VisionaryKunal/3DBall-MLAgents
By VisionaryKunal · ↓ 1 · ♥ 0 · Updated 2025-09-07
What's in this repo
This page lists the files in VisionaryKunal/3DBall-MLAgents, including 3DBall.onnx, 3DBall-499484.onnx, 3DBall-500484.onnx.
- 3 ONNX models (.onnx files)
- 17 files, including 3 .onnx, 3 .pt, 3 .mp4, 3 .json, 1 .0, 1 .md
- Last updated 2025-09-07
ONNX files
| File | Size | Download |
|---|---|---|
| 3DBall.onnx | — | Download |
| 3DBall/3DBall-499484.onnx | — | Download |
| 3DBall/3DBall-500484.onnx | — | Download |
How to use
These ONNX files can be loaded with ONNX Runtime. To inspect a model graph, upload the file to the Netron viewer. Download links point to the original files on Hugging Face.
README
3DBall Trained Agent
This is a trained model of a PPO agent playing the 3DBall environment, created using the Unity ML-Agents library. The agent learns to balance a ball on a moving platform for as long as possible.
Training Hyperparameters
The agent was trained using the following configuration from the 3DBall.yaml file:
behaviors:
3DBall:
trainer_type: ppo
hyperparameters:
learning_rate: 0.0003
learning_rate_schedule: linear
beta: 0.0005
epsilon: 0.2
lambd: 0.95
num_epoch: 3
buffer_size: 2048
batch_size: 256
time_horizon: 1024
network_settings:
normalize: false
hidden_units: 128
num_layers: 2
vis_encode_type: simple
reward_signals:
extrinsic:
gamma: 0.99
strength: 1.0
checkpoint_interval: 500000
threaded: true
Video Demo
Here is a video of the trained agent in action, demonstrating the learned behavior.
Your browser does not support the video tag.
Model info
Created: 2025-09-07
Updated: 2025-09-07
View on Hugging FaceFiles (17)
Related models
More from VisionaryKunal
No other models by this author.