README.md

---
library_name: lerobot
tags:
- model_hub_mixin
- pytorch_model_hub_mixin
- robotics
- dot
license: apache-2.0
datasets:
- lerobot/pusht
pipeline_tag: robotics
---

# Model Card for "Decoder Only Transformer (DOT) Policy" for PushT images dataset

Read more about the model and implementation details in the [DOT Policy repository](https://github.com/IliaLarchenko/dot_policy).

This model is trained using the [LeRobot library](https://huggingface.co/lerobot) and achieves state-of-the-art results on behavior cloning on the PushT images dataset. It achieves a 74.2% success rate (and 0.936 average max reward) vs. ~69% for the previous state-of-the-art model (Diffusion and VQ-BET perform the same).

This result is achieved without the checkpoint selection and is easy to reproduce.

You can use this model by installing LeRobot from [this branch](https://github.com/IliaLarchenko/lerobot/tree/dot)

To train the model:

```bash
python lerobot/scripts/train.py \
    --policy.type=dot \
    --dataset.repo_id=lerobot/pusht \
    --env.type=pusht \
    --env.task=PushT-v0 \
    --output_dir=outputs/train/pusht_images \
    --batch_size=24  \
    --log_freq=1000 \
    --eval_freq=10000 \
    --save_freq=50000 \
    --offline.steps=1000000 \
    --seed=100000 \
    --wandb.enable=true \
    --num_workers=24 \
    --use_amp=true \
    --device=cuda \
    --policy.return_every_n=2
```

To evaluate the model:

```bash
python lerobot/scripts/eval.py \
    --policy.path=IliaLarchenko/dot_pusht_images \
    --env.type=pusht \
    --env.task=PushT-v0 \
    --eval.n_episodes=1000 \
    --eval.batch_size=100 \
    --seed=1000000
```

Model size:
- Total parameters: 14.1m
- Trainable parameters: 2.9m