Comments (5)
We have conducted multiple tests and the issue did not occur. Could you furnish more particulars? Kindly attempt our default policy initially. We offer one trained policy as a baseline. In the event of any system setup disparities, ensure that the baseline policy functions properly. Subsequently, verify whether you continue to encounter the identical problem.
from humanoid-gym.
Hi, the default strategy works fine, but the model after training still doesn't work properly (simtosim.py), I didn't modify any code before training started. I look forward to your reply.
123.mp4
from humanoid-gym.
Do you just pull the project without making any modifications? How many iterations have you gone through in training? Please offer some training logs, like the last few training iterations. That would be helpful.
from humanoid-gym.
You are right, I found that iterating a few more times solved the problem. Thank you very much for your reply
from humanoid-gym.
I will close it as it has been solved. One comment is that more training steps do not guarantee a better sim2real result. After the reward function converges, training for a few hundred more steps is a common practice, and training for at least 1,000 steps is a good starting point.
from humanoid-gym.
Related Issues (20)
- About the simulation environment HOT 2
- How does policy handle privileged observations in real-world? HOT 3
- Mapping between urdf and mjcf HOT 1
- Version compatibility HOT 1
- Sim 2 sim transfer HOT 1
- Inquiry about Including Upper Body Joints in Reinforcement Learning for Humanoid Robot HOT 1
- Failed to create a PhysX CUDA Context Manager. Falling back to CPU. HOT 1
- when will the code and paper of denoising World Model Learning release? HOT 3
- Control a little 10-dof robot HOT 6
- load new URDF
- load new URDF HOT 1
- The kp setting for ankle roll and ankle pitc HOT 2
- I can't evaluate the Trained PPO Policy 'v1' HOT 1
- Pretrained model? HOT 1
- why the env_frictions and body_mass term in privileged_obs_buf is zero? HOT 2
- how to sim2real? HOT 1
- Is there a relationship between the target feet height and the initial robot joint position and the reference joint position? HOT 1
- kinematics and dynamics HOT 1
- install demand
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from humanoid-gym.