g1-dance

PyTorch checkpoint for a G1 humanoid locomotion policy trained with ADD (Adversarial Differential Discriminators).

Checkpoint info

Key	Value
Training iterations	10,900
Total environment samples	5,715,263,488
Number of parameters	4,371,991

Usage

import torch

checkpoint = torch.load("model.pt", map_location="cpu")
state_dict = checkpoint["model"]
# Load into your agent:
# agent.load_state_dict(state_dict)