YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
Neo Humanoid Dynamics v2
Overview
Neo Humanoid Dynamics v2 is a Hybrid CNN + Transformer model designed for humanoid robot motion control and action recognition.
Model Details
- Architecture: CNN + Transformer Encoder
- Framework: PyTorch
- Tasks: Action Classification & Motion Forecasting
- Number of Classes: 8
Input Format
- Sequence Length: 80 frames
- Joints: 24
- Features per Joint: 6 (x, y, z, vx, vy, vz)
Output
- Action Label
- Confidence Score
- Future Motion Frames
Supported Actions
- walking
- running
- jumping
- sitting
- standing
- grabbing_object
- pushing
- turning
Hyperparameters
- Learning Rate: 0.0002
- Batch Size: 64
- Epochs: 60
- Optimizer: AdamW
- Scheduler: Cosine Annealing
Dataset
- 15,000 motion sequences
- Balanced dataset
- Data augmentation with noise and rotation
License
Apache 2.0
- Downloads last month
- -
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support