p">(</span><span class="n">env_id</span><span class="p">,</span> <span class="n">render_mode</span><span class="o">=</span><span class="s1">'human'</span> <span class="k">if</span> <span class="n">render</span> <span class="k">else</span> <span class="kc">None</span><span class="p">)</span>
|
||||
<span class="linenos"> 25</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos"> 26</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="n">seed</span><span class="p">)</span>
|
||||
<span class="linenos"> 27</span> <span class="nb">print</span><span class="p">(</span><span class="s2">"Observation shape: "</span><span class="p">,</span> <span class="n">env</span><span class="o">.</span><span class="n">observation_space</span><span class="o">.</span><span class="n">shape</span><span class="p">)</span>
|
||||
@ -194,21 +194,21 @@
|
||||
<span class="linenos"> 85</span> <span class="c1"># do not return values above threshold</span>
|
||||
<span class="linenos"> 86</span> <span class="k">return</span> <span class="o">*</span><span class="nb">map</span><span class="p">(</span><span class="k">lambda</span> <span class="n">v</span><span class="p">:</span> <span class="n">np</span><span class="o">.</span><span class="n">stack</span><span class="p">(</span><span class="n">v</span><span class="p">)[:</span><span class="n">n_samples</span><span class="p">],</span> <span class="n">buffer</span><span class="o">.</span><span class="n">values</span><span class="p">()),</span>
|
||||
<span class="linenos"> 87</span>
|
||||
<span class="linenos"> 88</span>
|
||||
<span class="linenos"> 89</span><span class="k">if</span> <span class="vm">__name__</span> <span class="o">==</span> <span class="s1">'__main__'</span><span class="p">:</span>
|
||||
<span class="linenos"> 90</span> <span class="n">render</span> <span class="o">=</span> <span class="kc">True</span>
|
||||
<span class="linenos"> 88</span><span class="k">def</span> <span class="nf">main</span><span class="p">(</span><span class="n">render</span> <span class="o">=</span> <span class="kc">False</span><span class="p">):</span>
|
||||
<span class="linenos"> 89</span> <span class="c1"># Basic gym task</span>
|
||||
<span class="linenos"> 90</span> <span class="n">example_general</span><span class="p">(</span><span class="s2">"Pendulum-v1"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">200</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos"> 91</span>
|
||||
<span class="linenos"> 92</span> <span class="c1"># Basic gym task</span>
|
||||
<span class="linenos"> 93</span> <span class="n">example_general</span><span class="p">(</span><span class="s2">"Pendulum-v1"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">200</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos"> 92</span> <span class="c1"># Mujoco task from framework</span>
|
||||
<span class="linenos"> 93</span> <span class="n">example_general</span><span class="p">(</span><span class="s2">"fancy/Reacher5d-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">200</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos"> 94</span>
|
||||
<span class="linenos"> 95</span> <span class="c1"># Mujoco task from framework</span>
|
||||
<span class="linenos"> 96</span> <span class="n">example_general</span><span class="p">(</span><span class="s2">"fancy/Reacher5d-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">200</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos"> 95</span> <span class="c1"># # OpenAI Mujoco task</span>
|
||||
<span class="linenos"> 96</span> <span class="n">example_general</span><span class="p">(</span><span class="s2">"HalfCheetah-v2"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos"> 97</span>
|
||||
<span class="linenos"> 98</span> <span class="c1"># # OpenAI Mujoco task</span>
|
||||
<span class="linenos"> 99</span> <span class="n">example_general</span><span class="p">(</span><span class="s2">"HalfCheetah-v2"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos"> 98</span> <span class="c1"># Vectorized multiprocessing environments</span>
|
||||
<span class="linenos"> 99</span> <span class="c1"># example_async(env_id="HoleReacher-v0", n_cpu=2, seed=int('533D', 16), n_samples=2 * 200)</span>
|
||||
<span class="linenos">100</span>
|
||||
<span class="linenos">101</span> <span class="c1"># Vectorized multiprocessing environments</span>
|
||||
<span class="linenos">102</span> <span class="c1"># example_async(env_id="HoleReacher-v0", n_cpu=2, seed=int('533D', 16), n_samples=2 * 200)</span>
|
||||
<span class="linenos">101</span><span class="k">if</span> <span class="vm">__name__</span> <span class="o">==</span> <span class="s1">'__main__'</span><span class="p">:</span>
|
||||
<span class="linenos">102</span> <span class="n">main</span><span class="p">()</span>
|
||||
</pre></div>
|
||||
</div>
|
||||
</section>
|
||||
|
||||
205
docs/build/html/examples/metaworld.html
vendored
205
docs/build/html/examples/metaworld.html
vendored
@ -4,7 +4,7 @@
|
||||
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>Metaworld Examples — Fancy Gym 0.2 documentation</title>
|
||||
<title>Metaworld Examples — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/style.css" type="text/css" />
|
||||
@ -41,7 +41,7 @@
|
||||
<img src="../_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
|
||||
@ -111,7 +111,7 @@
|
||||
<span class="linenos"> 2</span><span class="kn">import</span> <span class="nn">fancy_gym</span>
|
||||
<span class="linenos"> 3</span>
|
||||
<span class="linenos"> 4</span>
|
||||
<span class="linenos"> 5</span><span class="k">def</span> <span class="nf">example_meta</span><span class="p">(</span><span class="n">env_id</span><span class="o">=</span><span class="s2">"fish-swim"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1000</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="kc">True</span><span class="p">):</span>
|
||||
<span class="linenos"> 5</span><span class="k">def</span> <span class="nf">example_meta</span><span class="p">(</span><span class="n">env_id</span><span class="o">=</span><span class="s2">"metaworld/button-press-v2"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1000</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="kc">True</span><span class="p">):</span>
|
||||
<span class="linenos"> 6</span><span class="w"> </span><span class="sd">"""</span>
|
||||
<span class="linenos"> 7</span><span class="sd"> Example for running a MetaWorld based env in the step based setting.</span>
|
||||
<span class="linenos"> 8</span><span class="sd"> The env_id has to be specified as `task_name-v2`. V1 versions are not supported and we always</span>
|
||||
@ -127,7 +127,7 @@
|
||||
<span class="linenos"> 18</span><span class="sd"> Returns:</span>
|
||||
<span class="linenos"> 19</span>
|
||||
<span class="linenos"> 20</span><span class="sd"> """</span>
|
||||
<span class="linenos"> 21</span> <span class="n">env</span> <span class="o">=</span> <span class="n">gym</span><span class="o">.</span><span class="n">make</span><span class="p">(</span><span class="n">env_id</span><span class="p">)</span>
|
||||
<span class="linenos"> 21</span> <span class="n">env</span> <span class="o">=</span> <span class="n">gym</span><span class="o">.</span><span class="n">make</span><span class="p">(</span><span class="n">env_id</span><span class="p">,</span> <span class="n">render_mode</span><span class="o">=</span><span class="s1">'human'</span> <span class="k">if</span> <span class="n">render</span> <span class="k">else</span> <span class="kc">None</span><span class="p">)</span>
|
||||
<span class="linenos"> 22</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos"> 23</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="n">seed</span><span class="p">)</span>
|
||||
<span class="linenos"> 24</span> <span class="nb">print</span><span class="p">(</span><span class="s2">"observation shape:"</span><span class="p">,</span> <span class="n">env</span><span class="o">.</span><span class="n">observation_space</span><span class="o">.</span><span class="n">shape</span><span class="p">)</span>
|
||||
@ -136,111 +136,104 @@
|
||||
<span class="linenos"> 27</span> <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">iterations</span><span class="p">):</span>
|
||||
<span class="linenos"> 28</span> <span class="n">ac</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">action_space</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
|
||||
<span class="linenos"> 29</span> <span class="k">if</span> <span class="n">render</span><span class="p">:</span>
|
||||
<span class="linenos"> 30</span> <span class="c1"># THIS NEEDS TO BE SET TO FALSE FOR NOW, BECAUSE THE INTERFACE FOR RENDERING IS DIFFERENT TO BASIC GYM</span>
|
||||
<span class="linenos"> 31</span> <span class="c1"># TODO: Remove this, when Metaworld fixes its interface.</span>
|
||||
<span class="linenos"> 32</span> <span class="n">env</span><span class="o">.</span><span class="n">render</span><span class="p">(</span><span class="kc">False</span><span class="p">)</span>
|
||||
<span class="linenos"> 33</span> <span class="n">obs</span><span class="p">,</span> <span class="n">reward</span><span class="p">,</span> <span class="n">terminated</span><span class="p">,</span> <span class="n">truncated</span><span class="p">,</span> <span class="n">info</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">step</span><span class="p">(</span><span class="n">ac</span><span class="p">)</span>
|
||||
<span class="linenos"> 34</span> <span class="n">rewards</span> <span class="o">+=</span> <span class="n">reward</span>
|
||||
<span class="linenos"> 35</span> <span class="k">if</span> <span class="n">terminated</span> <span class="ow">or</span> <span class="n">truncated</span><span class="p">:</span>
|
||||
<span class="linenos"> 36</span> <span class="nb">print</span><span class="p">(</span><span class="n">env_id</span><span class="p">,</span> <span class="n">rewards</span><span class="p">)</span>
|
||||
<span class="linenos"> 37</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos"> 38</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos"> 39</span>
|
||||
<span class="linenos"> 40</span> <span class="n">env</span><span class="o">.</span><span class="n">close</span><span class="p">()</span>
|
||||
<span class="linenos"> 41</span> <span class="k">del</span> <span class="n">env</span>
|
||||
<span class="linenos"> 42</span>
|
||||
<span class="linenos"> 43</span>
|
||||
<span class="linenos"> 44</span><span class="k">def</span> <span class="nf">example_custom_meta_and_mp</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="kc">True</span><span class="p">):</span>
|
||||
<span class="linenos"> 45</span><span class="w"> </span><span class="sd">"""</span>
|
||||
<span class="linenos"> 46</span><span class="sd"> Example for running a custom movement primitive based environments.</span>
|
||||
<span class="linenos"> 47</span><span class="sd"> Our already registered environments follow the same structure.</span>
|
||||
<span class="linenos"> 48</span><span class="sd"> Hence, this also allows to adjust hyperparameters of the movement primitives.</span>
|
||||
<span class="linenos"> 49</span><span class="sd"> Yet, we recommend the method above if you are just interested in chaining those parameters for existing tasks.</span>
|
||||
<span class="linenos"> 50</span><span class="sd"> We appreciate PRs for custom environments (especially MP wrappers of existing tasks)</span>
|
||||
<span class="linenos"> 51</span><span class="sd"> for our repo: https://github.com/ALRhub/fancy_gym/</span>
|
||||
<span class="linenos"> 52</span><span class="sd"> Args:</span>
|
||||
<span class="linenos"> 53</span><span class="sd"> seed: seed for deterministic behaviour (TODO: currently not working due to an issue in MetaWorld code)</span>
|
||||
<span class="linenos"> 54</span><span class="sd"> iterations: Number of rollout steps to run</span>
|
||||
<span class="linenos"> 55</span><span class="sd"> render: Render the episode (TODO: currently not working due to an issue in MetaWorld code)</span>
|
||||
<span class="linenos"> 30</span> <span class="n">env</span><span class="o">.</span><span class="n">render</span><span class="p">()</span>
|
||||
<span class="linenos"> 31</span> <span class="n">obs</span><span class="p">,</span> <span class="n">reward</span><span class="p">,</span> <span class="n">terminated</span><span class="p">,</span> <span class="n">truncated</span><span class="p">,</span> <span class="n">info</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">step</span><span class="p">(</span><span class="n">ac</span><span class="p">)</span>
|
||||
<span class="linenos"> 32</span> <span class="n">rewards</span> <span class="o">+=</span> <span class="n">reward</span>
|
||||
<span class="linenos"> 33</span> <span class="k">if</span> <span class="n">terminated</span> <span class="ow">or</span> <span class="n">truncated</span><span class="p">:</span>
|
||||
<span class="linenos"> 34</span> <span class="nb">print</span><span class="p">(</span><span class="n">env_id</span><span class="p">,</span> <span class="n">rewards</span><span class="p">)</span>
|
||||
<span class="linenos"> 35</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos"> 36</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="n">seed</span><span class="o">+</span><span class="n">i</span><span class="o">+</span><span class="mi">1</span><span class="p">)</span>
|
||||
<span class="linenos"> 37</span>
|
||||
<span class="linenos"> 38</span> <span class="n">env</span><span class="o">.</span><span class="n">close</span><span class="p">()</span>
|
||||
<span class="linenos"> 39</span> <span class="k">del</span> <span class="n">env</span>
|
||||
<span class="linenos"> 40</span>
|
||||
<span class="linenos"> 41</span>
|
||||
<span class="linenos"> 42</span><span class="k">def</span> <span class="nf">example_custom_meta_and_mp</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="kc">True</span><span class="p">):</span>
|
||||
<span class="linenos"> 43</span><span class="w"> </span><span class="sd">"""</span>
|
||||
<span class="linenos"> 44</span><span class="sd"> Example for running a custom movement primitive based environments.</span>
|
||||
<span class="linenos"> 45</span><span class="sd"> Our already registered environments follow the same structure.</span>
|
||||
<span class="linenos"> 46</span><span class="sd"> Hence, this also allows to adjust hyperparameters of the movement primitives.</span>
|
||||
<span class="linenos"> 47</span><span class="sd"> Yet, we recommend the method above if you are just interested in chaining those parameters for existing tasks.</span>
|
||||
<span class="linenos"> 48</span><span class="sd"> We appreciate PRs for custom environments (especially MP wrappers of existing tasks)</span>
|
||||
<span class="linenos"> 49</span><span class="sd"> for our repo: https://github.com/ALRhub/fancy_gym/</span>
|
||||
<span class="linenos"> 50</span><span class="sd"> Args:</span>
|
||||
<span class="linenos"> 51</span><span class="sd"> seed: seed for deterministic behaviour (TODO: currently not working due to an issue in MetaWorld code)</span>
|
||||
<span class="linenos"> 52</span><span class="sd"> iterations: Number of rollout steps to run</span>
|
||||
<span class="linenos"> 53</span><span class="sd"> render: Render the episode (TODO: currently not working due to an issue in MetaWorld code)</span>
|
||||
<span class="linenos"> 54</span>
|
||||
<span class="linenos"> 55</span><span class="sd"> Returns:</span>
|
||||
<span class="linenos"> 56</span>
|
||||
<span class="linenos"> 57</span><span class="sd"> Returns:</span>
|
||||
<span class="linenos"> 57</span><span class="sd"> """</span>
|
||||
<span class="linenos"> 58</span>
|
||||
<span class="linenos"> 59</span><span class="sd"> """</span>
|
||||
<span class="linenos"> 60</span>
|
||||
<span class="linenos"> 61</span> <span class="c1"># Base MetaWorld name, according to structure of above example</span>
|
||||
<span class="linenos"> 62</span> <span class="n">base_env_id</span> <span class="o">=</span> <span class="s2">"metaworld/button-press-v2"</span>
|
||||
<span class="linenos"> 63</span>
|
||||
<span class="linenos"> 64</span> <span class="c1"># Replace this wrapper with the custom wrapper for your environment by inheriting from the RawInterfaceWrapper.</span>
|
||||
<span class="linenos"> 65</span> <span class="c1"># You can also add other gym.Wrappers in case they are needed.</span>
|
||||
<span class="linenos"> 66</span> <span class="n">wrappers</span> <span class="o">=</span> <span class="p">[</span><span class="n">fancy_gym</span><span class="o">.</span><span class="n">meta</span><span class="o">.</span><span class="n">goal_object_change_mp_wrapper</span><span class="o">.</span><span class="n">MPWrapper</span><span class="p">]</span>
|
||||
<span class="linenos"> 67</span> <span class="c1"># # For a ProMP</span>
|
||||
<span class="linenos"> 68</span> <span class="c1"># trajectory_generator_kwargs = {'trajectory_generator_type': 'promp'}</span>
|
||||
<span class="linenos"> 69</span> <span class="c1"># phase_generator_kwargs = {'phase_generator_type': 'linear'}</span>
|
||||
<span class="linenos"> 70</span> <span class="c1"># controller_kwargs = {'controller_type': 'metaworld'}</span>
|
||||
<span class="linenos"> 71</span> <span class="c1"># basis_generator_kwargs = {'basis_generator_type': 'zero_rbf',</span>
|
||||
<span class="linenos"> 72</span> <span class="c1"># 'num_basis': 5,</span>
|
||||
<span class="linenos"> 73</span> <span class="c1"># 'num_basis_zero_start': 1</span>
|
||||
<span class="linenos"> 74</span> <span class="c1"># }</span>
|
||||
<span class="linenos"> 75</span>
|
||||
<span class="linenos"> 76</span> <span class="c1"># For a DMP</span>
|
||||
<span class="linenos"> 77</span> <span class="n">trajectory_generator_kwargs</span> <span class="o">=</span> <span class="p">{</span><span class="s1">'trajectory_generator_type'</span><span class="p">:</span> <span class="s1">'dmp'</span><span class="p">}</span>
|
||||
<span class="linenos"> 78</span> <span class="n">phase_generator_kwargs</span> <span class="o">=</span> <span class="p">{</span><span class="s1">'phase_generator_type'</span><span class="p">:</span> <span class="s1">'exp'</span><span class="p">,</span>
|
||||
<span class="linenos"> 79</span> <span class="s1">'alpha_phase'</span><span class="p">:</span> <span class="mi">2</span><span class="p">}</span>
|
||||
<span class="linenos"> 80</span> <span class="n">controller_kwargs</span> <span class="o">=</span> <span class="p">{</span><span class="s1">'controller_type'</span><span class="p">:</span> <span class="s1">'metaworld'</span><span class="p">}</span>
|
||||
<span class="linenos"> 81</span> <span class="n">basis_generator_kwargs</span> <span class="o">=</span> <span class="p">{</span><span class="s1">'basis_generator_type'</span><span class="p">:</span> <span class="s1">'rbf'</span><span class="p">,</span>
|
||||
<span class="linenos"> 82</span> <span class="s1">'num_basis'</span><span class="p">:</span> <span class="mi">5</span>
|
||||
<span class="linenos"> 83</span> <span class="p">}</span>
|
||||
<span class="linenos"> 84</span> <span class="n">env</span> <span class="o">=</span> <span class="n">fancy_gym</span><span class="o">.</span><span class="n">make_bb</span><span class="p">(</span><span class="n">env_id</span><span class="o">=</span><span class="n">base_env_id</span><span class="p">,</span> <span class="n">wrappers</span><span class="o">=</span><span class="n">wrappers</span><span class="p">,</span> <span class="n">black_box_kwargs</span><span class="o">=</span><span class="p">{},</span>
|
||||
<span class="linenos"> 85</span> <span class="n">traj_gen_kwargs</span><span class="o">=</span><span class="n">trajectory_generator_kwargs</span><span class="p">,</span> <span class="n">controller_kwargs</span><span class="o">=</span><span class="n">controller_kwargs</span><span class="p">,</span>
|
||||
<span class="linenos"> 86</span> <span class="n">phase_kwargs</span><span class="o">=</span><span class="n">phase_generator_kwargs</span><span class="p">,</span> <span class="n">basis_kwargs</span><span class="o">=</span><span class="n">basis_generator_kwargs</span><span class="p">,</span>
|
||||
<span class="linenos"> 87</span> <span class="n">seed</span><span class="o">=</span><span class="n">seed</span><span class="p">)</span>
|
||||
<span class="linenos"> 88</span>
|
||||
<span class="linenos"> 89</span> <span class="c1"># This renders the full MP trajectory</span>
|
||||
<span class="linenos"> 90</span> <span class="c1"># It is only required to call render() once in the beginning, which renders every consecutive trajectory.</span>
|
||||
<span class="linenos"> 91</span> <span class="c1"># Resetting to no rendering, can be achieved by render(mode=None).</span>
|
||||
<span class="linenos"> 92</span> <span class="c1"># It is also possible to change them mode multiple times when</span>
|
||||
<span class="linenos"> 93</span> <span class="c1"># e.g. only every nth trajectory should be displayed.</span>
|
||||
<span class="linenos"> 94</span> <span class="k">if</span> <span class="n">render</span><span class="p">:</span>
|
||||
<span class="linenos"> 95</span> <span class="k">raise</span> <span class="ne">ValueError</span><span class="p">(</span><span class="s2">"Metaworld render interface bug does not allow to render() fixes its interface. "</span>
|
||||
<span class="linenos"> 96</span> <span class="s2">"A temporary workaround is to alter their code in MujocoEnv render() from "</span>
|
||||
<span class="linenos"> 97</span> <span class="s2">"`if not offscreen` to `if not offscreen or offscreen == 'human'`."</span><span class="p">)</span>
|
||||
<span class="linenos"> 98</span> <span class="c1"># TODO: Remove this, when Metaworld fixes its interface.</span>
|
||||
<span class="linenos"> 99</span> <span class="c1"># env.render(mode="human")</span>
|
||||
<span class="linenos">100</span>
|
||||
<span class="linenos">101</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos">102</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">103</span>
|
||||
<span class="linenos">104</span> <span class="c1"># number of samples/full trajectories (multiple environment steps)</span>
|
||||
<span class="linenos">105</span> <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">iterations</span><span class="p">):</span>
|
||||
<span class="linenos">106</span> <span class="n">ac</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">action_space</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
|
||||
<span class="linenos">107</span> <span class="n">obs</span><span class="p">,</span> <span class="n">reward</span><span class="p">,</span> <span class="n">terminated</span><span class="p">,</span> <span class="n">truncated</span><span class="p">,</span> <span class="n">info</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">step</span><span class="p">(</span><span class="n">ac</span><span class="p">)</span>
|
||||
<span class="linenos">108</span> <span class="n">rewards</span> <span class="o">+=</span> <span class="n">reward</span>
|
||||
<span class="linenos"> 59</span> <span class="c1"># Base MetaWorld name, according to structure of above example</span>
|
||||
<span class="linenos"> 60</span> <span class="n">base_env_id</span> <span class="o">=</span> <span class="s2">"metaworld/button-press-v2"</span>
|
||||
<span class="linenos"> 61</span>
|
||||
<span class="linenos"> 62</span> <span class="c1"># Replace this wrapper with the custom wrapper for your environment by inheriting from the RawInterfaceWrapper.</span>
|
||||
<span class="linenos"> 63</span> <span class="c1"># You can also add other gym.Wrappers in case they are needed.</span>
|
||||
<span class="linenos"> 64</span> <span class="n">wrappers</span> <span class="o">=</span> <span class="p">[</span><span class="n">fancy_gym</span><span class="o">.</span><span class="n">meta</span><span class="o">.</span><span class="n">goal_object_change_mp_wrapper</span><span class="o">.</span><span class="n">MPWrapper</span><span class="p">]</span>
|
||||
<span class="linenos"> 65</span> <span class="c1"># # For a ProMP</span>
|
||||
<span class="linenos"> 66</span> <span class="c1"># trajectory_generator_kwargs = {'trajectory_generator_type': 'promp'}</span>
|
||||
<span class="linenos"> 67</span> <span class="c1"># phase_generator_kwargs = {'phase_generator_type': 'linear'}</span>
|
||||
<span class="linenos"> 68</span> <span class="c1"># controller_kwargs = {'controller_type': 'metaworld'}</span>
|
||||
<span class="linenos"> 69</span> <span class="c1"># basis_generator_kwargs = {'basis_generator_type': 'zero_rbf',</span>
|
||||
<span class="linenos"> 70</span> <span class="c1"># 'num_basis': 5,</span>
|
||||
<span class="linenos"> 71</span> <span class="c1"># 'num_basis_zero_start': 1</span>
|
||||
<span class="linenos"> 72</span> <span class="c1"># }</span>
|
||||
<span class="linenos"> 73</span>
|
||||
<span class="linenos"> 74</span> <span class="c1"># For a DMP</span>
|
||||
<span class="linenos"> 75</span> <span class="n">trajectory_generator_kwargs</span> <span class="o">=</span> <span class="p">{</span><span class="s1">'trajectory_generator_type'</span><span class="p">:</span> <span class="s1">'dmp'</span><span class="p">}</span>
|
||||
<span class="linenos"> 76</span> <span class="n">phase_generator_kwargs</span> <span class="o">=</span> <span class="p">{</span><span class="s1">'phase_generator_type'</span><span class="p">:</span> <span class="s1">'exp'</span><span class="p">,</span>
|
||||
<span class="linenos"> 77</span> <span class="s1">'alpha_phase'</span><span class="p">:</span> <span class="mi">2</span><span class="p">}</span>
|
||||
<span class="linenos"> 78</span> <span class="n">controller_kwargs</span> <span class="o">=</span> <span class="p">{</span><span class="s1">'controller_type'</span><span class="p">:</span> <span class="s1">'metaworld'</span><span class="p">}</span>
|
||||
<span class="linenos"> 79</span> <span class="n">basis_generator_kwargs</span> <span class="o">=</span> <span class="p">{</span><span class="s1">'basis_generator_type'</span><span class="p">:</span> <span class="s1">'rbf'</span><span class="p">,</span>
|
||||
<span class="linenos"> 80</span> <span class="s1">'num_basis'</span><span class="p">:</span> <span class="mi">5</span>
|
||||
<span class="linenos"> 81</span> <span class="p">}</span>
|
||||
<span class="linenos"> 82</span> <span class="n">base_env</span> <span class="o">=</span> <span class="n">gym</span><span class="o">.</span><span class="n">make</span><span class="p">(</span><span class="n">base_env_id</span><span class="p">,</span> <span class="n">render_mode</span><span class="o">=</span><span class="s1">'human'</span> <span class="k">if</span> <span class="n">render</span> <span class="k">else</span> <span class="kc">None</span><span class="p">)</span>
|
||||
<span class="linenos"> 83</span> <span class="n">env</span> <span class="o">=</span> <span class="n">fancy_gym</span><span class="o">.</span><span class="n">make_bb</span><span class="p">(</span><span class="n">env</span><span class="o">=</span><span class="n">base_env</span><span class="p">,</span> <span class="n">wrappers</span><span class="o">=</span><span class="n">wrappers</span><span class="p">,</span> <span class="n">black_box_kwargs</span><span class="o">=</span><span class="p">{},</span>
|
||||
<span class="linenos"> 84</span> <span class="n">traj_gen_kwargs</span><span class="o">=</span><span class="n">trajectory_generator_kwargs</span><span class="p">,</span> <span class="n">controller_kwargs</span><span class="o">=</span><span class="n">controller_kwargs</span><span class="p">,</span>
|
||||
<span class="linenos"> 85</span> <span class="n">phase_kwargs</span><span class="o">=</span><span class="n">phase_generator_kwargs</span><span class="p">,</span> <span class="n">basis_kwargs</span><span class="o">=</span><span class="n">basis_generator_kwargs</span><span class="p">,</span>
|
||||
<span class="linenos"> 86</span> <span class="n">seed</span><span class="o">=</span><span class="n">seed</span><span class="p">)</span>
|
||||
<span class="linenos"> 87</span>
|
||||
<span class="linenos"> 88</span> <span class="c1"># This renders the full MP trajectory</span>
|
||||
<span class="linenos"> 89</span> <span class="c1"># It is only required to call render() once in the beginning, which renders every consecutive trajectory.</span>
|
||||
<span class="linenos"> 90</span> <span class="c1"># Resetting to no rendering, can be achieved by render(mode=None).</span>
|
||||
<span class="linenos"> 91</span> <span class="c1"># It is also possible to change them mode multiple times when</span>
|
||||
<span class="linenos"> 92</span> <span class="c1"># e.g. only every nth trajectory should be displayed.</span>
|
||||
<span class="linenos"> 93</span> <span class="k">if</span> <span class="n">render</span><span class="p">:</span>
|
||||
<span class="linenos"> 94</span> <span class="n">env</span><span class="o">.</span><span class="n">render</span><span class="p">()</span>
|
||||
<span class="linenos"> 95</span>
|
||||
<span class="linenos"> 96</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos"> 97</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="n">seed</span><span class="p">)</span>
|
||||
<span class="linenos"> 98</span>
|
||||
<span class="linenos"> 99</span> <span class="c1"># number of samples/full trajectories (multiple environment steps)</span>
|
||||
<span class="linenos">100</span> <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">iterations</span><span class="p">):</span>
|
||||
<span class="linenos">101</span> <span class="n">ac</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">action_space</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
|
||||
<span class="linenos">102</span> <span class="n">obs</span><span class="p">,</span> <span class="n">reward</span><span class="p">,</span> <span class="n">terminated</span><span class="p">,</span> <span class="n">truncated</span><span class="p">,</span> <span class="n">info</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">step</span><span class="p">(</span><span class="n">ac</span><span class="p">)</span>
|
||||
<span class="linenos">103</span> <span class="n">rewards</span> <span class="o">+=</span> <span class="n">reward</span>
|
||||
<span class="linenos">104</span>
|
||||
<span class="linenos">105</span> <span class="k">if</span> <span class="n">terminated</span> <span class="ow">or</span> <span class="n">truncated</span><span class="p">:</span>
|
||||
<span class="linenos">106</span> <span class="nb">print</span><span class="p">(</span><span class="n">base_env_id</span><span class="p">,</span> <span class="n">rewards</span><span class="p">)</span>
|
||||
<span class="linenos">107</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos">108</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="n">seed</span><span class="o">+</span><span class="n">i</span><span class="o">+</span><span class="mi">1</span><span class="p">)</span>
|
||||
<span class="linenos">109</span>
|
||||
<span class="linenos">110</span> <span class="k">if</span> <span class="n">terminated</span> <span class="ow">or</span> <span class="n">truncated</span><span class="p">:</span>
|
||||
<span class="linenos">111</span> <span class="nb">print</span><span class="p">(</span><span class="n">base_env_id</span><span class="p">,</span> <span class="n">rewards</span><span class="p">)</span>
|
||||
<span class="linenos">112</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos">113</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">114</span>
|
||||
<span class="linenos">115</span> <span class="n">env</span><span class="o">.</span><span class="n">close</span><span class="p">()</span>
|
||||
<span class="linenos">116</span> <span class="k">del</span> <span class="n">env</span>
|
||||
<span class="linenos">117</span>
|
||||
<span class="linenos">118</span>
|
||||
<span class="linenos">119</span><span class="k">if</span> <span class="vm">__name__</span> <span class="o">==</span> <span class="s1">'__main__'</span><span class="p">:</span>
|
||||
<span class="linenos">120</span> <span class="c1"># Disclaimer: MetaWorld environments require the seed to be specified in the beginning.</span>
|
||||
<span class="linenos">121</span> <span class="c1"># Adjusting it afterwards with env.seed() is not recommended as it may not affect the underlying behavior.</span>
|
||||
<span class="linenos">122</span>
|
||||
<span class="linenos">123</span> <span class="c1"># For rendering it might be necessary to specify your OpenGL installation</span>
|
||||
<span class="linenos">124</span> <span class="c1"># export LD_PRELOAD=/usr/lib/x86_64-linux-gnu/libGLEW.so</span>
|
||||
<span class="linenos">125</span> <span class="n">render</span> <span class="o">=</span> <span class="kc">False</span>
|
||||
<span class="linenos">126</span>
|
||||
<span class="linenos">127</span> <span class="c1"># # Standard Meta world tasks</span>
|
||||
<span class="linenos">128</span> <span class="n">example_meta</span><span class="p">(</span><span class="s2">"metaworld/button-press-v2"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">500</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">129</span>
|
||||
<span class="linenos">130</span> <span class="c1"># # MP + MetaWorld hybrid task provided in the our framework</span>
|
||||
<span class="linenos">131</span> <span class="n">example_meta</span><span class="p">(</span><span class="s2">"metaworld_ProMP/ButtonPress-v2"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">132</span> <span class="c1">#</span>
|
||||
<span class="linenos">133</span> <span class="c1"># # Custom MetaWorld task</span>
|
||||
<span class="linenos">134</span> <span class="n">example_custom_meta_and_mp</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">110</span> <span class="n">env</span><span class="o">.</span><span class="n">close</span><span class="p">()</span>
|
||||
<span class="linenos">111</span> <span class="k">del</span> <span class="n">env</span>
|
||||
<span class="linenos">112</span>
|
||||
<span class="linenos">113</span><span class="k">def</span> <span class="nf">main</span><span class="p">(</span><span class="n">render</span> <span class="o">=</span> <span class="kc">False</span><span class="p">):</span>
|
||||
<span class="linenos">114</span> <span class="c1"># For rendering it might be necessary to specify your OpenGL installation</span>
|
||||
<span class="linenos">115</span> <span class="c1"># export LD_PRELOAD=/usr/lib/x86_64-linux-gnu/libGLEW.so</span>
|
||||
<span class="linenos">116</span>
|
||||
<span class="linenos">117</span> <span class="c1"># # Standard Meta world tasks</span>
|
||||
<span class="linenos">118</span> <span class="n">example_meta</span><span class="p">(</span><span class="s2">"metaworld/button-press-v2"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">500</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">119</span>
|
||||
<span class="linenos">120</span> <span class="c1"># # MP + MetaWorld hybrid task provided in the our framework</span>
|
||||
<span class="linenos">121</span> <span class="n">example_meta</span><span class="p">(</span><span class="s2">"metaworld_ProMP/button-press-v2"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">122</span> <span class="c1">#</span>
|
||||
<span class="linenos">123</span> <span class="c1"># # Custom MetaWorld task</span>
|
||||
<span class="linenos">124</span> <span class="n">example_custom_meta_and_mp</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">125</span>
|
||||
<span class="linenos">126</span><span class="k">if</span> <span class="vm">__name__</span> <span class="o">==</span> <span class="s1">'__main__'</span><span class="p">:</span>
|
||||
<span class="linenos">127</span> <span class="n">main</span><span class="p">()</span>
|
||||
</pre></div>
|
||||
</div>
|
||||
</section>
|
||||
|
||||
479
docs/build/html/examples/movement_primitives.html
vendored
479
docs/build/html/examples/movement_primitives.html
vendored
@ -4,7 +4,7 @@
|
||||
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>Movement Primitives Examples — Fancy Gym 0.2 documentation</title>
|
||||
<title>Movement Primitives Examples — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/style.css" type="text/css" />
|
||||
@ -41,7 +41,7 @@
|
||||
<img src="../_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
|
||||
@ -135,252 +135,253 @@
|
||||
<span class="linenos"> 26</span> <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">iterations</span><span class="p">):</span>
|
||||
<span class="linenos"> 27</span>
|
||||
<span class="linenos"> 28</span> <span class="k">if</span> <span class="n">render</span> <span class="ow">and</span> <span class="n">i</span> <span class="o">%</span> <span class="mi">1</span> <span class="o">==</span> <span class="mi">0</span><span class="p">:</span>
|
||||
<span class="linenos"> 29</span> <span class="n">env</span><span class="o">.</span><span class="n">render</span><span class="p">()</span>
|
||||
<span class="linenos"> 30</span>
|
||||
<span class="linenos"> 31</span> <span class="c1"># Now the action space is not the raw action but the parametrization of the trajectory generator,</span>
|
||||
<span class="linenos"> 32</span> <span class="c1"># such as a ProMP</span>
|
||||
<span class="linenos"> 33</span> <span class="n">ac</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">action_space</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
|
||||
<span class="linenos"> 34</span> <span class="c1"># This executes a full trajectory and gives back the context (obs) of the last step in the trajectory, or the</span>
|
||||
<span class="linenos"> 35</span> <span class="c1"># full observation space of the last step, if replanning/sub-trajectory learning is used. The 'reward' is equal</span>
|
||||
<span class="linenos"> 36</span> <span class="c1"># to the return of a trajectory. Default is the sum over the step-wise rewards.</span>
|
||||
<span class="linenos"> 37</span> <span class="n">obs</span><span class="p">,</span> <span class="n">reward</span><span class="p">,</span> <span class="n">terminated</span><span class="p">,</span> <span class="n">truncated</span><span class="p">,</span> <span class="n">info</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">step</span><span class="p">(</span><span class="n">ac</span><span class="p">)</span>
|
||||
<span class="linenos"> 38</span> <span class="c1"># Aggregated returns</span>
|
||||
<span class="linenos"> 39</span> <span class="n">returns</span> <span class="o">+=</span> <span class="n">reward</span>
|
||||
<span class="linenos"> 40</span>
|
||||
<span class="linenos"> 41</span> <span class="k">if</span> <span class="n">terminated</span> <span class="ow">or</span> <span class="n">truncated</span><span class="p">:</span>
|
||||
<span class="linenos"> 42</span> <span class="nb">print</span><span class="p">(</span><span class="n">reward</span><span class="p">)</span>
|
||||
<span class="linenos"> 43</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos"> 44</span> <span class="n">env</span><span class="o">.</span><span class="n">close</span><span class="p">()</span>
|
||||
<span class="linenos"> 45</span>
|
||||
<span class="linenos"> 46</span>
|
||||
<span class="linenos"> 47</span><span class="k">def</span> <span class="nf">example_custom_mp</span><span class="p">(</span><span class="n">env_name</span><span class="o">=</span><span class="s2">"fancy_ProMP/Reacher5d-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="kc">True</span><span class="p">):</span>
|
||||
<span class="linenos"> 48</span><span class="w"> </span><span class="sd">"""</span>
|
||||
<span class="linenos"> 49</span><span class="sd"> Example for running a custom movement primitive based environments.</span>
|
||||
<span class="linenos"> 50</span><span class="sd"> Our already registered environments follow the same structure.</span>
|
||||
<span class="linenos"> 51</span><span class="sd"> Hence, this also allows to adjust hyperparameters of the movement primitives.</span>
|
||||
<span class="linenos"> 52</span><span class="sd"> Yet, we recommend the method above if you are just interested in changing those parameters for existing tasks.</span>
|
||||
<span class="linenos"> 53</span><span class="sd"> We appreciate PRs for custom environments (especially MP wrappers of existing tasks) </span>
|
||||
<span class="linenos"> 54</span><span class="sd"> for our repo: https://github.com/ALRhub/fancy_gym/</span>
|
||||
<span class="linenos"> 55</span><span class="sd"> Args:</span>
|
||||
<span class="linenos"> 56</span><span class="sd"> seed: seed</span>
|
||||
<span class="linenos"> 57</span><span class="sd"> iterations: Number of rollout steps to run</span>
|
||||
<span class="linenos"> 58</span><span class="sd"> render: Render the episode</span>
|
||||
<span class="linenos"> 59</span>
|
||||
<span class="linenos"> 60</span><span class="sd"> Returns:</span>
|
||||
<span class="linenos"> 29</span> <span class="c1"># This renders the full MP trajectory</span>
|
||||
<span class="linenos"> 30</span> <span class="c1"># It is only required to call render() once in the beginning, which renders every consecutive trajectory.</span>
|
||||
<span class="linenos"> 31</span> <span class="n">env</span><span class="o">.</span><span class="n">render</span><span class="p">()</span>
|
||||
<span class="linenos"> 32</span>
|
||||
<span class="linenos"> 33</span> <span class="c1"># Now the action space is not the raw action but the parametrization of the trajectory generator,</span>
|
||||
<span class="linenos"> 34</span> <span class="c1"># such as a ProMP</span>
|
||||
<span class="linenos"> 35</span> <span class="n">ac</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">action_space</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
|
||||
<span class="linenos"> 36</span> <span class="c1"># This executes a full trajectory and gives back the context (obs) of the last step in the trajectory, or the</span>
|
||||
<span class="linenos"> 37</span> <span class="c1"># full observation space of the last step, if replanning/sub-trajectory learning is used. The 'reward' is equal</span>
|
||||
<span class="linenos"> 38</span> <span class="c1"># to the return of a trajectory. Default is the sum over the step-wise rewards.</span>
|
||||
<span class="linenos"> 39</span> <span class="n">obs</span><span class="p">,</span> <span class="n">reward</span><span class="p">,</span> <span class="n">terminated</span><span class="p">,</span> <span class="n">truncated</span><span class="p">,</span> <span class="n">info</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">step</span><span class="p">(</span><span class="n">ac</span><span class="p">)</span>
|
||||
<span class="linenos"> 40</span> <span class="c1"># Aggregated returns</span>
|
||||
<span class="linenos"> 41</span> <span class="n">returns</span> <span class="o">+=</span> <span class="n">reward</span>
|
||||
<span class="linenos"> 42</span>
|
||||
<span class="linenos"> 43</span> <span class="k">if</span> <span class="n">terminated</span> <span class="ow">or</span> <span class="n">truncated</span><span class="p">:</span>
|
||||
<span class="linenos"> 44</span> <span class="nb">print</span><span class="p">(</span><span class="n">reward</span><span class="p">)</span>
|
||||
<span class="linenos"> 45</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos"> 46</span> <span class="n">env</span><span class="o">.</span><span class="n">close</span><span class="p">()</span>
|
||||
<span class="linenos"> 47</span>
|
||||
<span class="linenos"> 48</span>
|
||||
<span class="linenos"> 49</span><span class="k">def</span> <span class="nf">example_custom_mp</span><span class="p">(</span><span class="n">env_name</span><span class="o">=</span><span class="s2">"fancy_ProMP/Reacher5d-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="kc">True</span><span class="p">):</span>
|
||||
<span class="linenos"> 50</span><span class="w"> </span><span class="sd">"""</span>
|
||||
<span class="linenos"> 51</span><span class="sd"> Example for running a custom movement primitive based environments.</span>
|
||||
<span class="linenos"> 52</span><span class="sd"> Our already registered environments follow the same structure.</span>
|
||||
<span class="linenos"> 53</span><span class="sd"> Hence, this also allows to adjust hyperparameters of the movement primitives.</span>
|
||||
<span class="linenos"> 54</span><span class="sd"> Yet, we recommend the method above if you are just interested in changing those parameters for existing tasks.</span>
|
||||
<span class="linenos"> 55</span><span class="sd"> We appreciate PRs for custom environments (especially MP wrappers of existing tasks) </span>
|
||||
<span class="linenos"> 56</span><span class="sd"> for our repo: https://github.com/ALRhub/fancy_gym/</span>
|
||||
<span class="linenos"> 57</span><span class="sd"> Args:</span>
|
||||
<span class="linenos"> 58</span><span class="sd"> seed: seed</span>
|
||||
<span class="linenos"> 59</span><span class="sd"> iterations: Number of rollout steps to run</span>
|
||||
<span class="linenos"> 60</span><span class="sd"> render: Render the episode</span>
|
||||
<span class="linenos"> 61</span>
|
||||
<span class="linenos"> 62</span><span class="sd"> """</span>
|
||||
<span class="linenos"> 63</span> <span class="c1"># Changing the arguments of the black box env is possible by providing them to gym through mp_config_override.</span>
|
||||
<span class="linenos"> 64</span> <span class="c1"># E.g. here for way to many basis functions</span>
|
||||
<span class="linenos"> 65</span> <span class="n">env</span> <span class="o">=</span> <span class="n">gym</span><span class="o">.</span><span class="n">make</span><span class="p">(</span><span class="n">env_name</span><span class="p">,</span> <span class="n">seed</span><span class="p">,</span> <span class="n">mp_config_override</span><span class="o">=</span><span class="p">{</span><span class="s1">'basis_generator_kwargs'</span><span class="p">:</span> <span class="p">{</span><span class="s1">'num_basis'</span><span class="p">:</span> <span class="mi">1000</span><span class="p">}},</span> <span class="n">render_mode</span><span class="o">=</span><span class="s1">'human'</span> <span class="k">if</span> <span class="n">render</span> <span class="k">else</span> <span class="kc">None</span><span class="p">)</span>
|
||||
<span class="linenos"> 66</span>
|
||||
<span class="linenos"> 67</span> <span class="n">returns</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos"> 68</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos"> 69</span>
|
||||
<span class="linenos"> 70</span> <span class="c1"># This time rendering every trajectory</span>
|
||||
<span class="linenos"> 71</span> <span class="k">if</span> <span class="n">render</span><span class="p">:</span>
|
||||
<span class="linenos"> 72</span> <span class="n">env</span><span class="o">.</span><span class="n">render</span><span class="p">()</span>
|
||||
<span class="linenos"> 73</span>
|
||||
<span class="linenos"> 74</span> <span class="c1"># number of samples/full trajectories (multiple environment steps)</span>
|
||||
<span class="linenos"> 75</span> <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">iterations</span><span class="p">):</span>
|
||||
<span class="linenos"> 76</span> <span class="n">ac</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">action_space</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
|
||||
<span class="linenos"> 77</span> <span class="n">obs</span><span class="p">,</span> <span class="n">reward</span><span class="p">,</span> <span class="n">terminated</span><span class="p">,</span> <span class="n">truncated</span><span class="p">,</span> <span class="n">info</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">step</span><span class="p">(</span><span class="n">ac</span><span class="p">)</span>
|
||||
<span class="linenos"> 78</span> <span class="n">returns</span> <span class="o">+=</span> <span class="n">reward</span>
|
||||
<span class="linenos"> 79</span>
|
||||
<span class="linenos"> 80</span> <span class="k">if</span> <span class="n">terminated</span> <span class="ow">or</span> <span class="n">truncated</span><span class="p">:</span>
|
||||
<span class="linenos"> 81</span> <span class="nb">print</span><span class="p">(</span><span class="n">i</span><span class="p">,</span> <span class="n">reward</span><span class="p">)</span>
|
||||
<span class="linenos"> 82</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos"> 83</span>
|
||||
<span class="linenos"> 84</span> <span class="n">env</span><span class="o">.</span><span class="n">close</span><span class="p">()</span>
|
||||
<span class="linenos"> 85</span> <span class="k">return</span> <span class="n">obs</span>
|
||||
<span class="linenos"> 86</span>
|
||||
<span class="linenos"> 87</span><span class="k">class</span> <span class="nc">Custom_MPWrapper</span><span class="p">(</span><span class="n">fancy_gym</span><span class="o">.</span><span class="n">envs</span><span class="o">.</span><span class="n">mujoco</span><span class="o">.</span><span class="n">reacher</span><span class="o">.</span><span class="n">MPWrapper</span><span class="p">):</span>
|
||||
<span class="linenos"> 88</span> <span class="n">mp_config</span> <span class="o">=</span> <span class="p">{</span>
|
||||
<span class="linenos"> 89</span> <span class="s1">'ProMP'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos"> 90</span> <span class="s1">'trajectory_generator_kwargs'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos"> 91</span> <span class="s1">'trajectory_generator_type'</span><span class="p">:</span> <span class="s1">'promp'</span><span class="p">,</span>
|
||||
<span class="linenos"> 92</span> <span class="s1">'weights_scale'</span><span class="p">:</span> <span class="mi">2</span>
|
||||
<span class="linenos"> 93</span> <span class="p">},</span>
|
||||
<span class="linenos"> 94</span> <span class="s1">'phase_generator_kwargs'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos"> 95</span> <span class="s1">'phase_generator_type'</span><span class="p">:</span> <span class="s1">'linear'</span>
|
||||
<span class="linenos"> 96</span> <span class="p">},</span>
|
||||
<span class="linenos"> 97</span> <span class="s1">'controller_kwargs'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos"> 98</span> <span class="s1">'controller_type'</span><span class="p">:</span> <span class="s1">'velocity'</span>
|
||||
<span class="linenos"> 99</span> <span class="p">},</span>
|
||||
<span class="linenos">100</span> <span class="s1">'basis_generator_kwargs'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos">101</span> <span class="s1">'basis_generator_type'</span><span class="p">:</span> <span class="s1">'zero_rbf'</span><span class="p">,</span>
|
||||
<span class="linenos">102</span> <span class="s1">'num_basis'</span><span class="p">:</span> <span class="mi">5</span><span class="p">,</span>
|
||||
<span class="linenos">103</span> <span class="s1">'num_basis_zero_start'</span><span class="p">:</span> <span class="mi">1</span>
|
||||
<span class="linenos">104</span> <span class="p">}</span>
|
||||
<span class="linenos">105</span> <span class="p">},</span>
|
||||
<span class="linenos">106</span> <span class="s1">'DMP'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos">107</span> <span class="s1">'trajectory_generator_kwargs'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos">108</span> <span class="s1">'trajectory_generator_type'</span><span class="p">:</span> <span class="s1">'dmp'</span><span class="p">,</span>
|
||||
<span class="linenos">109</span> <span class="s1">'weights_scale'</span><span class="p">:</span> <span class="mi">500</span>
|
||||
<span class="linenos">110</span> <span class="p">},</span>
|
||||
<span class="linenos">111</span> <span class="s1">'phase_generator_kwargs'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos">112</span> <span class="s1">'phase_generator_type'</span><span class="p">:</span> <span class="s1">'exp'</span><span class="p">,</span>
|
||||
<span class="linenos">113</span> <span class="s1">'alpha_phase'</span><span class="p">:</span> <span class="mf">2.5</span>
|
||||
<span class="linenos">114</span> <span class="p">},</span>
|
||||
<span class="linenos">115</span> <span class="s1">'controller_kwargs'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos">116</span> <span class="s1">'controller_type'</span><span class="p">:</span> <span class="s1">'velocity'</span>
|
||||
<span class="linenos">117</span> <span class="p">},</span>
|
||||
<span class="linenos">118</span> <span class="s1">'basis_generator_kwargs'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos">119</span> <span class="s1">'basis_generator_type'</span><span class="p">:</span> <span class="s1">'rbf'</span><span class="p">,</span>
|
||||
<span class="linenos">120</span> <span class="s1">'num_basis'</span><span class="p">:</span> <span class="mi">5</span>
|
||||
<span class="linenos">121</span> <span class="p">}</span>
|
||||
<span class="linenos">122</span> <span class="p">}</span>
|
||||
<span class="linenos"> 62</span><span class="sd"> Returns:</span>
|
||||
<span class="linenos"> 63</span>
|
||||
<span class="linenos"> 64</span><span class="sd"> """</span>
|
||||
<span class="linenos"> 65</span> <span class="c1"># Changing the arguments of the black box env is possible by providing them to gym through mp_config_override.</span>
|
||||
<span class="linenos"> 66</span> <span class="c1"># E.g. here for way to many basis functions</span>
|
||||
<span class="linenos"> 67</span> <span class="n">env</span> <span class="o">=</span> <span class="n">gym</span><span class="o">.</span><span class="n">make</span><span class="p">(</span><span class="n">env_name</span><span class="p">,</span> <span class="n">seed</span><span class="p">,</span> <span class="n">mp_config_override</span><span class="o">=</span><span class="p">{</span><span class="s1">'basis_generator_kwargs'</span><span class="p">:</span> <span class="p">{</span><span class="s1">'num_basis'</span><span class="p">:</span> <span class="mi">1000</span><span class="p">}},</span> <span class="n">render_mode</span><span class="o">=</span><span class="s1">'human'</span> <span class="k">if</span> <span class="n">render</span> <span class="k">else</span> <span class="kc">None</span><span class="p">)</span>
|
||||
<span class="linenos"> 68</span>
|
||||
<span class="linenos"> 69</span> <span class="n">returns</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos"> 70</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos"> 71</span>
|
||||
<span class="linenos"> 72</span> <span class="c1"># This time rendering every trajectory</span>
|
||||
<span class="linenos"> 73</span> <span class="k">if</span> <span class="n">render</span><span class="p">:</span>
|
||||
<span class="linenos"> 74</span> <span class="n">env</span><span class="o">.</span><span class="n">render</span><span class="p">()</span>
|
||||
<span class="linenos"> 75</span>
|
||||
<span class="linenos"> 76</span> <span class="c1"># number of samples/full trajectories (multiple environment steps)</span>
|
||||
<span class="linenos"> 77</span> <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">iterations</span><span class="p">):</span>
|
||||
<span class="linenos"> 78</span> <span class="n">ac</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">action_space</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
|
||||
<span class="linenos"> 79</span> <span class="n">obs</span><span class="p">,</span> <span class="n">reward</span><span class="p">,</span> <span class="n">terminated</span><span class="p">,</span> <span class="n">truncated</span><span class="p">,</span> <span class="n">info</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">step</span><span class="p">(</span><span class="n">ac</span><span class="p">)</span>
|
||||
<span class="linenos"> 80</span> <span class="n">returns</span> <span class="o">+=</span> <span class="n">reward</span>
|
||||
<span class="linenos"> 81</span>
|
||||
<span class="linenos"> 82</span> <span class="k">if</span> <span class="n">terminated</span> <span class="ow">or</span> <span class="n">truncated</span><span class="p">:</span>
|
||||
<span class="linenos"> 83</span> <span class="nb">print</span><span class="p">(</span><span class="n">i</span><span class="p">,</span> <span class="n">reward</span><span class="p">)</span>
|
||||
<span class="linenos"> 84</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos"> 85</span>
|
||||
<span class="linenos"> 86</span> <span class="n">env</span><span class="o">.</span><span class="n">close</span><span class="p">()</span>
|
||||
<span class="linenos"> 87</span> <span class="k">return</span> <span class="n">obs</span>
|
||||
<span class="linenos"> 88</span>
|
||||
<span class="linenos"> 89</span><span class="k">class</span> <span class="nc">Custom_MPWrapper</span><span class="p">(</span><span class="n">fancy_gym</span><span class="o">.</span><span class="n">envs</span><span class="o">.</span><span class="n">mujoco</span><span class="o">.</span><span class="n">reacher</span><span class="o">.</span><span class="n">MPWrapper</span><span class="p">):</span>
|
||||
<span class="linenos"> 90</span> <span class="n">mp_config</span> <span class="o">=</span> <span class="p">{</span>
|
||||
<span class="linenos"> 91</span> <span class="s1">'ProMP'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos"> 92</span> <span class="s1">'trajectory_generator_kwargs'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos"> 93</span> <span class="s1">'trajectory_generator_type'</span><span class="p">:</span> <span class="s1">'promp'</span><span class="p">,</span>
|
||||
<span class="linenos"> 94</span> <span class="s1">'weights_scale'</span><span class="p">:</span> <span class="mi">2</span>
|
||||
<span class="linenos"> 95</span> <span class="p">},</span>
|
||||
<span class="linenos"> 96</span> <span class="s1">'phase_generator_kwargs'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos"> 97</span> <span class="s1">'phase_generator_type'</span><span class="p">:</span> <span class="s1">'linear'</span>
|
||||
<span class="linenos"> 98</span> <span class="p">},</span>
|
||||
<span class="linenos"> 99</span> <span class="s1">'controller_kwargs'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos">100</span> <span class="s1">'controller_type'</span><span class="p">:</span> <span class="s1">'velocity'</span>
|
||||
<span class="linenos">101</span> <span class="p">},</span>
|
||||
<span class="linenos">102</span> <span class="s1">'basis_generator_kwargs'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos">103</span> <span class="s1">'basis_generator_type'</span><span class="p">:</span> <span class="s1">'zero_rbf'</span><span class="p">,</span>
|
||||
<span class="linenos">104</span> <span class="s1">'num_basis'</span><span class="p">:</span> <span class="mi">5</span><span class="p">,</span>
|
||||
<span class="linenos">105</span> <span class="s1">'num_basis_zero_start'</span><span class="p">:</span> <span class="mi">1</span>
|
||||
<span class="linenos">106</span> <span class="p">}</span>
|
||||
<span class="linenos">107</span> <span class="p">},</span>
|
||||
<span class="linenos">108</span> <span class="s1">'DMP'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos">109</span> <span class="s1">'trajectory_generator_kwargs'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos">110</span> <span class="s1">'trajectory_generator_type'</span><span class="p">:</span> <span class="s1">'dmp'</span><span class="p">,</span>
|
||||
<span class="linenos">111</span> <span class="s1">'weights_scale'</span><span class="p">:</span> <span class="mi">500</span>
|
||||
<span class="linenos">112</span> <span class="p">},</span>
|
||||
<span class="linenos">113</span> <span class="s1">'phase_generator_kwargs'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos">114</span> <span class="s1">'phase_generator_type'</span><span class="p">:</span> <span class="s1">'exp'</span><span class="p">,</span>
|
||||
<span class="linenos">115</span> <span class="s1">'alpha_phase'</span><span class="p">:</span> <span class="mf">2.5</span>
|
||||
<span class="linenos">116</span> <span class="p">},</span>
|
||||
<span class="linenos">117</span> <span class="s1">'controller_kwargs'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos">118</span> <span class="s1">'controller_type'</span><span class="p">:</span> <span class="s1">'velocity'</span>
|
||||
<span class="linenos">119</span> <span class="p">},</span>
|
||||
<span class="linenos">120</span> <span class="s1">'basis_generator_kwargs'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos">121</span> <span class="s1">'basis_generator_type'</span><span class="p">:</span> <span class="s1">'rbf'</span><span class="p">,</span>
|
||||
<span class="linenos">122</span> <span class="s1">'num_basis'</span><span class="p">:</span> <span class="mi">5</span>
|
||||
<span class="linenos">123</span> <span class="p">}</span>
|
||||
<span class="linenos">124</span>
|
||||
<span class="linenos">125</span>
|
||||
<span class="linenos">126</span><span class="k">def</span> <span class="nf">example_fully_custom_mp</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="kc">True</span><span class="p">):</span>
|
||||
<span class="linenos">127</span><span class="w"> </span><span class="sd">"""</span>
|
||||
<span class="linenos">128</span><span class="sd"> Example for running a custom movement primitive based environments.</span>
|
||||
<span class="linenos">129</span><span class="sd"> Our already registered environments follow the same structure.</span>
|
||||
<span class="linenos">130</span><span class="sd"> Hence, this also allows to adjust hyperparameters of the movement primitives.</span>
|
||||
<span class="linenos">131</span><span class="sd"> Yet, we recommend the method above if you are just interested in changing those parameters for existing tasks.</span>
|
||||
<span class="linenos">132</span><span class="sd"> We appreciate PRs for custom environments (especially MP wrappers of existing tasks) </span>
|
||||
<span class="linenos">133</span><span class="sd"> for our repo: https://github.com/ALRhub/fancy_gym/</span>
|
||||
<span class="linenos">134</span><span class="sd"> Args:</span>
|
||||
<span class="linenos">135</span><span class="sd"> seed: seed</span>
|
||||
<span class="linenos">136</span><span class="sd"> iterations: Number of rollout steps to run</span>
|
||||
<span class="linenos">137</span><span class="sd"> render: Render the episode</span>
|
||||
<span class="linenos">138</span>
|
||||
<span class="linenos">139</span><span class="sd"> Returns:</span>
|
||||
<span class="linenos">124</span> <span class="p">}</span>
|
||||
<span class="linenos">125</span> <span class="p">}</span>
|
||||
<span class="linenos">126</span>
|
||||
<span class="linenos">127</span>
|
||||
<span class="linenos">128</span><span class="k">def</span> <span class="nf">example_fully_custom_mp</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="kc">True</span><span class="p">):</span>
|
||||
<span class="linenos">129</span><span class="w"> </span><span class="sd">"""</span>
|
||||
<span class="linenos">130</span><span class="sd"> Example for running a custom movement primitive based environments.</span>
|
||||
<span class="linenos">131</span><span class="sd"> Our already registered environments follow the same structure.</span>
|
||||
<span class="linenos">132</span><span class="sd"> Hence, this also allows to adjust hyperparameters of the movement primitives.</span>
|
||||
<span class="linenos">133</span><span class="sd"> Yet, we recommend the method above if you are just interested in changing those parameters for existing tasks.</span>
|
||||
<span class="linenos">134</span><span class="sd"> We appreciate PRs for custom environments (especially MP wrappers of existing tasks) </span>
|
||||
<span class="linenos">135</span><span class="sd"> for our repo: https://github.com/ALRhub/fancy_gym/</span>
|
||||
<span class="linenos">136</span><span class="sd"> Args:</span>
|
||||
<span class="linenos">137</span><span class="sd"> seed: seed</span>
|
||||
<span class="linenos">138</span><span class="sd"> iterations: Number of rollout steps to run</span>
|
||||
<span class="linenos">139</span><span class="sd"> render: Render the episode</span>
|
||||
<span class="linenos">140</span>
|
||||
<span class="linenos">141</span><span class="sd"> """</span>
|
||||
<span class="linenos">141</span><span class="sd"> Returns:</span>
|
||||
<span class="linenos">142</span>
|
||||
<span class="linenos">143</span> <span class="n">base_env_id</span> <span class="o">=</span> <span class="s2">"fancy/Reacher5d-v0"</span>
|
||||
<span class="linenos">144</span> <span class="n">custom_env_id</span> <span class="o">=</span> <span class="s2">"fancy/Reacher5d-Custom-v0"</span>
|
||||
<span class="linenos">145</span> <span class="n">custom_env_id_DMP</span> <span class="o">=</span> <span class="s2">"fancy_DMP/Reacher5d-Custom-v0"</span>
|
||||
<span class="linenos">146</span> <span class="n">custom_env_id_ProMP</span> <span class="o">=</span> <span class="s2">"fancy_ProMP/Reacher5d-Custom-v0"</span>
|
||||
<span class="linenos">147</span>
|
||||
<span class="linenos">148</span> <span class="n">fancy_gym</span><span class="o">.</span><span class="n">upgrade</span><span class="p">(</span><span class="n">custom_env_id</span><span class="p">,</span> <span class="n">mp_wrapper</span><span class="o">=</span><span class="n">Custom_MPWrapper</span><span class="p">,</span> <span class="n">add_mp_types</span><span class="o">=</span><span class="p">[</span><span class="s1">'ProMP'</span><span class="p">,</span> <span class="s1">'DMP'</span><span class="p">],</span> <span class="n">base_id</span><span class="o">=</span><span class="n">base_env_id</span><span class="p">)</span>
|
||||
<span class="linenos">143</span><span class="sd"> """</span>
|
||||
<span class="linenos">144</span>
|
||||
<span class="linenos">145</span> <span class="n">base_env_id</span> <span class="o">=</span> <span class="s2">"fancy/Reacher5d-v0"</span>
|
||||
<span class="linenos">146</span> <span class="n">custom_env_id</span> <span class="o">=</span> <span class="s2">"fancy/Reacher5d-Custom-v0"</span>
|
||||
<span class="linenos">147</span> <span class="n">custom_env_id_DMP</span> <span class="o">=</span> <span class="s2">"fancy_DMP/Reacher5d-Custom-v0"</span>
|
||||
<span class="linenos">148</span> <span class="n">custom_env_id_ProMP</span> <span class="o">=</span> <span class="s2">"fancy_ProMP/Reacher5d-Custom-v0"</span>
|
||||
<span class="linenos">149</span>
|
||||
<span class="linenos">150</span> <span class="n">env</span> <span class="o">=</span> <span class="n">gym</span><span class="o">.</span><span class="n">make</span><span class="p">(</span><span class="n">custom_env_id_ProMP</span><span class="p">,</span> <span class="n">render_mode</span><span class="o">=</span><span class="s1">'human'</span> <span class="k">if</span> <span class="n">render</span> <span class="k">else</span> <span class="kc">None</span><span class="p">)</span>
|
||||
<span class="linenos">150</span> <span class="n">fancy_gym</span><span class="o">.</span><span class="n">upgrade</span><span class="p">(</span><span class="n">custom_env_id</span><span class="p">,</span> <span class="n">mp_wrapper</span><span class="o">=</span><span class="n">Custom_MPWrapper</span><span class="p">,</span> <span class="n">add_mp_types</span><span class="o">=</span><span class="p">[</span><span class="s1">'ProMP'</span><span class="p">,</span> <span class="s1">'DMP'</span><span class="p">],</span> <span class="n">base_id</span><span class="o">=</span><span class="n">base_env_id</span><span class="p">)</span>
|
||||
<span class="linenos">151</span>
|
||||
<span class="linenos">152</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos">153</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">154</span>
|
||||
<span class="linenos">155</span> <span class="k">if</span> <span class="n">render</span><span class="p">:</span>
|
||||
<span class="linenos">156</span> <span class="n">env</span><span class="o">.</span><span class="n">render</span><span class="p">()</span>
|
||||
<span class="linenos">157</span>
|
||||
<span class="linenos">158</span> <span class="c1"># number of samples/full trajectories (multiple environment steps)</span>
|
||||
<span class="linenos">159</span> <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">iterations</span><span class="p">):</span>
|
||||
<span class="linenos">160</span> <span class="n">ac</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">action_space</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
|
||||
<span class="linenos">161</span> <span class="n">obs</span><span class="p">,</span> <span class="n">reward</span><span class="p">,</span> <span class="n">terminated</span><span class="p">,</span> <span class="n">truncated</span><span class="p">,</span> <span class="n">info</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">step</span><span class="p">(</span><span class="n">ac</span><span class="p">)</span>
|
||||
<span class="linenos">162</span> <span class="n">rewards</span> <span class="o">+=</span> <span class="n">reward</span>
|
||||
<span class="linenos">163</span>
|
||||
<span class="linenos">164</span> <span class="k">if</span> <span class="n">terminated</span> <span class="ow">or</span> <span class="n">truncated</span><span class="p">:</span>
|
||||
<span class="linenos">165</span> <span class="nb">print</span><span class="p">(</span><span class="n">rewards</span><span class="p">)</span>
|
||||
<span class="linenos">166</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos">167</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">168</span>
|
||||
<span class="linenos">169</span> <span class="k">try</span><span class="p">:</span> <span class="c1"># Some mujoco-based envs don't correlcty implement .close</span>
|
||||
<span class="linenos">170</span> <span class="n">env</span><span class="o">.</span><span class="n">close</span><span class="p">()</span>
|
||||
<span class="linenos">171</span> <span class="k">except</span><span class="p">:</span>
|
||||
<span class="linenos">172</span> <span class="k">pass</span>
|
||||
<span class="linenos">173</span>
|
||||
<span class="linenos">174</span>
|
||||
<span class="linenos">175</span><span class="k">def</span> <span class="nf">example_fully_custom_mp_alternative</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="kc">True</span><span class="p">):</span>
|
||||
<span class="linenos">176</span><span class="w"> </span><span class="sd">"""</span>
|
||||
<span class="linenos">177</span><span class="sd"> Instead of defining the mp_args in a new custom MP_Wrapper, they can also be provided during registration.</span>
|
||||
<span class="linenos">178</span><span class="sd"> Args:</span>
|
||||
<span class="linenos">179</span><span class="sd"> seed: seed</span>
|
||||
<span class="linenos">180</span><span class="sd"> iterations: Number of rollout steps to run</span>
|
||||
<span class="linenos">181</span><span class="sd"> render: Render the episode</span>
|
||||
<span class="linenos">182</span>
|
||||
<span class="linenos">183</span><span class="sd"> Returns:</span>
|
||||
<span class="linenos">152</span> <span class="n">env</span> <span class="o">=</span> <span class="n">gym</span><span class="o">.</span><span class="n">make</span><span class="p">(</span><span class="n">custom_env_id_ProMP</span><span class="p">,</span> <span class="n">render_mode</span><span class="o">=</span><span class="s1">'human'</span> <span class="k">if</span> <span class="n">render</span> <span class="k">else</span> <span class="kc">None</span><span class="p">)</span>
|
||||
<span class="linenos">153</span>
|
||||
<span class="linenos">154</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos">155</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">156</span>
|
||||
<span class="linenos">157</span> <span class="k">if</span> <span class="n">render</span><span class="p">:</span>
|
||||
<span class="linenos">158</span> <span class="n">env</span><span class="o">.</span><span class="n">render</span><span class="p">()</span>
|
||||
<span class="linenos">159</span>
|
||||
<span class="linenos">160</span> <span class="c1"># number of samples/full trajectories (multiple environment steps)</span>
|
||||
<span class="linenos">161</span> <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">iterations</span><span class="p">):</span>
|
||||
<span class="linenos">162</span> <span class="n">ac</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">action_space</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
|
||||
<span class="linenos">163</span> <span class="n">obs</span><span class="p">,</span> <span class="n">reward</span><span class="p">,</span> <span class="n">terminated</span><span class="p">,</span> <span class="n">truncated</span><span class="p">,</span> <span class="n">info</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">step</span><span class="p">(</span><span class="n">ac</span><span class="p">)</span>
|
||||
<span class="linenos">164</span> <span class="n">rewards</span> <span class="o">+=</span> <span class="n">reward</span>
|
||||
<span class="linenos">165</span>
|
||||
<span class="linenos">166</span> <span class="k">if</span> <span class="n">terminated</span> <span class="ow">or</span> <span class="n">truncated</span><span class="p">:</span>
|
||||
<span class="linenos">167</span> <span class="nb">print</span><span class="p">(</span><span class="n">rewards</span><span class="p">)</span>
|
||||
<span class="linenos">168</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos">169</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">170</span>
|
||||
<span class="linenos">171</span> <span class="k">try</span><span class="p">:</span> <span class="c1"># Some mujoco-based envs don't correlcty implement .close</span>
|
||||
<span class="linenos">172</span> <span class="n">env</span><span class="o">.</span><span class="n">close</span><span class="p">()</span>
|
||||
<span class="linenos">173</span> <span class="k">except</span><span class="p">:</span>
|
||||
<span class="linenos">174</span> <span class="k">pass</span>
|
||||
<span class="linenos">175</span>
|
||||
<span class="linenos">176</span>
|
||||
<span class="linenos">177</span><span class="k">def</span> <span class="nf">example_fully_custom_mp_alternative</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="kc">True</span><span class="p">):</span>
|
||||
<span class="linenos">178</span><span class="w"> </span><span class="sd">"""</span>
|
||||
<span class="linenos">179</span><span class="sd"> Instead of defining the mp_args in a new custom MP_Wrapper, they can also be provided during registration.</span>
|
||||
<span class="linenos">180</span><span class="sd"> Args:</span>
|
||||
<span class="linenos">181</span><span class="sd"> seed: seed</span>
|
||||
<span class="linenos">182</span><span class="sd"> iterations: Number of rollout steps to run</span>
|
||||
<span class="linenos">183</span><span class="sd"> render: Render the episode</span>
|
||||
<span class="linenos">184</span>
|
||||
<span class="linenos">185</span><span class="sd"> """</span>
|
||||
<span class="linenos">185</span><span class="sd"> Returns:</span>
|
||||
<span class="linenos">186</span>
|
||||
<span class="linenos">187</span> <span class="n">base_env_id</span> <span class="o">=</span> <span class="s2">"fancy/Reacher5d-v0"</span>
|
||||
<span class="linenos">188</span> <span class="n">custom_env_id</span> <span class="o">=</span> <span class="s2">"fancy/Reacher5d-Custom-v0"</span>
|
||||
<span class="linenos">189</span> <span class="n">custom_env_id_ProMP</span> <span class="o">=</span> <span class="s2">"fancy_ProMP/Reacher5d-Custom-v0"</span>
|
||||
<span class="linenos">190</span>
|
||||
<span class="linenos">191</span> <span class="n">fancy_gym</span><span class="o">.</span><span class="n">upgrade</span><span class="p">(</span><span class="n">custom_env_id</span><span class="p">,</span> <span class="n">mp_wrapper</span><span class="o">=</span><span class="n">fancy_gym</span><span class="o">.</span><span class="n">envs</span><span class="o">.</span><span class="n">mujoco</span><span class="o">.</span><span class="n">reacher</span><span class="o">.</span><span class="n">MPWrapper</span><span class="p">,</span> <span class="n">add_mp_types</span><span class="o">=</span><span class="p">[</span><span class="s1">'ProMP'</span><span class="p">],</span> <span class="n">base_id</span><span class="o">=</span><span class="n">base_env_id</span><span class="p">,</span> <span class="n">mp_config_override</span><span class="o">=</span> <span class="p">{</span><span class="s1">'ProMP'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos">192</span> <span class="s1">'trajectory_generator_kwargs'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos">193</span> <span class="s1">'trajectory_generator_type'</span><span class="p">:</span> <span class="s1">'promp'</span><span class="p">,</span>
|
||||
<span class="linenos">194</span> <span class="s1">'weights_scale'</span><span class="p">:</span> <span class="mi">2</span>
|
||||
<span class="linenos">195</span> <span class="p">},</span>
|
||||
<span class="linenos">196</span> <span class="s1">'phase_generator_kwargs'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos">197</span> <span class="s1">'phase_generator_type'</span><span class="p">:</span> <span class="s1">'linear'</span>
|
||||
<span class="linenos">198</span> <span class="p">},</span>
|
||||
<span class="linenos">199</span> <span class="s1">'controller_kwargs'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos">200</span> <span class="s1">'controller_type'</span><span class="p">:</span> <span class="s1">'velocity'</span>
|
||||
<span class="linenos">201</span> <span class="p">},</span>
|
||||
<span class="linenos">202</span> <span class="s1">'basis_generator_kwargs'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos">203</span> <span class="s1">'basis_generator_type'</span><span class="p">:</span> <span class="s1">'zero_rbf'</span><span class="p">,</span>
|
||||
<span class="linenos">204</span> <span class="s1">'num_basis'</span><span class="p">:</span> <span class="mi">5</span><span class="p">,</span>
|
||||
<span class="linenos">205</span> <span class="s1">'num_basis_zero_start'</span><span class="p">:</span> <span class="mi">1</span>
|
||||
<span class="linenos">206</span> <span class="p">}</span>
|
||||
<span class="linenos">207</span> <span class="p">}})</span>
|
||||
<span class="linenos">208</span>
|
||||
<span class="linenos">209</span> <span class="n">env</span> <span class="o">=</span> <span class="n">gym</span><span class="o">.</span><span class="n">make</span><span class="p">(</span><span class="n">custom_env_id_ProMP</span><span class="p">,</span> <span class="n">render_mode</span><span class="o">=</span><span class="s1">'human'</span> <span class="k">if</span> <span class="n">render</span> <span class="k">else</span> <span class="kc">None</span><span class="p">)</span>
|
||||
<span class="linenos">187</span><span class="sd"> """</span>
|
||||
<span class="linenos">188</span>
|
||||
<span class="linenos">189</span> <span class="n">base_env_id</span> <span class="o">=</span> <span class="s2">"fancy/Reacher5d-v0"</span>
|
||||
<span class="linenos">190</span> <span class="n">custom_env_id</span> <span class="o">=</span> <span class="s2">"fancy/Reacher5d-Custom-v0"</span>
|
||||
<span class="linenos">191</span> <span class="n">custom_env_id_ProMP</span> <span class="o">=</span> <span class="s2">"fancy_ProMP/Reacher5d-Custom-v0"</span>
|
||||
<span class="linenos">192</span>
|
||||
<span class="linenos">193</span> <span class="n">fancy_gym</span><span class="o">.</span><span class="n">upgrade</span><span class="p">(</span><span class="n">custom_env_id</span><span class="p">,</span> <span class="n">mp_wrapper</span><span class="o">=</span><span class="n">fancy_gym</span><span class="o">.</span><span class="n">envs</span><span class="o">.</span><span class="n">mujoco</span><span class="o">.</span><span class="n">reacher</span><span class="o">.</span><span class="n">MPWrapper</span><span class="p">,</span> <span class="n">add_mp_types</span><span class="o">=</span><span class="p">[</span><span class="s1">'ProMP'</span><span class="p">],</span> <span class="n">base_id</span><span class="o">=</span><span class="n">base_env_id</span><span class="p">,</span> <span class="n">mp_config_override</span><span class="o">=</span> <span class="p">{</span><span class="s1">'ProMP'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos">194</span> <span class="s1">'trajectory_generator_kwargs'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos">195</span> <span class="s1">'trajectory_generator_type'</span><span class="p">:</span> <span class="s1">'promp'</span><span class="p">,</span>
|
||||
<span class="linenos">196</span> <span class="s1">'weights_scale'</span><span class="p">:</span> <span class="mi">2</span>
|
||||
<span class="linenos">197</span> <span class="p">},</span>
|
||||
<span class="linenos">198</span> <span class="s1">'phase_generator_kwargs'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos">199</span> <span class="s1">'phase_generator_type'</span><span class="p">:</span> <span class="s1">'linear'</span>
|
||||
<span class="linenos">200</span> <span class="p">},</span>
|
||||
<span class="linenos">201</span> <span class="s1">'controller_kwargs'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos">202</span> <span class="s1">'controller_type'</span><span class="p">:</span> <span class="s1">'velocity'</span>
|
||||
<span class="linenos">203</span> <span class="p">},</span>
|
||||
<span class="linenos">204</span> <span class="s1">'basis_generator_kwargs'</span><span class="p">:</span> <span class="p">{</span>
|
||||
<span class="linenos">205</span> <span class="s1">'basis_generator_type'</span><span class="p">:</span> <span class="s1">'zero_rbf'</span><span class="p">,</span>
|
||||
<span class="linenos">206</span> <span class="s1">'num_basis'</span><span class="p">:</span> <span class="mi">5</span><span class="p">,</span>
|
||||
<span class="linenos">207</span> <span class="s1">'num_basis_zero_start'</span><span class="p">:</span> <span class="mi">1</span>
|
||||
<span class="linenos">208</span> <span class="p">}</span>
|
||||
<span class="linenos">209</span> <span class="p">}})</span>
|
||||
<span class="linenos">210</span>
|
||||
<span class="linenos">211</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos">212</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">213</span>
|
||||
<span class="linenos">214</span> <span class="k">if</span> <span class="n">render</span><span class="p">:</span>
|
||||
<span class="linenos">215</span> <span class="n">env</span><span class="o">.</span><span class="n">render</span><span class="p">()</span>
|
||||
<span class="linenos">216</span>
|
||||
<span class="linenos">217</span> <span class="c1"># number of samples/full trajectories (multiple environment steps)</span>
|
||||
<span class="linenos">218</span> <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">iterations</span><span class="p">):</span>
|
||||
<span class="linenos">219</span> <span class="n">ac</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">action_space</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
|
||||
<span class="linenos">220</span> <span class="n">obs</span><span class="p">,</span> <span class="n">reward</span><span class="p">,</span> <span class="n">terminated</span><span class="p">,</span> <span class="n">truncated</span><span class="p">,</span> <span class="n">info</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">step</span><span class="p">(</span><span class="n">ac</span><span class="p">)</span>
|
||||
<span class="linenos">221</span> <span class="n">rewards</span> <span class="o">+=</span> <span class="n">reward</span>
|
||||
<span class="linenos">222</span>
|
||||
<span class="linenos">223</span> <span class="k">if</span> <span class="n">terminated</span> <span class="ow">or</span> <span class="n">truncated</span><span class="p">:</span>
|
||||
<span class="linenos">224</span> <span class="nb">print</span><span class="p">(</span><span class="n">rewards</span><span class="p">)</span>
|
||||
<span class="linenos">225</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos">226</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">227</span>
|
||||
<span class="linenos">228</span> <span class="k">if</span> <span class="n">render</span><span class="p">:</span>
|
||||
<span class="linenos">229</span> <span class="n">env</span><span class="o">.</span><span class="n">render</span><span class="p">()</span>
|
||||
<span class="linenos">230</span>
|
||||
<span class="linenos">231</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos">232</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">233</span>
|
||||
<span class="linenos">234</span> <span class="c1"># number of samples/full trajectories (multiple environment steps)</span>
|
||||
<span class="linenos">235</span> <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">iterations</span><span class="p">):</span>
|
||||
<span class="linenos">236</span> <span class="n">ac</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">action_space</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
|
||||
<span class="linenos">237</span> <span class="n">obs</span><span class="p">,</span> <span class="n">reward</span><span class="p">,</span> <span class="n">terminated</span><span class="p">,</span> <span class="n">truncated</span><span class="p">,</span> <span class="n">info</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">step</span><span class="p">(</span><span class="n">ac</span><span class="p">)</span>
|
||||
<span class="linenos">238</span> <span class="n">rewards</span> <span class="o">+=</span> <span class="n">reward</span>
|
||||
<span class="linenos">239</span>
|
||||
<span class="linenos">240</span> <span class="k">if</span> <span class="n">terminated</span> <span class="ow">or</span> <span class="n">truncated</span><span class="p">:</span>
|
||||
<span class="linenos">241</span> <span class="nb">print</span><span class="p">(</span><span class="n">rewards</span><span class="p">)</span>
|
||||
<span class="linenos">242</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos">243</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">244</span>
|
||||
<span class="linenos">245</span> <span class="k">try</span><span class="p">:</span> <span class="c1"># Some mujoco-based envs don't correlcty implement .close</span>
|
||||
<span class="linenos">246</span> <span class="n">env</span><span class="o">.</span><span class="n">close</span><span class="p">()</span>
|
||||
<span class="linenos">247</span> <span class="k">except</span><span class="p">:</span>
|
||||
<span class="linenos">248</span> <span class="k">pass</span>
|
||||
<span class="linenos">249</span>
|
||||
<span class="linenos">250</span>
|
||||
<span class="linenos">251</span><span class="k">def</span> <span class="nf">main</span><span class="p">():</span>
|
||||
<span class="linenos">252</span> <span class="n">render</span> <span class="o">=</span> <span class="kc">False</span>
|
||||
<span class="linenos">253</span> <span class="c1"># DMP</span>
|
||||
<span class="linenos">254</span> <span class="n">example_mp</span><span class="p">(</span><span class="s2">"fancy_DMP/HoleReacher-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">5</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">255</span>
|
||||
<span class="linenos">256</span> <span class="c1"># ProMP</span>
|
||||
<span class="linenos">257</span> <span class="n">example_mp</span><span class="p">(</span><span class="s2">"fancy_ProMP/HoleReacher-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">5</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">258</span> <span class="n">example_mp</span><span class="p">(</span><span class="s2">"fancy_ProMP/BoxPushingTemporalSparse-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">259</span> <span class="n">example_mp</span><span class="p">(</span><span class="s2">"fancy_ProMP/TableTennis4D-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">20</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">260</span>
|
||||
<span class="linenos">261</span> <span class="c1"># ProDMP with Replanning</span>
|
||||
<span class="linenos">262</span> <span class="n">example_mp</span><span class="p">(</span><span class="s2">"fancy_ProDMP/BoxPushingDenseReplan-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">4</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">263</span> <span class="n">example_mp</span><span class="p">(</span><span class="s2">"fancy_ProDMP/TableTennis4DReplan-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">20</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">264</span> <span class="n">example_mp</span><span class="p">(</span><span class="s2">"fancy_ProDMP/TableTennisWindReplan-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">20</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">265</span>
|
||||
<span class="linenos">266</span> <span class="c1"># Altered basis functions</span>
|
||||
<span class="linenos">267</span> <span class="n">obs1</span> <span class="o">=</span> <span class="n">example_custom_mp</span><span class="p">(</span><span class="s2">"fancy_ProMP/Reacher5d-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">268</span>
|
||||
<span class="linenos">269</span> <span class="c1"># Custom MP</span>
|
||||
<span class="linenos">270</span> <span class="n">example_fully_custom_mp</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">271</span> <span class="n">example_fully_custom_mp_alternative</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">272</span>
|
||||
<span class="linenos">273</span><span class="k">if</span> <span class="vm">__name__</span><span class="o">==</span><span class="s1">'__main__'</span><span class="p">:</span>
|
||||
<span class="linenos">274</span> <span class="n">main</span><span class="p">()</span>
|
||||
<span class="linenos">211</span> <span class="n">env</span> <span class="o">=</span> <span class="n">gym</span><span class="o">.</span><span class="n">make</span><span class="p">(</span><span class="n">custom_env_id_ProMP</span><span class="p">,</span> <span class="n">render_mode</span><span class="o">=</span><span class="s1">'human'</span> <span class="k">if</span> <span class="n">render</span> <span class="k">else</span> <span class="kc">None</span><span class="p">)</span>
|
||||
<span class="linenos">212</span>
|
||||
<span class="linenos">213</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos">214</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">215</span>
|
||||
<span class="linenos">216</span> <span class="k">if</span> <span class="n">render</span><span class="p">:</span>
|
||||
<span class="linenos">217</span> <span class="n">env</span><span class="o">.</span><span class="n">render</span><span class="p">()</span>
|
||||
<span class="linenos">218</span>
|
||||
<span class="linenos">219</span> <span class="c1"># number of samples/full trajectories (multiple environment steps)</span>
|
||||
<span class="linenos">220</span> <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">iterations</span><span class="p">):</span>
|
||||
<span class="linenos">221</span> <span class="n">ac</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">action_space</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
|
||||
<span class="linenos">222</span> <span class="n">obs</span><span class="p">,</span> <span class="n">reward</span><span class="p">,</span> <span class="n">terminated</span><span class="p">,</span> <span class="n">truncated</span><span class="p">,</span> <span class="n">info</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">step</span><span class="p">(</span><span class="n">ac</span><span class="p">)</span>
|
||||
<span class="linenos">223</span> <span class="n">rewards</span> <span class="o">+=</span> <span class="n">reward</span>
|
||||
<span class="linenos">224</span>
|
||||
<span class="linenos">225</span> <span class="k">if</span> <span class="n">terminated</span> <span class="ow">or</span> <span class="n">truncated</span><span class="p">:</span>
|
||||
<span class="linenos">226</span> <span class="nb">print</span><span class="p">(</span><span class="n">rewards</span><span class="p">)</span>
|
||||
<span class="linenos">227</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos">228</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">229</span>
|
||||
<span class="linenos">230</span> <span class="k">if</span> <span class="n">render</span><span class="p">:</span>
|
||||
<span class="linenos">231</span> <span class="n">env</span><span class="o">.</span><span class="n">render</span><span class="p">()</span>
|
||||
<span class="linenos">232</span>
|
||||
<span class="linenos">233</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos">234</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">235</span>
|
||||
<span class="linenos">236</span> <span class="c1"># number of samples/full trajectories (multiple environment steps)</span>
|
||||
<span class="linenos">237</span> <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">iterations</span><span class="p">):</span>
|
||||
<span class="linenos">238</span> <span class="n">ac</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">action_space</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
|
||||
<span class="linenos">239</span> <span class="n">obs</span><span class="p">,</span> <span class="n">reward</span><span class="p">,</span> <span class="n">terminated</span><span class="p">,</span> <span class="n">truncated</span><span class="p">,</span> <span class="n">info</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">step</span><span class="p">(</span><span class="n">ac</span><span class="p">)</span>
|
||||
<span class="linenos">240</span> <span class="n">rewards</span> <span class="o">+=</span> <span class="n">reward</span>
|
||||
<span class="linenos">241</span>
|
||||
<span class="linenos">242</span> <span class="k">if</span> <span class="n">terminated</span> <span class="ow">or</span> <span class="n">truncated</span><span class="p">:</span>
|
||||
<span class="linenos">243</span> <span class="nb">print</span><span class="p">(</span><span class="n">rewards</span><span class="p">)</span>
|
||||
<span class="linenos">244</span> <span class="n">rewards</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos">245</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">246</span>
|
||||
<span class="linenos">247</span> <span class="k">try</span><span class="p">:</span> <span class="c1"># Some mujoco-based envs don't correlcty implement .close</span>
|
||||
<span class="linenos">248</span> <span class="n">env</span><span class="o">.</span><span class="n">close</span><span class="p">()</span>
|
||||
<span class="linenos">249</span> <span class="k">except</span><span class="p">:</span>
|
||||
<span class="linenos">250</span> <span class="k">pass</span>
|
||||
<span class="linenos">251</span>
|
||||
<span class="linenos">252</span>
|
||||
<span class="linenos">253</span><span class="k">def</span> <span class="nf">main</span><span class="p">(</span><span class="n">render</span><span class="o">=</span><span class="kc">False</span><span class="p">):</span>
|
||||
<span class="linenos">254</span> <span class="c1"># DMP</span>
|
||||
<span class="linenos">255</span> <span class="n">example_mp</span><span class="p">(</span><span class="s2">"fancy_DMP/HoleReacher-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">5</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">256</span>
|
||||
<span class="linenos">257</span> <span class="c1"># ProMP</span>
|
||||
<span class="linenos">258</span> <span class="n">example_mp</span><span class="p">(</span><span class="s2">"fancy_ProMP/HoleReacher-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">5</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">259</span> <span class="n">example_mp</span><span class="p">(</span><span class="s2">"fancy_ProMP/BoxPushingTemporalSparse-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">260</span> <span class="n">example_mp</span><span class="p">(</span><span class="s2">"fancy_ProMP/TableTennis4D-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">20</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">261</span>
|
||||
<span class="linenos">262</span> <span class="c1"># ProDMP with Replanning</span>
|
||||
<span class="linenos">263</span> <span class="n">example_mp</span><span class="p">(</span><span class="s2">"fancy_ProDMP/BoxPushingDenseReplan-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">4</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">264</span> <span class="n">example_mp</span><span class="p">(</span><span class="s2">"fancy_ProDMP/TableTennis4DReplan-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">20</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">265</span> <span class="n">example_mp</span><span class="p">(</span><span class="s2">"fancy_ProDMP/TableTennisWindReplan-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">20</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">266</span>
|
||||
<span class="linenos">267</span> <span class="c1"># Altered basis functions</span>
|
||||
<span class="linenos">268</span> <span class="n">obs1</span> <span class="o">=</span> <span class="n">example_custom_mp</span><span class="p">(</span><span class="s2">"fancy_ProMP/Reacher5d-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">269</span>
|
||||
<span class="linenos">270</span> <span class="c1"># Custom MP</span>
|
||||
<span class="linenos">271</span> <span class="n">example_fully_custom_mp</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">272</span> <span class="n">example_fully_custom_mp_alternative</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="mi">10</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">273</span>
|
||||
<span class="linenos">274</span><span class="k">if</span> <span class="vm">__name__</span><span class="o">==</span><span class="s1">'__main__'</span><span class="p">:</span>
|
||||
<span class="linenos">275</span> <span class="n">main</span><span class="p">()</span>
|
||||
</pre></div>
|
||||
</div>
|
||||
</section>
|
||||
|
||||
@ -4,7 +4,7 @@
|
||||
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>MP Params Tuning Example — Fancy Gym 0.2 documentation</title>
|
||||
<title>MP Params Tuning Example — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/style.css" type="text/css" />
|
||||
@ -41,7 +41,7 @@
|
||||
<img src="../_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
|
||||
|
||||
30
docs/build/html/examples/open_ai.html
vendored
30
docs/build/html/examples/open_ai.html
vendored
@ -4,7 +4,7 @@
|
||||
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>OpenAI Envs Examples — Fancy Gym 0.2 documentation</title>
|
||||
<title>OpenAI Envs Examples — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/style.css" type="text/css" />
|
||||
@ -41,7 +41,7 @@
|
||||
<img src="../_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
|
||||
@ -122,27 +122,27 @@
|
||||
<span class="linenos">13</span><span class="sd"> Returns:</span>
|
||||
<span class="linenos">14</span>
|
||||
<span class="linenos">15</span><span class="sd"> """</span>
|
||||
<span class="linenos">16</span> <span class="n">env</span> <span class="o">=</span> <span class="n">gym</span><span class="o">.</span><span class="n">make</span><span class="p">(</span><span class="n">env_name</span><span class="p">)</span>
|
||||
<span class="linenos">16</span> <span class="n">env</span> <span class="o">=</span> <span class="n">gym</span><span class="o">.</span><span class="n">make</span><span class="p">(</span><span class="n">env_name</span><span class="p">,</span> <span class="n">render_mode</span><span class="o">=</span><span class="s1">'human'</span> <span class="k">if</span> <span class="n">render</span> <span class="k">else</span> <span class="kc">None</span><span class="p">)</span>
|
||||
<span class="linenos">17</span>
|
||||
<span class="linenos">18</span> <span class="n">returns</span> <span class="o">=</span> <span class="mi">0</span>
|
||||
<span class="linenos">19</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="n">seed</span><span class="p">)</span>
|
||||
<span class="linenos">20</span> <span class="c1"># number of samples/full trajectories (multiple environment steps)</span>
|
||||
<span class="linenos">21</span> <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="mi">10</span><span class="p">):</span>
|
||||
<span class="linenos">22</span> <span class="k">if</span> <span class="n">render</span> <span class="ow">and</span> <span class="n">i</span> <span class="o">%</span> <span class="mi">2</span> <span class="o">==</span> <span class="mi">0</span><span class="p">:</span>
|
||||
<span class="linenos">23</span> <span class="n">env</span><span class="o">.</span><span class="n">render</span><span class="p">(</span><span class="n">mode</span><span class="o">=</span><span class="s2">"human"</span><span class="p">)</span>
|
||||
<span class="linenos">24</span> <span class="k">else</span><span class="p">:</span>
|
||||
<span class="linenos">25</span> <span class="n">env</span><span class="o">.</span><span class="n">render</span><span class="p">()</span>
|
||||
<span class="linenos">26</span> <span class="n">ac</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">action_space</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
|
||||
<span class="linenos">27</span> <span class="n">obs</span><span class="p">,</span> <span class="n">reward</span><span class="p">,</span> <span class="n">terminated</span><span class="p">,</span> <span class="n">truncated</span><span class="p">,</span> <span class="n">info</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">step</span><span class="p">(</span><span class="n">ac</span><span class="p">)</span>
|
||||
<span class="linenos">28</span> <span class="n">returns</span> <span class="o">+=</span> <span class="n">reward</span>
|
||||
<span class="linenos">29</span>
|
||||
<span class="linenos">30</span> <span class="k">if</span> <span class="n">terminated</span> <span class="ow">or</span> <span class="n">truncated</span><span class="p">:</span>
|
||||
<span class="linenos">31</span> <span class="nb">print</span><span class="p">(</span><span class="n">returns</span><span class="p">)</span>
|
||||
<span class="linenos">32</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">33</span>
|
||||
<span class="linenos">23</span> <span class="n">env</span><span class="o">.</span><span class="n">render</span><span class="p">()</span>
|
||||
<span class="linenos">24</span> <span class="n">ac</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">action_space</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
|
||||
<span class="linenos">25</span> <span class="n">obs</span><span class="p">,</span> <span class="n">reward</span><span class="p">,</span> <span class="n">terminated</span><span class="p">,</span> <span class="n">truncated</span><span class="p">,</span> <span class="n">info</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">step</span><span class="p">(</span><span class="n">ac</span><span class="p">)</span>
|
||||
<span class="linenos">26</span> <span class="n">returns</span> <span class="o">+=</span> <span class="n">reward</span>
|
||||
<span class="linenos">27</span>
|
||||
<span class="linenos">28</span> <span class="k">if</span> <span class="n">terminated</span> <span class="ow">or</span> <span class="n">truncated</span><span class="p">:</span>
|
||||
<span class="linenos">29</span> <span class="nb">print</span><span class="p">(</span><span class="n">returns</span><span class="p">)</span>
|
||||
<span class="linenos">30</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">31</span>
|
||||
<span class="linenos">32</span><span class="k">def</span> <span class="nf">main</span><span class="p">(</span><span class="n">render</span><span class="o">=</span><span class="kc">True</span><span class="p">):</span>
|
||||
<span class="linenos">33</span> <span class="n">example_mp</span><span class="p">(</span><span class="s2">"gym_ProMP/Reacher-v2"</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">34</span>
|
||||
<span class="linenos">35</span><span class="k">if</span> <span class="vm">__name__</span> <span class="o">==</span> <span class="s1">'__main__'</span><span class="p">:</span>
|
||||
<span class="linenos">36</span> <span class="n">example_mp</span><span class="p">(</span><span class="s2">"gym_ProMP/Reacher-v2"</span><span class="p">)</span>
|
||||
<span class="linenos">36</span> <span class="n">main</span><span class="p">()</span>
|
||||
</pre></div>
|
||||
</div>
|
||||
</section>
|
||||
|
||||
@ -4,7 +4,7 @@
|
||||
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>PD Control Gain Tuning Example — Fancy Gym 0.2 documentation</title>
|
||||
<title>PD Control Gain Tuning Example — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/style.css" type="text/css" />
|
||||
@ -41,7 +41,7 @@
|
||||
<img src="../_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
|
||||
|
||||
69
docs/build/html/examples/replanning_envs.html
vendored
69
docs/build/html/examples/replanning_envs.html
vendored
@ -4,7 +4,7 @@
|
||||
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>Replanning Example — Fancy Gym 0.2 documentation</title>
|
||||
<title>Replanning Example — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/style.css" type="text/css" />
|
||||
@ -41,7 +41,7 @@
|
||||
<img src="../_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
|
||||
@ -112,24 +112,24 @@
|
||||
<span class="linenos"> 3</span>
|
||||
<span class="linenos"> 4</span>
|
||||
<span class="linenos"> 5</span><span class="k">def</span> <span class="nf">example_run_replanning_env</span><span class="p">(</span><span class="n">env_name</span><span class="o">=</span><span class="s2">"fancy_ProDMP/BoxPushingDenseReplan-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="kc">False</span><span class="p">):</span>
|
||||
<span class="linenos"> 6</span> <span class="n">env</span> <span class="o">=</span> <span class="n">gym</span><span class="o">.</span><span class="n">make</span><span class="p">(</span><span class="n">env_name</span><span class="p">)</span>
|
||||
<span class="linenos"> 6</span> <span class="n">env</span> <span class="o">=</span> <span class="n">gym</span><span class="o">.</span><span class="n">make</span><span class="p">(</span><span class="n">env_name</span><span class="p">,</span> <span class="n">render_mode</span><span class="o">=</span><span class="s1">'human'</span> <span class="k">if</span> <span class="n">render</span> <span class="k">else</span> <span class="kc">None</span><span class="p">)</span>
|
||||
<span class="linenos"> 7</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="n">seed</span><span class="p">)</span>
|
||||
<span class="linenos"> 8</span> <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">iterations</span><span class="p">):</span>
|
||||
<span class="linenos"> 9</span> <span class="n">done</span> <span class="o">=</span> <span class="kc">False</span>
|
||||
<span class="linenos">10</span> <span class="k">while</span> <span class="n">done</span> <span class="ow">is</span> <span class="kc">False</span><span class="p">:</span>
|
||||
<span class="linenos">11</span> <span class="n">ac</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">action_space</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
|
||||
<span class="linenos">12</span> <span class="n">obs</span><span class="p">,</span> <span class="n">reward</span><span class="p">,</span> <span class="n">terminated</span><span class="p">,</span> <span class="n">truncated</span><span class="p">,</span> <span class="n">info</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">step</span><span class="p">(</span><span class="n">ac</span><span class="p">)</span>
|
||||
<span class="linenos">13</span> <span class="k">if</span> <span class="n">render</span><span class="p">:</span>
|
||||
<span class="linenos">14</span> <span class="n">env</span><span class="o">.</span><span class="n">render</span><span class="p">(</span><span class="n">mode</span><span class="o">=</span><span class="s2">"human"</span><span class="p">)</span>
|
||||
<span class="linenos">15</span> <span class="k">if</span> <span class="n">terminated</span> <span class="ow">or</span> <span class="n">truncated</span><span class="p">:</span>
|
||||
<span class="linenos">16</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos"> 9</span> <span class="k">while</span> <span class="kc">True</span><span class="p">:</span>
|
||||
<span class="linenos">10</span> <span class="n">ac</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">action_space</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
|
||||
<span class="linenos">11</span> <span class="n">obs</span><span class="p">,</span> <span class="n">reward</span><span class="p">,</span> <span class="n">terminated</span><span class="p">,</span> <span class="n">truncated</span><span class="p">,</span> <span class="n">info</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">step</span><span class="p">(</span><span class="n">ac</span><span class="p">)</span>
|
||||
<span class="linenos">12</span> <span class="k">if</span> <span class="n">render</span><span class="p">:</span>
|
||||
<span class="linenos">13</span> <span class="n">env</span><span class="o">.</span><span class="n">render</span><span class="p">()</span>
|
||||
<span class="linenos">14</span> <span class="k">if</span> <span class="n">terminated</span> <span class="ow">or</span> <span class="n">truncated</span><span class="p">:</span>
|
||||
<span class="linenos">15</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">16</span> <span class="k">break</span>
|
||||
<span class="linenos">17</span> <span class="n">env</span><span class="o">.</span><span class="n">close</span><span class="p">()</span>
|
||||
<span class="linenos">18</span> <span class="k">del</span> <span class="n">env</span>
|
||||
<span class="linenos">19</span>
|
||||
<span class="linenos">20</span>
|
||||
<span class="linenos">21</span><span class="k">def</span> <span class="nf">example_custom_replanning_envs</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span> <span class="n">iteration</span><span class="o">=</span><span class="mi">100</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="kc">True</span><span class="p">):</span>
|
||||
<span class="linenos">22</span> <span class="c1"># id for a step-based environment</span>
|
||||
<span class="linenos">23</span> <span class="n">base_env_id</span> <span class="o">=</span> <span class="s2">"BoxPushingDense-v0"</span>
|
||||
<span class="linenos">23</span> <span class="n">base_env_id</span> <span class="o">=</span> <span class="s2">"fancy/BoxPushingDense-v0"</span>
|
||||
<span class="linenos">24</span>
|
||||
<span class="linenos">25</span> <span class="n">wrappers</span> <span class="o">=</span> <span class="p">[</span><span class="n">fancy_gym</span><span class="o">.</span><span class="n">envs</span><span class="o">.</span><span class="n">mujoco</span><span class="o">.</span><span class="n">box_pushing</span><span class="o">.</span><span class="n">mp_wrapper</span><span class="o">.</span><span class="n">MPWrapper</span><span class="p">]</span>
|
||||
<span class="linenos">26</span>
|
||||
@ -147,31 +147,34 @@
|
||||
<span class="linenos">38</span> <span class="s1">'replanning_schedule'</span><span class="p">:</span> <span class="k">lambda</span> <span class="n">pos</span><span class="p">,</span> <span class="n">vel</span><span class="p">,</span> <span class="n">obs</span><span class="p">,</span> <span class="n">action</span><span class="p">,</span> <span class="n">t</span><span class="p">:</span> <span class="n">t</span> <span class="o">%</span> <span class="mi">25</span> <span class="o">==</span> <span class="mi">0</span><span class="p">,</span>
|
||||
<span class="linenos">39</span> <span class="s1">'condition_on_desired'</span><span class="p">:</span> <span class="kc">True</span><span class="p">}</span>
|
||||
<span class="linenos">40</span>
|
||||
<span class="linenos">41</span> <span class="n">env</span> <span class="o">=</span> <span class="n">fancy_gym</span><span class="o">.</span><span class="n">make_bb</span><span class="p">(</span><span class="n">env_id</span><span class="o">=</span><span class="n">base_env_id</span><span class="p">,</span> <span class="n">wrappers</span><span class="o">=</span><span class="n">wrappers</span><span class="p">,</span> <span class="n">black_box_kwargs</span><span class="o">=</span><span class="n">black_box_kwargs</span><span class="p">,</span>
|
||||
<span class="linenos">42</span> <span class="n">traj_gen_kwargs</span><span class="o">=</span><span class="n">trajectory_generator_kwargs</span><span class="p">,</span> <span class="n">controller_kwargs</span><span class="o">=</span><span class="n">controller_kwargs</span><span class="p">,</span>
|
||||
<span class="linenos">43</span> <span class="n">phase_kwargs</span><span class="o">=</span><span class="n">phase_generator_kwargs</span><span class="p">,</span> <span class="n">basis_kwargs</span><span class="o">=</span><span class="n">basis_generator_kwargs</span><span class="p">,</span>
|
||||
<span class="linenos">44</span> <span class="n">seed</span><span class="o">=</span><span class="n">seed</span><span class="p">)</span>
|
||||
<span class="linenos">45</span> <span class="k">if</span> <span class="n">render</span><span class="p">:</span>
|
||||
<span class="linenos">46</span> <span class="n">env</span><span class="o">.</span><span class="n">render</span><span class="p">(</span><span class="n">mode</span><span class="o">=</span><span class="s2">"human"</span><span class="p">)</span>
|
||||
<span class="linenos">47</span>
|
||||
<span class="linenos">48</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">49</span>
|
||||
<span class="linenos">50</span> <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">iteration</span><span class="p">):</span>
|
||||
<span class="linenos">51</span> <span class="n">ac</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">action_space</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
|
||||
<span class="linenos">52</span> <span class="n">obs</span><span class="p">,</span> <span class="n">reward</span><span class="p">,</span> <span class="n">terminated</span><span class="p">,</span> <span class="n">truncated</span><span class="p">,</span> <span class="n">info</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">step</span><span class="p">(</span><span class="n">ac</span><span class="p">)</span>
|
||||
<span class="linenos">53</span> <span class="k">if</span> <span class="n">terminated</span> <span class="ow">or</span> <span class="n">truncated</span><span class="p">:</span>
|
||||
<span class="linenos">54</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">55</span>
|
||||
<span class="linenos">56</span> <span class="n">env</span><span class="o">.</span><span class="n">close</span><span class="p">()</span>
|
||||
<span class="linenos">57</span> <span class="k">del</span> <span class="n">env</span>
|
||||
<span class="linenos">58</span>
|
||||
<span class="linenos">41</span> <span class="n">base_env</span> <span class="o">=</span> <span class="n">gym</span><span class="o">.</span><span class="n">make</span><span class="p">(</span><span class="n">base_env_id</span><span class="p">,</span> <span class="n">render_mode</span><span class="o">=</span><span class="s1">'human'</span> <span class="k">if</span> <span class="n">render</span> <span class="k">else</span> <span class="kc">None</span><span class="p">)</span>
|
||||
<span class="linenos">42</span> <span class="n">env</span> <span class="o">=</span> <span class="n">fancy_gym</span><span class="o">.</span><span class="n">make_bb</span><span class="p">(</span><span class="n">env</span><span class="o">=</span><span class="n">base_env</span><span class="p">,</span> <span class="n">wrappers</span><span class="o">=</span><span class="n">wrappers</span><span class="p">,</span> <span class="n">black_box_kwargs</span><span class="o">=</span><span class="n">black_box_kwargs</span><span class="p">,</span>
|
||||
<span class="linenos">43</span> <span class="n">traj_gen_kwargs</span><span class="o">=</span><span class="n">trajectory_generator_kwargs</span><span class="p">,</span> <span class="n">controller_kwargs</span><span class="o">=</span><span class="n">controller_kwargs</span><span class="p">,</span>
|
||||
<span class="linenos">44</span> <span class="n">phase_kwargs</span><span class="o">=</span><span class="n">phase_generator_kwargs</span><span class="p">,</span> <span class="n">basis_kwargs</span><span class="o">=</span><span class="n">basis_generator_kwargs</span><span class="p">,</span>
|
||||
<span class="linenos">45</span> <span class="n">seed</span><span class="o">=</span><span class="n">seed</span><span class="p">)</span>
|
||||
<span class="linenos">46</span> <span class="k">if</span> <span class="n">render</span><span class="p">:</span>
|
||||
<span class="linenos">47</span> <span class="n">env</span><span class="o">.</span><span class="n">render</span><span class="p">()</span>
|
||||
<span class="linenos">48</span>
|
||||
<span class="linenos">49</span> <span class="n">obs</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">50</span>
|
||||
<span class="linenos">51</span> <span class="k">for</span> <span class="n">i</span> <span class="ow">in</span> <span class="nb">range</span><span class="p">(</span><span class="n">iteration</span><span class="p">):</span>
|
||||
<span class="linenos">52</span> <span class="n">ac</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">action_space</span><span class="o">.</span><span class="n">sample</span><span class="p">()</span>
|
||||
<span class="linenos">53</span> <span class="n">obs</span><span class="p">,</span> <span class="n">reward</span><span class="p">,</span> <span class="n">terminated</span><span class="p">,</span> <span class="n">truncated</span><span class="p">,</span> <span class="n">info</span> <span class="o">=</span> <span class="n">env</span><span class="o">.</span><span class="n">step</span><span class="p">(</span><span class="n">ac</span><span class="p">)</span>
|
||||
<span class="linenos">54</span> <span class="k">if</span> <span class="n">terminated</span> <span class="ow">or</span> <span class="n">truncated</span><span class="p">:</span>
|
||||
<span class="linenos">55</span> <span class="n">env</span><span class="o">.</span><span class="n">reset</span><span class="p">()</span>
|
||||
<span class="linenos">56</span>
|
||||
<span class="linenos">57</span> <span class="n">env</span><span class="o">.</span><span class="n">close</span><span class="p">()</span>
|
||||
<span class="linenos">58</span> <span class="k">del</span> <span class="n">env</span>
|
||||
<span class="linenos">59</span>
|
||||
<span class="linenos">60</span><span class="k">if</span> <span class="vm">__name__</span> <span class="o">==</span> <span class="s2">"__main__"</span><span class="p">:</span>
|
||||
<span class="linenos">60</span><span class="k">def</span> <span class="nf">main</span><span class="p">(</span><span class="n">render</span><span class="o">=</span><span class="kc">False</span><span class="p">):</span>
|
||||
<span class="linenos">61</span> <span class="c1"># run a registered replanning environment</span>
|
||||
<span class="linenos">62</span> <span class="n">example_run_replanning_env</span><span class="p">(</span><span class="n">env_name</span><span class="o">=</span><span class="s2">"fancy_ProDMP/BoxPushingDenseReplan-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="kc">False</span><span class="p">)</span>
|
||||
<span class="linenos">62</span> <span class="n">example_run_replanning_env</span><span class="p">(</span><span class="n">env_name</span><span class="o">=</span><span class="s2">"fancy_ProDMP/BoxPushingDenseReplan-v0"</span><span class="p">,</span> <span class="n">seed</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">iterations</span><span class="o">=</span><span class="mi">1</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">63</span>
|
||||
<span class="linenos">64</span> <span class="c1"># run a custom replanning environment</span>
|
||||
<span class="linenos">65</span> <span class="n">example_custom_replanning_envs</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span> <span class="n">iteration</span><span class="o">=</span><span class="mi">8</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="kc">True</span><span class="p">)</span>
|
||||
<span class="linenos">65</span> <span class="n">example_custom_replanning_envs</span><span class="p">(</span><span class="n">seed</span><span class="o">=</span><span class="mi">0</span><span class="p">,</span> <span class="n">iteration</span><span class="o">=</span><span class="mi">8</span><span class="p">,</span> <span class="n">render</span><span class="o">=</span><span class="n">render</span><span class="p">)</span>
|
||||
<span class="linenos">66</span>
|
||||
<span class="linenos">67</span><span class="k">if</span> <span class="vm">__name__</span> <span class="o">==</span> <span class="s2">"__main__"</span><span class="p">:</span>
|
||||
<span class="linenos">68</span> <span class="n">main</span><span class="p">()</span>
|
||||
</pre></div>
|
||||
</div>
|
||||
</section>
|
||||
|
||||
@ -4,7 +4,7 @@
|
||||
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>fancy_gym.envs — Fancy Gym 0.2 documentation</title>
|
||||
<title>fancy_gym.envs — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/style.css" type="text/css" />
|
||||
@ -39,7 +39,7 @@
|
||||
<img src="../_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
|
||||
|
||||
@ -4,7 +4,7 @@
|
||||
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>fancy_gym.register — Fancy Gym 0.2 documentation</title>
|
||||
<title>fancy_gym.register — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/style.css" type="text/css" />
|
||||
@ -41,7 +41,7 @@
|
||||
<img src="../_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
|
||||
|
||||
@ -4,7 +4,7 @@
|
||||
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>fancy_gym.upgrade — Fancy Gym 0.2 documentation</title>
|
||||
<title>fancy_gym.upgrade — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/style.css" type="text/css" />
|
||||
@ -40,7 +40,7 @@
|
||||
<img src="../_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
|
||||
|
||||
4
docs/build/html/genindex.html
vendored
4
docs/build/html/genindex.html
vendored
@ -3,7 +3,7 @@
|
||||
<head>
|
||||
<meta charset="utf-8" />
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>Index — Fancy Gym 0.2 documentation</title>
|
||||
<title>Index — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="_static/style.css" type="text/css" />
|
||||
@ -38,7 +38,7 @@
|
||||
<img src="_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="search.html" method="get">
|
||||
|
||||
4
docs/build/html/guide/basic_usage.html
vendored
4
docs/build/html/guide/basic_usage.html
vendored
@ -4,7 +4,7 @@
|
||||
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>Basic Usage — Fancy Gym 0.2 documentation</title>
|
||||
<title>Basic Usage — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/style.css" type="text/css" />
|
||||
@ -41,7 +41,7 @@
|
||||
<img src="../_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
|
||||
|
||||
4
docs/build/html/guide/episodic_rl.html
vendored
4
docs/build/html/guide/episodic_rl.html
vendored
@ -4,7 +4,7 @@
|
||||
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>What is Episodic RL? — Fancy Gym 0.2 documentation</title>
|
||||
<title>What is Episodic RL? — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/style.css" type="text/css" />
|
||||
@ -41,7 +41,7 @@
|
||||
<img src="../_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
|
||||
|
||||
8
docs/build/html/guide/installation.html
vendored
8
docs/build/html/guide/installation.html
vendored
@ -4,7 +4,7 @@
|
||||
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>Installation — Fancy Gym 0.2 documentation</title>
|
||||
<title>Installation — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/style.css" type="text/css" />
|
||||
@ -41,7 +41,7 @@
|
||||
<img src="../_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
|
||||
@ -135,7 +135,7 @@ pip<span class="w"> </span>install<span class="w"> </span><span class="s1">'
|
||||
</div>
|
||||
<p>Pip can not automatically install up-to-date versions of metaworld,
|
||||
since they are not avaible on PyPI yet. Install metaworld via</p>
|
||||
<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>pip<span class="w"> </span>install<span class="w"> </span>metaworld@git+https://github.com/Farama-Foundation/Metaworld.git@d155d0051630bb365ea6a824e02c66c068947439#egg<span class="o">=</span>metaworld
|
||||
<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>pip<span class="w"> </span>install<span class="w"> </span>metaworld@git+https://github.com/Farama-Foundation/Metaworld.git@c822f28f582ba1ad49eb5dcf61016566f28003ba#egg<span class="o">=</span>metaworld
|
||||
</pre></div>
|
||||
</div>
|
||||
</section>
|
||||
@ -169,7 +169,7 @@ pip<span class="w"> </span>install<span class="w"> </span>-e<span class="w"> </s
|
||||
</pre></div>
|
||||
</div>
|
||||
<p>Metaworld has to be installed manually with</p>
|
||||
<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>pip<span class="w"> </span>install<span class="w"> </span>metaworld@git+https://github.com/Farama-Foundation/Metaworld.git@d155d0051630bb365ea6a824e02c66c068947439#egg<span class="o">=</span>metaworld
|
||||
<div class="highlight-bash notranslate"><div class="highlight"><pre><span></span>pip<span class="w"> </span>install<span class="w"> </span>metaworld@git+https://github.com/Farama-Foundation/Metaworld.git@c822f28f582ba1ad49eb5dcf61016566f28003ba#egg<span class="o">=</span>metaworld
|
||||
</pre></div>
|
||||
</div>
|
||||
</section>
|
||||
|
||||
4
docs/build/html/guide/upgrading_envs.html
vendored
4
docs/build/html/guide/upgrading_envs.html
vendored
@ -4,7 +4,7 @@
|
||||
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>Creating new MP Environments — Fancy Gym 0.2 documentation</title>
|
||||
<title>Creating new MP Environments — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="../_static/style.css" type="text/css" />
|
||||
@ -41,7 +41,7 @@
|
||||
<img src="../_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="../search.html" method="get">
|
||||
|
||||
4
docs/build/html/index.html
vendored
4
docs/build/html/index.html
vendored
@ -4,7 +4,7 @@
|
||||
<meta charset="utf-8" /><meta name="generator" content="Docutils 0.19: https://docutils.sourceforge.io/" />
|
||||
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>Fancy Gym — Fancy Gym 0.2 documentation</title>
|
||||
<title>Fancy Gym — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="_static/style.css" type="text/css" />
|
||||
@ -40,7 +40,7 @@
|
||||
<img src="_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="search.html" method="get">
|
||||
|
||||
BIN
docs/build/html/objects.inv
vendored
BIN
docs/build/html/objects.inv
vendored
Binary file not shown.
4
docs/build/html/py-modindex.html
vendored
4
docs/build/html/py-modindex.html
vendored
@ -3,7 +3,7 @@
|
||||
<head>
|
||||
<meta charset="utf-8" />
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>Python Module Index — Fancy Gym 0.2 documentation</title>
|
||||
<title>Python Module Index — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="_static/style.css" type="text/css" />
|
||||
@ -41,7 +41,7 @@
|
||||
<img src="_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="search.html" method="get">
|
||||
|
||||
4
docs/build/html/search.html
vendored
4
docs/build/html/search.html
vendored
@ -3,7 +3,7 @@
|
||||
<head>
|
||||
<meta charset="utf-8" />
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>Search — Fancy Gym 0.2 documentation</title>
|
||||
<title>Search — Fancy Gym 0.3.0 documentation</title>
|
||||
<link rel="stylesheet" href="_static/pygments.css" type="text/css" />
|
||||
<link rel="stylesheet" href="_static/css/theme.css" type="text/css" />
|
||||
<link rel="stylesheet" href="_static/style.css" type="text/css" />
|
||||
@ -41,7 +41,7 @@
|
||||
<img src="_static/icon.svg" class="logo" alt="Logo"/>
|
||||
</a>
|
||||
<div class="version">
|
||||
0.2
|
||||
0.3.0
|
||||
</div>
|
||||
<div role="search">
|
||||
<form id="rtd-search-form" class="wy-form" action="#" method="get">
|
||||
|
||||
2
docs/build/html/searchindex.js
vendored
2
docs/build/html/searchindex.js
vendored
File diff suppressed because one or more lines are too long
@ -18,6 +18,12 @@ A composite reward function serves as the performance metric for the RL system.
|
||||
|
||||
Variations of this environment are available, differing in reward structures and the optionality of randomizing the box's initial position. These variations are purposefully designed to challenge RL algorithms, enhancing their generalization and adaptation capabilities. Temporally sparse environments only provide a reward at the last timestep. Spatially sparse environments only provide a reward, if the goal is almost reached, the box is close enought to the goal and somewhat correctly aligned.
|
||||
|
||||
These environments all provide smoothness metrics as part of the return infos:
|
||||
|
||||
- mean_squared_jerk: Averages the square of jerk (rate of acceleration change) across the motion. Lower values indicate smoother movement.
|
||||
- maximum_jerk: Identifies the highest jerk value encountered.
|
||||
- dimensionless_jerk: Normalizes the summed squared jerk over the motion's duration and peak velocity, offering a scale-independent metric of smoothness
|
||||
|
||||
| Name | Description | Horizon | Action Dimension | Observation Dimension |
|
||||
| ------------------------------------------ | -------------------------------------------------------------------- | ------- | ---------------- | --------------------- |
|
||||
| `fancy/BoxPushingDense-v0` | Custom Box-pushing task with dense rewards | 100 | 3 | 13 |
|
||||
@ -49,6 +55,9 @@ Variations of the table tennis environment are available to cater to different r
|
||||
| `fancy/TableTennisWind-v0` | Table Tennis task with wind effects, based on a custom environment for table tennis | 350 | 7 | 19 |
|
||||
| `fancy/TableTennisGoalSwitching-v0` | Table Tennis task with goal switching, based on a custom environment for table tennis | 350 | 7 | 19 |
|
||||
| `fancy/TableTennisWindReplan-v0` | Table Tennis task with wind effects and replanning, based on a custom environment for table tennis | 350 | 7 | 19 |
|
||||
| `fancy/TableTennisRndRobot-v0` | Table Tennis task with random initial robot joint positions \* | 350 | 7 | 19 |
|
||||
|
||||
\* Random initialization of robot joint position and speed can be enabled by providing `random_pos_scale` / `random_vel_scale` to make. `TableTennisRndRobot` is equivalent to `TableTennis4D` except, that `random_pos_scale` is set to 0.1 instead of 0 per default.
|
||||
|
||||
---
|
||||
|
||||
@ -89,8 +98,9 @@ A successful throw in this task is determined by the ball landing in the cup at
|
||||
| `fancy/Reacher5dSparse-v0` | Sparse Reacher task with 5 links, based on Gymnasium's `gym.envs.mujoco.ReacherEnv` | 200 | 5 | 20 |
|
||||
| `fancy/Reacher7d-v0` | Reacher task with 7 links, based on Gymnasium's `gym.envs.mujoco.ReacherEnv` | 200 | 7 | 22 |
|
||||
| `fancy/Reacher7dSparse-v0` | Sparse Reacher task with 7 links, based on Gymnasium's `gym.envs.mujoco.ReacherEnv` | 200 | 7 | 22 |
|
||||
| `fancy/HopperJumpSparse-v0` | Hopper Jump task with sparse rewards, based on Gymnasium's `gym.envs.mujoco.Hopper` | 250 | 3 | 15 / 16\* |
|
||||
| `fancy/HopperJump-v0` | Hopper Jump task with continuous rewards, based on Gymnasium's `gym.envs.mujoco.Hopper` | 250 | 3 | 15 / 16\* |
|
||||
| `fancy/HopperJumpMarkov-v0` | `fancy/HopperJump-v0`, but with an alternative reward that is markovian. | 250 | 3 | 15 / 16\* |
|
||||
| `fancy/HopperJumpSparse-v0` | Hopper Jump task with sparse rewards, based on Gymnasium's `gym.envs.mujoco.Hopper` | 250 | 3 | 15 / 16\* |
|
||||
| `fancy/AntJump-v0` | Ant Jump task, based on Gymnasium's `gym.envs.mujoco.Ant` | 200 | 8 | 119 |
|
||||
| `fancy/HalfCheetahJump-v0` | HalfCheetah Jump task, based on Gymnasium's `gym.envs.mujoco.HalfCheetah` | 100 | 6 | 112 |
|
||||
| `fancy/HopperJumpOnBox-v0` | Hopper Jump on Box task, based on Gymnasium's `gym.envs.mujoco.Hopper` | 250 | 4 | 16 / 100\* |
|
||||
|
||||
@ -25,10 +25,11 @@ from .mujoco.hopper_throw.hopper_throw_in_basket import MAX_EPISODE_STEPS_HOPPER
|
||||
from .mujoco.walker_2d_jump.walker_2d_jump import MAX_EPISODE_STEPS_WALKERJUMP
|
||||
from .mujoco.box_pushing.box_pushing_env import BoxPushingDense, BoxPushingTemporalSparse, \
|
||||
BoxPushingTemporalSpatialSparse, MAX_EPISODE_STEPS_BOX_PUSHING
|
||||
from .mujoco.table_tennis.table_tennis_env import TableTennisEnv, TableTennisWind, TableTennisGoalSwitching, \
|
||||
MAX_EPISODE_STEPS_TABLE_TENNIS
|
||||
from .mujoco.table_tennis.table_tennis_env import TableTennisEnv, TableTennisWind, TableTennisGoalSwitching, TableTennisMarkov, \
|
||||
MAX_EPISODE_STEPS_TABLE_TENNIS, MAX_EPISODE_STEPS_TABLE_TENNIS_MARKOV_VER
|
||||
from .mujoco.table_tennis.mp_wrapper import TT_MPWrapper as MPWrapper_TableTennis
|
||||
from .mujoco.table_tennis.mp_wrapper import TT_MPWrapper_Replan as MPWrapper_TableTennis_Replan
|
||||
from .mujoco.table_tennis.mp_wrapper import TTRndRobot_MPWrapper as MPWrapper_TableTennis_Rnd
|
||||
from .mujoco.table_tennis.mp_wrapper import TTVelObs_MPWrapper as MPWrapper_TableTennis_VelObs
|
||||
from .mujoco.table_tennis.mp_wrapper import TTVelObs_MPWrapper_Replan as MPWrapper_TableTennis_VelObs_Replan
|
||||
|
||||
@ -135,6 +136,19 @@ register(
|
||||
}
|
||||
)
|
||||
|
||||
register(
|
||||
id='fancy/HopperJumpMarkov-v0',
|
||||
entry_point='fancy_gym.envs.mujoco:HopperJumpMarkovRew',
|
||||
mp_wrapper=mujoco.hopper_jump.MPWrapper,
|
||||
max_episode_steps=MAX_EPISODE_STEPS_HOPPERJUMP,
|
||||
kwargs={
|
||||
"sparse": False,
|
||||
"healthy_reward": 1.0,
|
||||
"contact_weight": 0.0,
|
||||
"height_weight": 3.0,
|
||||
}
|
||||
)
|
||||
|
||||
# TODO: Add [MPs] later when finished (old TODO I moved here during refactor)
|
||||
register(
|
||||
id='fancy/AntJump-v0',
|
||||
@ -290,6 +304,37 @@ register(
|
||||
}
|
||||
)
|
||||
|
||||
register(
|
||||
id='fancy/TableTennisRndRobot-v0',
|
||||
entry_point='fancy_gym.envs.mujoco:TableTennisRandomInit',
|
||||
mp_wrapper=MPWrapper_TableTennis_Rnd,
|
||||
max_episode_steps=MAX_EPISODE_STEPS_TABLE_TENNIS,
|
||||
kwargs={
|
||||
'random_pos_scale': 0.1,
|
||||
'random_vel_scale': 0.0,
|
||||
}
|
||||
)
|
||||
|
||||
register(
|
||||
id='fancy/TableTennisMarkov-v0',
|
||||
mp_wrapper=MPWrapper_TableTennis,
|
||||
entry_point='fancy_gym.envs.mujoco:TableTennisMarkov',
|
||||
max_episode_steps=MAX_EPISODE_STEPS_TABLE_TENNIS_MARKOV_VER,
|
||||
kwargs={
|
||||
}
|
||||
)
|
||||
|
||||
register(
|
||||
id='fancy/TableTennisRndRobotMarkov-v0',
|
||||
mp_wrapper=MPWrapper_TableTennis_Rnd,
|
||||
entry_point='fancy_gym.envs.mujoco:TableTennisMarkov',
|
||||
max_episode_steps=MAX_EPISODE_STEPS_TABLE_TENNIS_MARKOV_VER,
|
||||
kwargs={
|
||||
'random_pos_scale': 0.1,
|
||||
'random_vel_scale': 0.0,
|
||||
}
|
||||
)
|
||||
|
||||
# Air Hockey environments
|
||||
for env_mode in ["7dof-hit", "7dof-defend", "3dof-hit", "3dof-defend", "7dof-hit-airhockit2023", "7dof-defend-airhockit2023"]:
|
||||
register(
|
||||
|
||||
@ -1,14 +1,14 @@
|
||||
from .ant_jump.ant_jump import AntJumpEnv
|
||||
from .beerpong.beerpong import BeerPongEnv, BeerPongEnvStepBasedEpisodicReward
|
||||
from .half_cheetah_jump.half_cheetah_jump import HalfCheetahJumpEnv
|
||||
from .hopper_jump.hopper_jump import HopperJumpEnv
|
||||
from .hopper_jump.hopper_jump import HopperJumpEnv, HopperJumpMarkovRew
|
||||
from .hopper_jump.hopper_jump_on_box import HopperJumpOnBoxEnv
|
||||
from .hopper_throw.hopper_throw import HopperThrowEnv
|
||||
from .hopper_throw.hopper_throw_in_basket import HopperThrowInBasketEnv
|
||||
from .reacher.reacher import ReacherEnv
|
||||
from .walker_2d_jump.walker_2d_jump import Walker2dJumpEnv
|
||||
from .box_pushing.box_pushing_env import BoxPushingDense, BoxPushingTemporalSparse, BoxPushingTemporalSpatialSparse
|
||||
from .table_tennis.table_tennis_env import TableTennisEnv, TableTennisWind, TableTennisGoalSwitching
|
||||
from .table_tennis.table_tennis_env import TableTennisEnv, TableTennisWind, TableTennisGoalSwitching, TableTennisMarkov, TableTennisRandomInit
|
||||
|
||||
try:
|
||||
from .air_hockey.air_hockey_env_wrapper import AirHockeyEnv
|
||||
|
||||
@ -7,6 +7,7 @@ from fancy_gym.envs.mujoco.box_pushing.box_pushing_utils import rot_to_quat, get
|
||||
from fancy_gym.envs.mujoco.box_pushing.box_pushing_utils import rot_to_quat, get_quaternion_error, rotation_distance
|
||||
from fancy_gym.envs.mujoco.box_pushing.box_pushing_utils import q_max, q_min, q_dot_max, q_torque_max
|
||||
from fancy_gym.envs.mujoco.box_pushing.box_pushing_utils import desired_rod_quat
|
||||
from fancy_gym.envs.mujoco.box_pushing.box_pushing_utils import calculate_jerk_profile, calculate_mean_squared_jerk, calculate_dimensionless_jerk, calculate_maximum_jerk
|
||||
|
||||
import mujoco
|
||||
|
||||
@ -50,6 +51,7 @@ class BoxPushingEnvBase(MujocoEnv, utils.EzPickle):
|
||||
self._desired_rod_quat = desired_rod_quat
|
||||
|
||||
self._episode_energy = 0.
|
||||
self.velocity_profile = []
|
||||
|
||||
self.observation_space = spaces.Box(
|
||||
low=-np.inf, high=np.inf, shape=(28,), dtype=np.float64
|
||||
@ -69,6 +71,8 @@ class BoxPushingEnvBase(MujocoEnv, utils.EzPickle):
|
||||
|
||||
unstable_simulation = False
|
||||
|
||||
self.velocity_profile.append(self.data.qvel[:7].copy())
|
||||
|
||||
try:
|
||||
self.do_simulation(resultant_action, self.frame_skip)
|
||||
except Exception as e:
|
||||
@ -98,11 +102,15 @@ class BoxPushingEnvBase(MujocoEnv, utils.EzPickle):
|
||||
obs = self._get_obs()
|
||||
box_goal_pos_dist = 0. if not episode_end else np.linalg.norm(box_pos - target_pos)
|
||||
box_goal_quat_dist = 0. if not episode_end else rotation_distance(box_quat, target_quat)
|
||||
mean_squared_jerk, maximum_jerk, dimensionless_jerk = (0.0,0.0,0.0) if not episode_end else self.calculate_smoothness_metrics(np.array(self.velocity_profile), self.dt)
|
||||
infos = {
|
||||
'episode_end': episode_end,
|
||||
'box_goal_pos_dist': box_goal_pos_dist,
|
||||
'box_goal_rot_dist': box_goal_quat_dist,
|
||||
'episode_energy': 0. if not episode_end else self._episode_energy,
|
||||
'mean_squared_jerk': mean_squared_jerk,
|
||||
'maximum_jerk': maximum_jerk,
|
||||
'dimensionless_jerk': dimensionless_jerk,
|
||||
'is_success': True if episode_end and box_goal_pos_dist < 0.05 and box_goal_quat_dist < 0.5 else False,
|
||||
'num_steps': self._steps
|
||||
}
|
||||
@ -119,6 +127,26 @@ class BoxPushingEnvBase(MujocoEnv, utils.EzPickle):
|
||||
self.render_active = True
|
||||
return super().render()
|
||||
|
||||
def calculate_smoothness_metrics(self, velocity_profile, dt):
|
||||
"""
|
||||
Calculates the smoothness metrics for the given velocity profile.
|
||||
param velocity_profile: np.array
|
||||
The array containing the movement velocity profile.
|
||||
param dt: float
|
||||
The sampling time interval of the data.
|
||||
return mean_squared_jerk: float
|
||||
The mean squared jerk estimate of the given movement's smoothness.
|
||||
return maximum_jerk: float
|
||||
The maximum jerk estimate of the given movement's smoothness.
|
||||
return dimensionless_jerk: float
|
||||
The dimensionless jerk estimate of the given movement's smoothness.
|
||||
"""
|
||||
jerk_profile = calculate_jerk_profile(velocity_profile, dt)
|
||||
mean_squared_jerk = calculate_mean_squared_jerk(jerk_profile)
|
||||
maximum_jerk = calculate_maximum_jerk(jerk_profile)
|
||||
dimensionless_jerk = calculate_dimensionless_jerk(jerk_profile, velocity_profile, dt)
|
||||
return mean_squared_jerk, maximum_jerk, dimensionless_jerk
|
||||
|
||||
def reset_model(self):
|
||||
# rest box to initial position
|
||||
self.set_state(self.init_qpos_box_pushing, self.init_qvel_box_pushing)
|
||||
|
||||
@ -51,3 +51,19 @@ def rot_to_quat(theta, axis):
|
||||
quant[0] = np.sin(theta / 2.)
|
||||
quant[1:] = np.cos(theta / 2.) * axis
|
||||
return quant
|
||||
|
||||
def calculate_jerk_profile(velocity_profile, dt):
|
||||
jerk = np.diff(velocity_profile, 2, 0) / pow(dt, 2)
|
||||
return jerk
|
||||
|
||||
def calculate_mean_squared_jerk(jerk_profile):
|
||||
return np.mean(pow(jerk_profile, 2))
|
||||
|
||||
def calculate_maximum_jerk(jerk_profile):
|
||||
return np.max(abs(jerk_profile))
|
||||
|
||||
def calculate_dimensionless_jerk(jerk_profile, velocity_profile, dt):
|
||||
sum_squared_jerk = np.sum(pow(jerk_profile, 2), 0)
|
||||
duration = len(velocity_profile) * dt
|
||||
peak_velocity = np.max(abs(velocity_profile), 0)
|
||||
return np.mean(sum_squared_jerk * pow(duration, 3) / pow(peak_velocity, 2))
|
||||
@ -272,76 +272,100 @@ class HopperJumpEnv(HopperEnvCustomXML):
|
||||
return True
|
||||
return False
|
||||
|
||||
# # TODO is that needed? if so test it
|
||||
# class HopperJumpStepEnv(HopperJumpEnv):
|
||||
#
|
||||
# def __init__(self,
|
||||
# xml_file='hopper_jump.xml',
|
||||
# forward_reward_weight=1.0,
|
||||
# ctrl_cost_weight=1e-3,
|
||||
# healthy_reward=1.0,
|
||||
# height_weight=3,
|
||||
# dist_weight=3,
|
||||
# terminate_when_unhealthy=False,
|
||||
# healthy_state_range=(-100.0, 100.0),
|
||||
# healthy_z_range=(0.5, float('inf')),
|
||||
# healthy_angle_range=(-float('inf'), float('inf')),
|
||||
# reset_noise_scale=5e-3,
|
||||
# exclude_current_positions_from_observation=False
|
||||
# ):
|
||||
#
|
||||
# self._height_weight = height_weight
|
||||
# self._dist_weight = dist_weight
|
||||
# super().__init__(xml_file, forward_reward_weight, ctrl_cost_weight, healthy_reward, terminate_when_unhealthy,
|
||||
# healthy_state_range, healthy_z_range, healthy_angle_range, reset_noise_scale,
|
||||
# exclude_current_positions_from_observation)
|
||||
#
|
||||
# def step(self, action):
|
||||
# self._steps += 1
|
||||
#
|
||||
# self.do_simulation(action, self.frame_skip)
|
||||
#
|
||||
# height_after = self.get_body_com("torso")[2]
|
||||
# site_pos_after = self.data.site('foot_site').xpos.copy()
|
||||
# self.max_height = max(height_after, self.max_height)
|
||||
#
|
||||
# ctrl_cost = self.control_cost(action)
|
||||
# healthy_reward = self.healthy_reward
|
||||
# height_reward = self._height_weight * height_after
|
||||
# goal_dist = np.linalg.norm(site_pos_after - np.array([self.goal, 0, 0]))
|
||||
# goal_dist_reward = -self._dist_weight * goal_dist
|
||||
# dist_reward = self._forward_reward_weight * (goal_dist_reward + height_reward)
|
||||
#
|
||||
# rewards = dist_reward + healthy_reward
|
||||
# costs = ctrl_cost
|
||||
# done = False
|
||||
#
|
||||
# # This is only for logging the distance to goal when first having the contact
|
||||
# has_floor_contact = self._is_floor_foot_contact() if not self.contact_with_floor else False
|
||||
#
|
||||
# if not self.init_floor_contact:
|
||||
# self.init_floor_contact = has_floor_contact
|
||||
# if self.init_floor_contact and not self.has_left_floor:
|
||||
# self.has_left_floor = not has_floor_contact
|
||||
# if not self.contact_with_floor and self.has_left_floor:
|
||||
# self.contact_with_floor = has_floor_contact
|
||||
#
|
||||
# if self.contact_dist is None and self.contact_with_floor:
|
||||
# self.contact_dist = goal_dist
|
||||
#
|
||||
# ##############################################################
|
||||
#
|
||||
# observation = self._get_obs()
|
||||
# reward = rewards - costs
|
||||
# info = {
|
||||
# 'height': height_after,
|
||||
# 'x_pos': site_pos_after,
|
||||
# 'max_height': copy.copy(self.max_height),
|
||||
# 'goal': copy.copy(self.goal),
|
||||
# 'goal_dist': goal_dist,
|
||||
# 'height_rew': height_reward,
|
||||
# 'healthy_reward': healthy_reward,
|
||||
# 'healthy': copy.copy(self.is_healthy),
|
||||
# 'contact_dist': copy.copy(self.contact_dist) or 0
|
||||
# }
|
||||
# return observation, reward, done, info
|
||||
class HopperJumpMarkovRew(HopperJumpEnv):
|
||||
def step(self, action):
|
||||
self._steps += 1
|
||||
|
||||
self.do_simulation(action, self.frame_skip)
|
||||
|
||||
height_after = self.get_body_com("torso")[2]
|
||||
# site_pos_after = self.data.get_site_xpos('foot_site')
|
||||
site_pos_after = self.data.site('foot_site').xpos
|
||||
self.max_height = max(height_after, self.max_height)
|
||||
|
||||
has_floor_contact = self._is_floor_foot_contact() if not self.contact_with_floor else False
|
||||
|
||||
if not self.init_floor_contact:
|
||||
self.init_floor_contact = has_floor_contact
|
||||
if self.init_floor_contact and not self.has_left_floor:
|
||||
self.has_left_floor = not has_floor_contact
|
||||
if not self.contact_with_floor and self.has_left_floor:
|
||||
self.contact_with_floor = has_floor_contact
|
||||
|
||||
ctrl_cost = self.control_cost(action)
|
||||
costs = ctrl_cost
|
||||
terminated = False
|
||||
truncated = False
|
||||
|
||||
goal_dist = np.linalg.norm(site_pos_after - self.goal)
|
||||
if self.contact_dist is None and self.contact_with_floor:
|
||||
self.contact_dist = goal_dist
|
||||
|
||||
rewards = 0
|
||||
if not self.sparse or (self.sparse and self._steps >= MAX_EPISODE_STEPS_HOPPERJUMP):
|
||||
healthy_reward = self.healthy_reward
|
||||
distance_reward = -goal_dist * self._dist_weight
|
||||
height_reward = (self.max_height if self.sparse else height_after) * self._height_weight
|
||||
contact_reward = -(self.contact_dist or 5) * self._contact_weight
|
||||
rewards = self._forward_reward_weight * (distance_reward + height_reward + contact_reward + healthy_reward)
|
||||
|
||||
observation = self._get_obs()
|
||||
|
||||
# While loop to simulate the process after jump to make the task Markovian
|
||||
if self.sparse and self.has_left_floor:
|
||||
while self._steps < MAX_EPISODE_STEPS_HOPPERJUMP:
|
||||
# Simulate to the end of the episode
|
||||
self._steps += 1
|
||||
|
||||
try:
|
||||
self.do_simulation(np.zeros_like(action), self.frame_skip)
|
||||
except Exception as e:
|
||||
print(e)
|
||||
|
||||
height_after = self.get_body_com("torso")[2]
|
||||
#site_pos_after = self.data.get_site_xpos('foot_site')
|
||||
site_pos_after = self.data.site('foot_site').xpos
|
||||
self.max_height = max(height_after, self.max_height)
|
||||
|
||||
has_floor_contact = self._is_floor_foot_contact() if not self.contact_with_floor else False
|
||||
|
||||
if not self.init_floor_contact:
|
||||
self.init_floor_contact = has_floor_contact
|
||||
if self.init_floor_contact and not self.has_left_floor:
|
||||
self.has_left_floor = not has_floor_contact
|
||||
if not self.contact_with_floor and self.has_left_floor:
|
||||
self.contact_with_floor = has_floor_contact
|
||||
|
||||
ctrl_cost = self.control_cost(action)
|
||||
costs = ctrl_cost
|
||||
done = False
|
||||
|
||||
goal_dist = np.linalg.norm(site_pos_after - self.goal)
|
||||
if self.contact_dist is None and self.contact_with_floor:
|
||||
self.contact_dist = goal_dist
|
||||
|
||||
rewards = 0
|
||||
|
||||
# Task has reached the end, compute the sparse reward
|
||||
done = True
|
||||
healthy_reward = self.healthy_reward
|
||||
distance_reward = -goal_dist * self._dist_weight
|
||||
height_reward = (self.max_height if self.sparse else height_after) * self._height_weight
|
||||
contact_reward = -(self.contact_dist or 5) * self._contact_weight
|
||||
rewards = self._forward_reward_weight * (distance_reward + height_reward + contact_reward + healthy_reward)
|
||||
|
||||
reward = rewards - costs
|
||||
info = dict(
|
||||
height=height_after,
|
||||
x_pos=site_pos_after,
|
||||
max_height=self.max_height,
|
||||
goal=self.goal[:1],
|
||||
goal_dist=goal_dist,
|
||||
height_rew=self.max_height,
|
||||
healthy_reward=self.healthy_reward,
|
||||
healthy=self.is_healthy,
|
||||
contact_dist=self.contact_dist or 0,
|
||||
num_steps=self._steps,
|
||||
has_left_floor=self.has_left_floor
|
||||
)
|
||||
return observation, reward, terminated, truncated, info
|
||||
|
||||
@ -151,3 +151,15 @@ class TTVelObs_MPWrapper_Replan(TT_MPWrapper_Replan):
|
||||
[True] * 2, # target landing position
|
||||
# [True] * 1, # time
|
||||
])
|
||||
|
||||
class TTRndRobot_MPWrapper(TT_MPWrapper):
|
||||
@property
|
||||
def context_mask(self):
|
||||
return np.hstack([
|
||||
[True] * 7, # joints position
|
||||
[False] * 7, # joints velocity
|
||||
[True] * 2, # position ball x, y
|
||||
[False] * 1, # position ball z
|
||||
[True] * 2, # target landing position
|
||||
# [True] * 1, # time
|
||||
])
|
||||
@ -5,11 +5,12 @@ from gymnasium import utils, spaces
|
||||
from gymnasium.envs.mujoco import MujocoEnv
|
||||
|
||||
from fancy_gym.envs.mujoco.table_tennis.table_tennis_utils import is_init_state_valid, magnus_force
|
||||
from fancy_gym.envs.mujoco.table_tennis.table_tennis_utils import jnt_pos_low, jnt_pos_high
|
||||
from fancy_gym.envs.mujoco.table_tennis.table_tennis_utils import jnt_pos_low, jnt_pos_high, jnt_vel_low, jnt_vel_high
|
||||
|
||||
import mujoco
|
||||
|
||||
MAX_EPISODE_STEPS_TABLE_TENNIS = 350
|
||||
MAX_EPISODE_STEPS_TABLE_TENNIS_MARKOV_VER = 300
|
||||
|
||||
CONTEXT_BOUNDS_2DIMS = np.array([[-1.0, -0.65], [-0.2, 0.65]])
|
||||
CONTEXT_BOUNDS_4DIMS = np.array([[-1.0, -0.65, -1.0, -0.65],
|
||||
@ -18,6 +19,9 @@ CONTEXT_BOUNDS_SWICHING = np.array([[-1.0, -0.65, -1.0, 0.],
|
||||
[-0.2, 0.65, -0.2, 0.65]])
|
||||
|
||||
|
||||
DEFAULT_ROBOT_INIT_POS = np.array([0.0, 0.0, 0.0, 1.5, 0.0, 0.0, 1.5])
|
||||
DEFAULT_ROBOT_INIT_VEL = np.array([0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0])
|
||||
|
||||
class TableTennisEnv(MujocoEnv, utils.EzPickle):
|
||||
"""
|
||||
7 DoF table tennis environment
|
||||
@ -34,7 +38,11 @@ class TableTennisEnv(MujocoEnv, utils.EzPickle):
|
||||
|
||||
def __init__(self, ctxt_dim: int = 4, frame_skip: int = 4,
|
||||
goal_switching_step: int = None,
|
||||
enable_artificial_wind: bool = False, **kwargs):
|
||||
enable_artificial_wind: bool = False,
|
||||
random_pos_scale: float = 0.0,
|
||||
random_vel_scale: float = 0.0,
|
||||
**kwargs,
|
||||
):
|
||||
utils.EzPickle.__init__(**locals())
|
||||
self._steps = 0
|
||||
|
||||
@ -48,6 +56,10 @@ class TableTennisEnv(MujocoEnv, utils.EzPickle):
|
||||
|
||||
self._id_set = False
|
||||
|
||||
# initial robot state
|
||||
self._random_pos_scale = random_pos_scale
|
||||
self._random_vel_scale = random_vel_scale
|
||||
|
||||
# reward calculation
|
||||
self.ball_landing_pos = None
|
||||
self._goal_pos = np.zeros(2)
|
||||
@ -158,7 +170,7 @@ class TableTennisEnv(MujocoEnv, utils.EzPickle):
|
||||
"num_steps": self._steps,
|
||||
}
|
||||
|
||||
terminated, truncated = self._terminated, False
|
||||
terminated, truncated = self._terminated, self._steps == MAX_EPISODE_STEPS_TABLE_TENNIS
|
||||
|
||||
if self.render_active and self.render_mode=='human':
|
||||
self.render()
|
||||
@ -176,6 +188,17 @@ class TableTennisEnv(MujocoEnv, utils.EzPickle):
|
||||
return True
|
||||
return False
|
||||
|
||||
def get_initial_robot_state(self):
|
||||
|
||||
robot_init_pos = DEFAULT_ROBOT_INIT_POS + \
|
||||
self.np_random.uniform(-1.0, 1.0, size=7) *\
|
||||
np.array([5.2, 4.0, 5.6, 4.0, 6.1, 3.2, 4.4]) *\
|
||||
self._random_pos_scale
|
||||
|
||||
robot_init_vel = DEFAULT_ROBOT_INIT_VEL + self.np_random.uniform(-1.0, 1.0, size=7) * self._random_vel_scale
|
||||
|
||||
return np.clip(robot_init_pos, jnt_pos_low, jnt_pos_high), np.clip(robot_init_vel, jnt_vel_low, jnt_vel_high)
|
||||
|
||||
def reset_model(self):
|
||||
self._steps = 0
|
||||
self._init_ball_state = self._generate_valid_init_ball(random_pos=True, random_vel=False)
|
||||
@ -192,8 +215,10 @@ class TableTennisEnv(MujocoEnv, utils.EzPickle):
|
||||
|
||||
self.model.body_pos[5] = np.concatenate([self._goal_pos, [0.77]])
|
||||
|
||||
self.data.qpos[:7] = np.array([0., 0., 0., 1.5, 0., 0., 1.5])
|
||||
self.data.qvel[:7] = np.zeros(7)
|
||||
robot_init_pos, robot_init_vel = self.get_initial_robot_state()
|
||||
|
||||
self.data.qpos[:7] = robot_init_pos
|
||||
self.data.qvel[:7] = robot_init_vel
|
||||
|
||||
mujoco.mj_forward(self.model, self.data)
|
||||
|
||||
@ -266,7 +291,7 @@ class TableTennisEnv(MujocoEnv, utils.EzPickle):
|
||||
def get_invalid_traj_step_return(self, action, pos_traj, contextual_obs, tau_bound, delay_bound):
|
||||
obs = self._get_obs() if contextual_obs else np.concatenate([self._get_obs(), np.array([0])]) # 0 for invalid traj
|
||||
penalty = self._get_traj_invalid_penalty(action, pos_traj, tau_bound, delay_bound)
|
||||
return obs, penalty, True, False, {
|
||||
return obs, penalty, False, True, {
|
||||
"hit_ball": [False],
|
||||
"ball_returned_success": [False],
|
||||
"land_dist_error": [10.],
|
||||
@ -283,6 +308,179 @@ class TableTennisEnv(MujocoEnv, utils.EzPickle):
|
||||
return False, pos_traj, vel_traj
|
||||
return True, pos_traj, vel_traj
|
||||
|
||||
class TableTennisMarkov(TableTennisEnv):
|
||||
def _get_reward2(self, hit_now, land_now):
|
||||
|
||||
# Phase 1 not hit ball
|
||||
if not self._hit_ball:
|
||||
# Not hit ball
|
||||
min_r_b_dist = np.min(np.linalg.norm(np.array(self._ball_traj) - np.array(self._racket_traj), axis=1))
|
||||
return 0.005 * (1 - np.tanh(min_r_b_dist**2))
|
||||
|
||||
# Phase 2 hit ball now
|
||||
elif self._hit_ball and hit_now:
|
||||
return 2
|
||||
|
||||
# Phase 3 hit ball already and not land yet
|
||||
elif self._hit_ball and self._ball_landing_pos is None:
|
||||
min_b_des_b_dist = np.min(np.linalg.norm(np.array(self._ball_traj)[:,:2] - self._goal_pos[:2], axis=1))
|
||||
return 0.02 * (1 - np.tanh(min_b_des_b_dist**2))
|
||||
|
||||
# Phase 4 hit ball already and land now
|
||||
elif self._hit_ball and land_now:
|
||||
over_net_bonus = int(self._ball_landing_pos[0] < 0)
|
||||
min_b_des_b_land_dist = np.linalg.norm(self._goal_pos[:2] - self._ball_landing_pos[:2])
|
||||
return 4 * (1 - np.tanh(min_b_des_b_land_dist ** 2)) + over_net_bonus
|
||||
|
||||
# Phase 5 hit ball already and land already
|
||||
elif self._hit_ball and not land_now and self._ball_landing_pos is not None:
|
||||
return 0
|
||||
|
||||
else:
|
||||
raise NotImplementedError
|
||||
|
||||
def _get_reward(self, terminated):
|
||||
# if not terminated:
|
||||
# return 0
|
||||
|
||||
min_r_b_dist = np.min(np.linalg.norm(np.array(self._ball_traj) - np.array(self._racket_traj), axis=1))
|
||||
if not self._hit_ball:
|
||||
# Not hit ball
|
||||
return 0.2 * (1 - np.tanh(min_r_b_dist**2))
|
||||
elif self._ball_landing_pos is None:
|
||||
# Hit ball but not landing pos
|
||||
min_b_des_b_dist = np.min(np.linalg.norm(np.array(self._ball_traj)[:,:2] - self._goal_pos[:2], axis=1))
|
||||
return 2 + (1 - np.tanh(min_b_des_b_dist**2))
|
||||
else:
|
||||
# Hit ball and land
|
||||
min_b_des_b_land_dist = np.linalg.norm(self._goal_pos[:2] - self._ball_landing_pos[:2])
|
||||
over_net_bonus = int(self._ball_landing_pos[0] < 0)
|
||||
return 2 + 4 * (1 - np.tanh(min_b_des_b_land_dist ** 2)) + over_net_bonus
|
||||
|
||||
|
||||
def _get_traj_invalid_penalty(self, action, pos_traj, tau_bound, delay_bound):
|
||||
tau_invalid_penalty = 3 * (np.max([0, action[0] - tau_bound[1]]) + np.max([0, tau_bound[0] - action[0]]))
|
||||
delay_invalid_penalty = 3 * (np.max([0, action[1] - delay_bound[1]]) + np.max([0, delay_bound[0] - action[1]]))
|
||||
violate_high_bound_error = np.mean(np.maximum(pos_traj - jnt_pos_high, 0))
|
||||
violate_low_bound_error = np.mean(np.maximum(jnt_pos_low - pos_traj, 0))
|
||||
invalid_penalty = tau_invalid_penalty + delay_invalid_penalty + \
|
||||
violate_high_bound_error + violate_low_bound_error
|
||||
return -invalid_penalty
|
||||
|
||||
def get_invalid_traj_step_penalty(self, pos_traj):
|
||||
violate_high_bound_error = (
|
||||
np.maximum(pos_traj - jnt_pos_high, 0).mean())
|
||||
violate_low_bound_error = (
|
||||
np.maximum(jnt_pos_low - pos_traj, 0).mean())
|
||||
invalid_penalty = violate_high_bound_error + violate_low_bound_error
|
||||
|
||||
|
||||
def _update_game_state(self, action):
|
||||
for _ in range(self.frame_skip):
|
||||
if self._enable_artificial_wind:
|
||||
self.data.qfrc_applied[-2] = self._artificial_force
|
||||
try:
|
||||
self.do_simulation(action, 1)
|
||||
except Exception as e:
|
||||
print("Simulation get unstable return with MujocoException: ", e)
|
||||
unstable_simulation = True
|
||||
self._terminated = True
|
||||
break
|
||||
|
||||
# Update game state
|
||||
if not self._terminated:
|
||||
if not self._hit_ball:
|
||||
self._hit_ball = self._contact_checker(self._ball_contact_id, self._bat_front_id) or \
|
||||
self._contact_checker(self._ball_contact_id, self._bat_back_id)
|
||||
if not self._hit_ball:
|
||||
ball_land_on_floor_no_hit = self._contact_checker(self._ball_contact_id, self._floor_contact_id)
|
||||
if ball_land_on_floor_no_hit:
|
||||
self._ball_landing_pos = self.data.body("target_ball").xpos.copy()
|
||||
self._terminated = True
|
||||
if self._hit_ball and not self._ball_contact_after_hit:
|
||||
if self._contact_checker(self._ball_contact_id, self._floor_contact_id): # first check contact with floor
|
||||
self._ball_contact_after_hit = True
|
||||
self._ball_landing_pos = self.data.geom("target_ball_contact").xpos.copy()
|
||||
self._terminated = True
|
||||
elif self._contact_checker(self._ball_contact_id, self._table_contact_id): # second check contact with table
|
||||
self._ball_contact_after_hit = True
|
||||
self._ball_landing_pos = self.data.geom("target_ball_contact").xpos.copy()
|
||||
if self._ball_landing_pos[0] < 0.: # ball lands on the opponent side
|
||||
self._ball_return_success = True
|
||||
self._terminated = True
|
||||
|
||||
# update ball trajectory & racket trajectory
|
||||
self._ball_traj.append(self.data.body("target_ball").xpos.copy())
|
||||
self._racket_traj.append(self.data.geom("bat").xpos.copy())
|
||||
|
||||
def ball_racket_contact(self):
|
||||
return self._contact_checker(self._ball_contact_id, self._bat_front_id) or \
|
||||
self._contact_checker(self._ball_contact_id, self._bat_back_id)
|
||||
|
||||
def step(self, action):
|
||||
if not self._id_set:
|
||||
self._set_ids()
|
||||
|
||||
unstable_simulation = False
|
||||
hit_already = self._hit_ball
|
||||
if self._steps == self._goal_switching_step and self.np_random.uniform() < 0.5:
|
||||
new_goal_pos = self._generate_goal_pos(random=True)
|
||||
new_goal_pos[1] = -new_goal_pos[1]
|
||||
self._goal_pos = new_goal_pos
|
||||
self.model.body_pos[5] = np.concatenate([self._goal_pos, [0.77]])
|
||||
mujoco.mj_forward(self.model, self.data)
|
||||
|
||||
self._update_game_state(action)
|
||||
self._steps += 1
|
||||
|
||||
obs = self._get_obs()
|
||||
|
||||
# Compute reward
|
||||
if unstable_simulation:
|
||||
reward = -25
|
||||
else:
|
||||
# reward = self._get_reward(self._terminated)
|
||||
# hit_now = not hit_already and self._hit_ball
|
||||
hit_finish = self._hit_ball and not self.ball_racket_contact()
|
||||
|
||||
if hit_finish:
|
||||
# Clean the ball and racket traj before hit
|
||||
self._ball_traj = []
|
||||
self._racket_traj = []
|
||||
|
||||
# Simulate the rest of the traj
|
||||
reward = self._get_reward2(True, False)
|
||||
while self._steps < MAX_EPISODE_STEPS_TABLE_TENNIS_MARKOV_VER:
|
||||
land_already = self._ball_landing_pos is not None
|
||||
self._update_game_state(np.zeros_like(action))
|
||||
self._steps += 1
|
||||
|
||||
land_now = (not land_already
|
||||
and self._ball_landing_pos is not None)
|
||||
temp_reward = self._get_reward2(False, land_now)
|
||||
# print(temp_reward)
|
||||
reward += temp_reward
|
||||
|
||||
# Uncomment the line below to visualize the sim after hit
|
||||
# self.render(mode="human")
|
||||
else:
|
||||
reward = self._get_reward2(False, False)
|
||||
|
||||
# Update ball landing error
|
||||
land_dist_err = np.linalg.norm(self._ball_landing_pos[:-1] - self._goal_pos) \
|
||||
if self._ball_landing_pos is not None else 10.
|
||||
|
||||
info = {
|
||||
"hit_ball": self._hit_ball,
|
||||
"ball_returned_success": self._ball_return_success,
|
||||
"land_dist_error": land_dist_err,
|
||||
"is_success": self._ball_return_success and land_dist_err < 0.2,
|
||||
"num_steps": self._steps,
|
||||
}
|
||||
|
||||
terminated, truncated = self._terminated, self._steps == MAX_EPISODE_STEPS_TABLE_TENNIS_MARKOV_VER
|
||||
|
||||
return obs, reward, terminated, truncated, info
|
||||
|
||||
class TableTennisWind(TableTennisEnv):
|
||||
def __init__(self, ctxt_dim: int = 4, frame_skip: int = 4, **kwargs):
|
||||
@ -305,7 +503,17 @@ class TableTennisWind(TableTennisEnv):
|
||||
])
|
||||
return obs
|
||||
|
||||
|
||||
class TableTennisGoalSwitching(TableTennisEnv):
|
||||
def __init__(self, frame_skip: int = 4, goal_switching_step: int = 99, **kwargs):
|
||||
super().__init__(frame_skip=frame_skip, goal_switching_step=goal_switching_step, **kwargs)
|
||||
|
||||
|
||||
class TableTennisRandomInit(TableTennisEnv):
|
||||
def __init__(self, ctxt_dim: int = 4, frame_skip: int = 4,
|
||||
random_pos_scale: float = 1.0,
|
||||
random_vel_scale: float = 0.0,
|
||||
**kwargs):
|
||||
super().__init__(ctxt_dim=ctxt_dim, frame_skip=frame_skip,
|
||||
random_pos_scale=random_pos_scale,
|
||||
random_vel_scale=random_vel_scale,
|
||||
**kwargs)
|
||||
@ -2,6 +2,10 @@ import numpy as np
|
||||
|
||||
jnt_pos_low = np.array([-2.6, -2.0, -2.8, -0.9, -4.8, -1.6, -2.2])
|
||||
jnt_pos_high = np.array([2.6, 2.0, 2.8, 3.1, 1.3, 1.6, 2.2])
|
||||
|
||||
jnt_vel_low = np.ones(7) * -7
|
||||
jnt_vel_high = np.ones(7) * 7
|
||||
|
||||
delay_bound = [0.05, 0.15]
|
||||
tau_bound = [0.5, 1.5]
|
||||
|
||||
|
||||
@ -52,7 +52,7 @@ class FixMetaworldIgnoresSeedOnResetWrapper(gym.Wrapper, gym.utils.RecordConstru
|
||||
|
||||
def reset(self, **kwargs):
|
||||
if 'seed' in kwargs:
|
||||
print('[!] You just called .reset on a Metaworld env and supplied a seed. Metaworld curretly does not correctly implement seeding. Do not rely on deterministic behavior.')
|
||||
print('[Fancy Gym] You just called .reset on a Metaworld env and supplied a seed. Metaworld curretly does not correctly implement seeding. Do not rely on deterministic behavior.')
|
||||
self.env.seed(kwargs['seed'])
|
||||
return self.env.reset(**kwargs)
|
||||
|
||||
|
||||
Loading…
Reference in New Issue
Block a user