File size: 2,118 Bytes
7c0ded9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5cf5555
7c0ded9
 
 
 
b2f6b92
 
 
 
 
 
 
 
 
 
 
 
 
 
3c4dfdb
b2f6b92
 
 
 
 
 
 
 
 
 
 
 
7c0ded9
 
 
 
 
b2f6b92
 
 
 
 
 
 
 
7c0ded9
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
---
library_name: lerobot
tags:
- model_hub_mixin
- pytorch_model_hub_mixin
- robotics
- dot
license: apache-2.0
datasets:
- lerobot/aloha_sim_insertion_human
pipeline_tag: robotics
---

# Model Card for "Decoder Only Transformer (DOT) Policy" for ALOHA bimanual insert problem

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 ALOHA bimanual insert dataset. It achieves 29.6% success rate vs. 21% for the previous state-of-the-art model (ACT).

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_new_config)

To train the model:

```bash
python lerobot/scripts/train.py \
    --policy.type=dot \
    --dataset.repo_id=lerobot/aloha_sim_insertion_human \
    --env.type=aloha \
    --env.task=AlohaInsertion-v0 \
    --env.episode_length=500 \
    --output_dir=outputs/train/pusht_aloha_insert \
    --batch_size=24  \
    --log_freq=1000 \
    --eval_freq=10000 \
    --save_freq=10000 \
    --offline.steps=100000 \
    --seed=100000 \
    --wandb.enable=true \
    --num_workers=24 \
    --use_amp=true \
    --device=cuda \
    --policy.optimizer_lr=0.00003 \
    --policy.optimizer_min_lr=0.00001 \
    --policy.optimizer_lr_cycle_steps=100000 \
    --policy.train_horizon=150 \
    --policy.inference_horizon=100 \
    --policy.lookback_obs_steps=30 \
    --policy.lookback_aug=5 \
    --policy.rescale_shape="[480,640]" \
    --policy.alpha=0.98 \
    --policy.train_alpha=0.99
```

To evaluate the model:

```bash
python lerobot/scripts/eval.py \
    --policy.path=IliaLarchenko/dot_bimanual_insert \
    --env.type=aloha \
    --env.task=AlohaInsertion-v0 \
    --env.episode_length=500 \
    --eval.n_episodes=1000 \
    --eval.batch_size=100 \
    --seed=1000000
```

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