# emo-gym

**Repository Path**: rock3t/emo-gym

## Basic Information

- **Project Name**: emo-gym
- **Description**: 一个强化学习双足框架，基于humanoid-gym，用自己的urdf文件实现
- **Primary Language**: Unknown
- **License**: Not specified
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 1
- **Forks**: 0
- **Created**: 2025-08-19
- **Last Updated**: 2025-08-25

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# emo-gym

## how to use

### train in isaac gym
~~~bash
python train.py --task=emo --max_iterations=1000 --headless --num_envs=4096
~~~

Train 1000 iterations, the robot can stand and walk, it's enough for a simple test

### play in isaac gym


If just to play, set 
~~~python
EXPORT_POLICY = False
RENDER = False
FIX_COMMAND = True
PRINT_OBS = False
~~~
In the last few lines in play.py
~~~bash
python play.py --task=emo --load_run=xxx(optional)
~~~

###  export policy and play in mujoco

first, set the flags in play.py
~~~python
EXPORT_POLICY = True #to export policy
RENDER = False
FIX_COMMAND = True
PRINT_OBS = False
~~~

then, find the policy in the floder

change the policy path below in **sim2sim_emo_test.py**

~~~python
policy = torch.jit.load('your model path') 
~~~

then run 
~~~bash
python sim2sim_emo_test.py 
~~~
to check the result

## tools
In play.py you can set 
~~~python
PRINT_OBS = True
~~~
to print all the obs buf

In sim2sim_emo_test.py,the obs will be print automatically