| library_name: lerobot | |
| tags: | |
| - act | |
| - pusht | |
| - reinforcement-learning | |
| - robot | |
| # ACT Policy for pusht | |
| This model was trained using [LeRobot](https://github.com/huggingface/lerobot). | |
| ## Model Details | |
| - **Repository**: [vvrs/act-pusht](https://huggingface.co/vvrs/act-pusht) | |
| - **Policy Type**: act | |
| - **Environment**: pusht | |
| ## Hyperparameters | |
| - **n_action_steps**: 10 | |
| - **chunk_size**: 50 | |
| - **dim_model**: 512 | |
| - **n_encoder_layers**: 4 | |
| - **dropout**: 0.1 | |
| - **kl_weight**: 0.26090695687163257 | |
| - **optimizer_lr**: 7.765556097859912e-05 | |
| ## Usage | |
| ```python | |
| from lerobot.common.policies.act.modeling_act import ACTPolicy | |
| from lerobot.common.policies.act.configuration_act import ACTConfig | |
| # Load the policy from the Hub | |
| policy = ACTPolicy.from_pretrained("vvrs/act-pusht") | |
| # Use the policy for inference... | |
| ``` | |