Hugging Face's logo

kimyg119
/
pnp_sim_only

Robotics
LeRobot
Safetensors
sim_only
DP
diffusion
ai_worker
Model card Files
xet
 Community
kimyg119 commited on
Commit
e887be6
·
verified ·
1 Parent(s): d8b372b

Upload policy weights, train config and readme

Browse files
Files changed (4) hide show
  1. README.md +65 -0
  2. config.json +101 -0
  3. model.safetensors +3 -0
  4. train_config.json +211 -0
README.md ADDED
@@ -0,0 +1,65 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ datasets: kimyg119/251105_soup_can_sim_50_lighting_augmented_trellis
3
+ library_name: lerobot
4
+ license: apache-2.0
5
+ model_name: diffusion
6
+ pipeline_tag: robotics
7
+ tags:
8
+ - sim_only
9
+ - DP
10
+ - diffusion
11
+ - lerobot
12
+ - robotics
13
+ - ai_worker
14
+ ---
15
+
16
+ # Model Card for diffusion
17
+
18
+ <!-- Provide a quick summary of what the model is/does. -->
19
+
20
+
21
+ [Diffusion Policy](https://huggingface.co/papers/2303.04137) treats visuomotor control as a generative diffusion process, producing smooth, multi-step action trajectories that excel at contact-rich manipulation.
22
+
23
+
24
+ This policy has been trained and pushed to the Hub using [LeRobot](https://github.com/huggingface/lerobot).
25
+ See the full documentation at [LeRobot Docs](https://huggingface.co/docs/lerobot/index).
26
+
27
+ ---
28
+
29
+ ## How to Get Started with the Model
30
+
31
+ For a complete walkthrough, see the [training guide](https://huggingface.co/docs/lerobot/il_robots#train-a-policy).
32
+ Below is the short version on how to train and run inference/eval:
33
+
34
+ ### Train from scratch
35
+
36
+ ```bash
37
+ python -m lerobot.scripts.train \
38
+ --dataset.repo_id=${HF_USER}/<dataset> \
39
+ --policy.type=act \
40
+ --output_dir=outputs/train/<desired_policy_repo_id> \
41
+ --job_name=lerobot_training \
42
+ --policy.device=cuda \
43
+ --policy.repo_id=${HF_USER}/<desired_policy_repo_id>
44
+ --wandb.enable=true
45
+ ```
46
+
47
+ _Writes checkpoints to `outputs/train/<desired_policy_repo_id>/checkpoints/`._
48
+
49
+ ### Evaluate the policy/run inference
50
+
51
+ ```bash
52
+ python -m lerobot.record \
53
+ --robot.type=so100_follower \
54
+ --dataset.repo_id=<hf_user>/eval_<dataset> \
55
+ --policy.path=<hf_user>/<desired_policy_repo_id> \
56
+ --episodes=10
57
+ ```
58
+
59
+ Prefix the dataset repo with **eval\_** and supply `--policy.path` pointing to a local or hub checkpoint.
60
+
61
+ ---
62
+
63
+ ## Model Details
64
+
65
+ - **License:** apache-2.0
config.json ADDED
@@ -0,0 +1,101 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "diffusion",
3
+ "n_obs_steps": 2,
4
+ "normalization_mapping": {
5
+ "VISUAL": "MEAN_STD",
6
+ "STATE": "MIN_MAX",
7
+ "ACTION": "MIN_MAX"
8
+ },
9
+ "input_features": {
10
+ "observation.state": {
11
+ "type": "STATE",
12
+ "shape": [
13
+ 16
14
+ ]
15
+ },
16
+ "observation.images.cam_head": {
17
+ "type": "VISUAL",
18
+ "shape": [
19
+ 3,
20
+ 480,
21
+ 640
22
+ ]
23
+ },
24
+ "observation.images.cam_wrist_left": {
25
+ "type": "VISUAL",
26
+ "shape": [
27
+ 3,
28
+ 480,
29
+ 640
30
+ ]
31
+ },
32
+ "observation.images.cam_wrist_right": {
33
+ "type": "VISUAL",
34
+ "shape": [
35
+ 3,
36
+ 480,
37
+ 640
38
+ ]
39
+ }
40
+ },
41
+ "output_features": {
42
+ "action": {
43
+ "type": "ACTION",
44
+ "shape": [
45
+ 16
46
+ ]
47
+ }
48
+ },
49
+ "device": "cuda",
50
+ "use_amp": false,
51
+ "push_to_hub": true,
52
+ "repo_id": "kimyg119/pnp_sim_only",
53
+ "private": null,
54
+ "tags": [
55
+ "DP",
56
+ "ai_worker",
57
+ "sim_only"
58
+ ],
59
+ "license": null,
60
+ "horizon": 16,
61
+ "n_action_steps": 8,
62
+ "drop_n_last_frames": 7,
63
+ "vision_backbone": "resnet18",
64
+ "crop_shape": [
65
+ 84,
66
+ 84
67
+ ],
68
+ "crop_is_random": true,
69
+ "pretrained_backbone_weights": null,
70
+ "use_group_norm": true,
71
+ "spatial_softmax_num_keypoints": 32,
72
+ "use_separate_rgb_encoder_per_camera": false,
73
+ "down_dims": [
74
+ 512,
75
+ 1024,
76
+ 2048
77
+ ],
78
+ "kernel_size": 5,
79
+ "n_groups": 8,
80
+ "diffusion_step_embed_dim": 128,
81
+ "use_film_scale_modulation": true,
82
+ "noise_scheduler_type": "DDPM",
83
+ "num_train_timesteps": 100,
84
+ "beta_schedule": "squaredcos_cap_v2",
85
+ "beta_start": 0.0001,
86
+ "beta_end": 0.02,
87
+ "prediction_type": "epsilon",
88
+ "clip_sample": true,
89
+ "clip_sample_range": 1.0,
90
+ "num_inference_steps": null,
91
+ "do_mask_loss_for_padding": false,
92
+ "optimizer_lr": 0.0001,
93
+ "optimizer_betas": [
94
+ 0.95,
95
+ 0.999
96
+ ],
97
+ "optimizer_eps": 1e-08,
98
+ "optimizer_weight_decay": 1e-06,
99
+ "scheduler_name": "cosine",
100
+ "scheduler_warmup_steps": 500
101
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:be3714aaefa38810684deacd40faa9773f585e63bf8a4f4f84435b1de7a7c2ce
3
+ size 1083635528
train_config.json ADDED
@@ -0,0 +1,211 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "kimyg119/251105_soup_can_sim_50_lighting_augmented_trellis",
4
+ "root": null,
5
+ "episodes": null,
6
+ "image_transforms": {
7
+ "enable": false,
8
+ "max_num_transforms": 3,
9
+ "random_order": false,
10
+ "tfs": {
11
+ "brightness": {
12
+ "weight": 1.0,
13
+ "type": "ColorJitter",
14
+ "kwargs": {
15
+ "brightness": [
16
+ 0.8,
17
+ 1.2
18
+ ]
19
+ }
20
+ },
21
+ "contrast": {
22
+ "weight": 1.0,
23
+ "type": "ColorJitter",
24
+ "kwargs": {
25
+ "contrast": [
26
+ 0.8,
27
+ 1.2
28
+ ]
29
+ }
30
+ },
31
+ "saturation": {
32
+ "weight": 1.0,
33
+ "type": "ColorJitter",
34
+ "kwargs": {
35
+ "saturation": [
36
+ 0.5,
37
+ 1.5
38
+ ]
39
+ }
40
+ },
41
+ "hue": {
42
+ "weight": 1.0,
43
+ "type": "ColorJitter",
44
+ "kwargs": {
45
+ "hue": [
46
+ -0.05,
47
+ 0.05
48
+ ]
49
+ }
50
+ },
51
+ "sharpness": {
52
+ "weight": 1.0,
53
+ "type": "SharpnessJitter",
54
+ "kwargs": {
55
+ "sharpness": [
56
+ 0.5,
57
+ 1.5
58
+ ]
59
+ }
60
+ }
61
+ }
62
+ },
63
+ "revision": null,
64
+ "use_imagenet_stats": true,
65
+ "video_backend": "torchcodec"
66
+ },
67
+ "env": null,
68
+ "policy": {
69
+ "type": "diffusion",
70
+ "n_obs_steps": 2,
71
+ "normalization_mapping": {
72
+ "VISUAL": "MEAN_STD",
73
+ "STATE": "MIN_MAX",
74
+ "ACTION": "MIN_MAX"
75
+ },
76
+ "input_features": {
77
+ "observation.state": {
78
+ "type": "STATE",
79
+ "shape": [
80
+ 16
81
+ ]
82
+ },
83
+ "observation.images.cam_head": {
84
+ "type": "VISUAL",
85
+ "shape": [
86
+ 3,
87
+ 480,
88
+ 640
89
+ ]
90
+ },
91
+ "observation.images.cam_wrist_left": {
92
+ "type": "VISUAL",
93
+ "shape": [
94
+ 3,
95
+ 480,
96
+ 640
97
+ ]
98
+ },
99
+ "observation.images.cam_wrist_right": {
100
+ "type": "VISUAL",
101
+ "shape": [
102
+ 3,
103
+ 480,
104
+ 640
105
+ ]
106
+ }
107
+ },
108
+ "output_features": {
109
+ "action": {
110
+ "type": "ACTION",
111
+ "shape": [
112
+ 16
113
+ ]
114
+ }
115
+ },
116
+ "device": "cuda",
117
+ "use_amp": false,
118
+ "push_to_hub": true,
119
+ "repo_id": "kimyg119/pnp_sim_only",
120
+ "private": null,
121
+ "tags": [
122
+ "DP",
123
+ "ai_worker",
124
+ "sim_only"
125
+ ],
126
+ "license": null,
127
+ "horizon": 16,
128
+ "n_action_steps": 8,
129
+ "drop_n_last_frames": 7,
130
+ "vision_backbone": "resnet18",
131
+ "crop_shape": [
132
+ 84,
133
+ 84
134
+ ],
135
+ "crop_is_random": true,
136
+ "pretrained_backbone_weights": null,
137
+ "use_group_norm": true,
138
+ "spatial_softmax_num_keypoints": 32,
139
+ "use_separate_rgb_encoder_per_camera": false,
140
+ "down_dims": [
141
+ 512,
142
+ 1024,
143
+ 2048
144
+ ],
145
+ "kernel_size": 5,
146
+ "n_groups": 8,
147
+ "diffusion_step_embed_dim": 128,
148
+ "use_film_scale_modulation": true,
149
+ "noise_scheduler_type": "DDPM",
150
+ "num_train_timesteps": 100,
151
+ "beta_schedule": "squaredcos_cap_v2",
152
+ "beta_start": 0.0001,
153
+ "beta_end": 0.02,
154
+ "prediction_type": "epsilon",
155
+ "clip_sample": true,
156
+ "clip_sample_range": 1.0,
157
+ "num_inference_steps": null,
158
+ "do_mask_loss_for_padding": false,
159
+ "optimizer_lr": 0.0001,
160
+ "optimizer_betas": [
161
+ 0.95,
162
+ 0.999
163
+ ],
164
+ "optimizer_eps": 1e-08,
165
+ "optimizer_weight_decay": 1e-06,
166
+ "scheduler_name": "cosine",
167
+ "scheduler_warmup_steps": 500
168
+ },
169
+ "output_dir": "outputs/train/pnp_sim_only",
170
+ "job_name": "diffusion",
171
+ "resume": false,
172
+ "seed": 1000,
173
+ "num_workers": 4,
174
+ "batch_size": 64,
175
+ "steps": 100000,
176
+ "eval_freq": -1,
177
+ "log_freq": 1000,
178
+ "save_checkpoint": true,
179
+ "save_freq": 20000,
180
+ "use_policy_training_preset": true,
181
+ "optimizer": {
182
+ "type": "adam",
183
+ "lr": 0.0001,
184
+ "weight_decay": 1e-06,
185
+ "grad_clip_norm": 10.0,
186
+ "betas": [
187
+ 0.95,
188
+ 0.999
189
+ ],
190
+ "eps": 1e-08
191
+ },
192
+ "scheduler": {
193
+ "type": "diffuser",
194
+ "num_warmup_steps": 500,
195
+ "name": "cosine"
196
+ },
197
+ "eval": {
198
+ "n_episodes": 50,
199
+ "batch_size": 50,
200
+ "use_async_envs": false
201
+ },
202
+ "wandb": {
203
+ "enable": false,
204
+ "disable_artifact": false,
205
+ "project": "lerobot",
206
+ "entity": null,
207
+ "notes": null,
208
+ "run_id": null,
209
+ "mode": null
210
+ }
211
+ }