Hugging Face's logo

DAVIAN-Robotics
/
DiscoDemo-Stage1_RL-PnP_Banana

Reinforcement Learning
robotics
discodemo
franka
Model card Files
xet
 Community
DiscoDemo-Stage1_RL-PnP_Banana
File size: 2,869 Bytes
a3bfe84
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
project_name: DiscoDemo
entity_name: null
group_name: pnp_banana
exp_name: discodemo-pnp_banana
seed: 0
logger_type: wandb
save_path: libs/FlashSAC/models/pnp_banana/discodemo-pnp_banana/seed0-TIMESTAMP
agent_load_path: null
num_env_steps: 50000000
num_train_envs: 2048
updates_per_interaction_step: 64
gamma: 0.99
n_step: 8
env:
  workcell: libs/real2sim/assets/workcell.json
  task: libs/real2sim/real2sim/tasks/pnp_banana.py
  env_version: fr3-pnp_banana
  gripper_lock: null
success_hold_steps: 20
agent:
  agent_type: flashSAC
  seed: 0
  device_type: cuda
  buffer_max_length: 100000000
  buffer_min_length: 5000
  buffer_device_type: cuda
  sample_batch_size: 2048
  normalize_reward: true
  normalized_G_max: 5.0
  asymmetric_observation: false
  learning_rate_init: 0.0003
  learning_rate_peak: 0.0003
  learning_rate_end: 0.00015
  learning_rate_warmup_rate: 1.0e-06
  learning_rate_warmup_step: 50
  learning_rate_decay_rate: 1.0
  learning_rate_decay_step: 50000000
  actor_num_blocks: 2
  actor_hidden_dim: 128
  actor_bc_alpha: 0
  actor_noise_zeta_mu: 2.0
  actor_noise_zeta_max: 16
  actor_update_period: 2
  critic_num_blocks: 2
  critic_hidden_dim: 256
  critic_num_bins: 101
  critic_min_v: -5
  critic_max_v: 5
  critic_target_update_tau: 0.01
  temp_initial_value: 0.01
  temp_target_sigma: 0.15
  temp_target_entropy: null
  gamma: 0.99
  n_step: 8
  use_compile: true
  compile_mode: auto
  use_amp: true
  load_optimizer: true
  load_reward_normalizer: true
rfcl:
  demo_ratio: 0.5
  truncate_after_success: 50
  reverse:
    reverse_step_size: 48
    advance_threshold: 0.9
    frontier_window: 10
    minimum_episode_steps: 30
  solved_frac_threshold: 0.5
  demo_path: data/demo_banks/pnp_banana/human_demos.h5
  init_pose: box:0.4125,0.6375,-0.15,0.15,180,0.20
eval_worker:
  num_envs: 50
skill:
  phi_hidden:
  - 256
  - 256
  phi_lr: 0.0001
  dual_slack: 0.001
  lam_init: 1.0
  alpha: 0.3
  z_dim: 3
safety_penalty:
  illegal_contact:
    weight: 0.0001
    sat: 100
    robot_bodies:
    - panda_link5
    - panda_link6
    - panda_link7
    - panda_hand
    - panda_leftfinger
    - panda_rightfinger
    scene_bodies:
    - table
    - bowl
    - banana
    allowed_pairs:
    - - panda_leftfinger
      - banana
    - - panda_rightfinger
      - banana
obs_bounds:
  object_z_m:
  - 0.0
  - 0.6
  eef_box_m:
    x:
    - 0.2
    - 0.85
    'y':
    - -0.45
    - 0.45
    z:
    - 0.0
    - 0.7
  object_xy:
  - - 0.2
    - 0.85
  - - -0.4
    - 0.4
task_name: pnp_banana
jerk_limited_action:
  max_joint_velocity:
  - 0.22215
  - 0.51715
  - 0.19244
  - 0.66218
  - 0.31884
  - 0.41767
  - 0.40726
  max_joint_acceleration:
  - 0.46616
  - 0.95012
  - 0.46407
  - 1.28545
  - 1.25775
  - 1.18572
  - 1.33469
  max_joint_jerk:
  - 2.65668
  - 4.40304
  - 2.82091
  - 5.53682
  - 9.90014
  - 6.67391
  - 10.55754
method_name: discodemo