metadata
license: other
tags:
- reinforcement-learning
- robot
- g1
- bfmzero
BFMZero - 29DoF Checkpoint
Trained checkpoint for BFMZero (FBcprAuxAgent / FBcprAuxModel) with 29 degrees of freedom.
Training Status
- Steps: 144,000,000
- Config: Residual MLP with z_dim=128, hidden_dim=1024, 4 hidden layers
- Device: CUDA
Files
| File | Description |
|---|---|
model.safetensors |
Model weights (SafeTensors format) |
optimizers.pth |
Optimizer states |
config.json |
Model architecture configuration |
init_kwargs.json |
Initialization parameters |
train_status.json |
Training step counter |
Usage
from huggingface_hub import hf_hub_download
import torch
# Download model weights
model_path = hf_hub_download(
repo_id="Kennyp-Chen/bfmzero-29dof",
filename="model.safetensors"
)