Hugging Face's logo

notmahi
/
act-maze-ball-1

Transformers
Safetensors
pytorch_model_hub_mixin
model_hub_mixin
Inference Endpoints
Model card Files Community
act-maze-ball-1
File size: 2,051 Bytes 
7e6bb5a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
 
resume: false
device: cuda
use_amp: false
seed: 1000
dataset_repo_id: notmahi/tutorial-ball
video_backend: pyav
training:
  offline_steps: 80000
  online_steps: 0
  online_steps_between_rollouts: 1
  online_sampling_ratio: 0.5
  online_env_seed: ???
  eval_freq: 5000
  log_freq: 250
  save_checkpoint: true
  save_freq: 5000
  num_workers: 4
  batch_size: 128
  image_transforms:
    enable: false
    max_num_transforms: 3
    random_order: false
    brightness:
      weight: 1
      min_max:
      - 0.8
      - 1.2
    contrast:
      weight: 1
      min_max:
      - 0.8
      - 1.2
    saturation:
      weight: 1
      min_max:
      - 0.5
      - 1.5
    hue:
      weight: 1
      min_max:
      - -0.05
      - 0.05
    sharpness:
      weight: 1
      min_max:
      - 0.8
      - 1.2
  lr: 1.0e-05
  lr_backbone: 1.0e-05
  weight_decay: 0.0001
  grad_clip_norm: 10
  delta_timestamps:
    action:
    - 0.0
    - 0.016666666666666666
    - 0.03333333333333333
    - 0.05
    - 0.06666666666666667
eval:
  n_episodes: 10
  batch_size: 10
  use_async_envs: false
wandb:
  enable: true
  disable_artifact: false
  project: lerobot
  notes: ''
fps: 60
env:
  name: ballgame
  task: Ballgame-v0
  state_dim: 4
  action_dim: 2
  fps: ${fps}
  episode_length: 1000
  gym:
    fps: ${fps}
    obs_type: pixels_agent_pos
    timeout: 1000
policy:
  name: act
  n_obs_steps: 1
  chunk_size: 5
  n_action_steps: 5
  input_shapes:
    observation.state:
    - ${env.state_dim}
  output_shapes:
    action:
    - ${env.action_dim}
  input_normalization_modes:
    observation.state: mean_std
  output_normalization_modes:
    action: mean_std
  vision_backbone: resnet18
  pretrained_backbone_weights: ResNet18_Weights.IMAGENET1K_V1
  replace_final_stride_with_dilation: false
  pre_norm: false
  dim_model: 512
  n_heads: 8
  dim_feedforward: 3200
  feedforward_activation: relu
  n_encoder_layers: 4
  n_decoder_layers: 2
  use_vae: true
  latent_dim: 32
  n_vae_encoder_layers: 4
  temporal_ensemble_momentum: null
  dropout: 0.1
  kl_weight: 10.0