Hugging Face's logo

KrishKrosh
/
diffusion-tshirt-folding

Safetensors
Model card Files
xet
 Community
KrishKrosh commited on
Commit
b69e7a5
·
verified ·
1 Parent(s): 711770c

Diffusion policy checkpoint at step 50000

Browse files
Files changed (7) hide show
  1. config.json +101 -0
  2. model.safetensors +3 -0
  3. policy_postprocessor.json +32 -0
  4. policy_postprocessor_step_0_unnormalizer_processor.safetensors +3 -0
  5. policy_preprocessor.json +72 -0
  6. policy_preprocessor_step_3_normalizer_processor.safetensors +3 -0
  7. train_config.json +182 -0
config.json ADDED
@@ -0,0 +1,101 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "diffusion",
3
+ "n_obs_steps": 2,
4
+ "input_features": {
5
+ "observation.state": {
6
+ "type": "STATE",
7
+ "shape": [
8
+ 14
9
+ ]
10
+ },
11
+ "observation.images.top": {
12
+ "type": "VISUAL",
13
+ "shape": [
14
+ 3,
15
+ 96,
16
+ 128
17
+ ]
18
+ },
19
+ "observation.images.left_wrist": {
20
+ "type": "VISUAL",
21
+ "shape": [
22
+ 3,
23
+ 96,
24
+ 128
25
+ ]
26
+ },
27
+ "observation.images.right_wrist": {
28
+ "type": "VISUAL",
29
+ "shape": [
30
+ 3,
31
+ 96,
32
+ 128
33
+ ]
34
+ }
35
+ },
36
+ "output_features": {
37
+ "action": {
38
+ "type": "ACTION",
39
+ "shape": [
40
+ 28
41
+ ]
42
+ }
43
+ },
44
+ "device": "cuda",
45
+ "use_amp": false,
46
+ "use_peft": false,
47
+ "push_to_hub": true,
48
+ "repo_id": "KrishKrosh/diffusion-tshirt-folding",
49
+ "private": null,
50
+ "tags": null,
51
+ "license": null,
52
+ "pretrained_path": null,
53
+ "horizon": 16,
54
+ "n_action_steps": 8,
55
+ "normalization_mapping": {
56
+ "VISUAL": "MEAN_STD",
57
+ "STATE": "MIN_MAX",
58
+ "ACTION": "MIN_MAX"
59
+ },
60
+ "drop_n_last_frames": 7,
61
+ "vision_backbone": "resnet18",
62
+ "crop_shape": [
63
+ 84,
64
+ 84
65
+ ],
66
+ "crop_is_random": true,
67
+ "pretrained_backbone_weights": null,
68
+ "use_group_norm": true,
69
+ "spatial_softmax_num_keypoints": 32,
70
+ "use_separate_rgb_encoder_per_camera": false,
71
+ "down_dims": [
72
+ 512,
73
+ 1024,
74
+ 2048
75
+ ],
76
+ "kernel_size": 5,
77
+ "n_groups": 8,
78
+ "diffusion_step_embed_dim": 128,
79
+ "use_film_scale_modulation": true,
80
+ "noise_scheduler_type": "DDPM",
81
+ "num_train_timesteps": 100,
82
+ "beta_schedule": "squaredcos_cap_v2",
83
+ "beta_start": 0.0001,
84
+ "beta_end": 0.02,
85
+ "prediction_type": "epsilon",
86
+ "clip_sample": true,
87
+ "clip_sample_range": 1.0,
88
+ "num_inference_steps": null,
89
+ "compile_model": false,
90
+ "compile_mode": "reduce-overhead",
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 sha256:05c812d604d10b67eb615398fc1f93045f2f50a34133aa5cc2f918b6e04b6400
3
+ size 1083346944
policy_postprocessor.json ADDED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "name": "policy_postprocessor",
3
+ "steps": [
4
+ {
5
+ "registry_name": "unnormalizer_processor",
6
+ "config": {
7
+ "eps": 1e-08,
8
+ "features": {
9
+ "action": {
10
+ "type": "ACTION",
11
+ "shape": [
12
+ 28
13
+ ]
14
+ }
15
+ },
16
+ "norm_map": {
17
+ "VISUAL": "MEAN_STD",
18
+ "STATE": "MIN_MAX",
19
+ "ACTION": "MIN_MAX"
20
+ }
21
+ },
22
+ "state_file": "policy_postprocessor_step_0_unnormalizer_processor.safetensors"
23
+ },
24
+ {
25
+ "registry_name": "device_processor",
26
+ "config": {
27
+ "device": "cpu",
28
+ "float_dtype": null
29
+ }
30
+ }
31
+ ]
32
+ }
policy_postprocessor_step_0_unnormalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:be756f317fa7659062d7745ea831097ebf15c7689efd89a8d75fd8dd0ee4b740
3
+ size 9920
policy_preprocessor.json ADDED
@@ -0,0 +1,72 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "name": "policy_preprocessor",
3
+ "steps": [
4
+ {
5
+ "registry_name": "rename_observations_processor",
6
+ "config": {
7
+ "rename_map": {}
8
+ }
9
+ },
10
+ {
11
+ "registry_name": "to_batch_processor",
12
+ "config": {}
13
+ },
14
+ {
15
+ "registry_name": "device_processor",
16
+ "config": {
17
+ "device": "cuda",
18
+ "float_dtype": null
19
+ }
20
+ },
21
+ {
22
+ "registry_name": "normalizer_processor",
23
+ "config": {
24
+ "eps": 1e-08,
25
+ "features": {
26
+ "observation.state": {
27
+ "type": "STATE",
28
+ "shape": [
29
+ 14
30
+ ]
31
+ },
32
+ "observation.images.top": {
33
+ "type": "VISUAL",
34
+ "shape": [
35
+ 3,
36
+ 96,
37
+ 128
38
+ ]
39
+ },
40
+ "observation.images.left_wrist": {
41
+ "type": "VISUAL",
42
+ "shape": [
43
+ 3,
44
+ 96,
45
+ 128
46
+ ]
47
+ },
48
+ "observation.images.right_wrist": {
49
+ "type": "VISUAL",
50
+ "shape": [
51
+ 3,
52
+ 96,
53
+ 128
54
+ ]
55
+ },
56
+ "action": {
57
+ "type": "ACTION",
58
+ "shape": [
59
+ 28
60
+ ]
61
+ }
62
+ },
63
+ "norm_map": {
64
+ "VISUAL": "MEAN_STD",
65
+ "STATE": "MIN_MAX",
66
+ "ACTION": "MIN_MAX"
67
+ }
68
+ },
69
+ "state_file": "policy_preprocessor_step_3_normalizer_processor.safetensors"
70
+ }
71
+ ]
72
+ }
policy_preprocessor_step_3_normalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:be756f317fa7659062d7745ea831097ebf15c7689efd89a8d75fd8dd0ee4b740
3
+ size 9920
train_config.json ADDED
@@ -0,0 +1,182 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "Gongsta/trlc_tshirt_folding",
4
+ "root": null,
5
+ "episodes": null,
6
+ "image_transforms": {
7
+ "enable": true,
8
+ "max_num_transforms": 1,
9
+ "random_order": false,
10
+ "tfs": {
11
+ "resize": {
12
+ "weight": 1.0,
13
+ "type": "Resize",
14
+ "kwargs": {
15
+ "size": [
16
+ 96,
17
+ 128
18
+ ]
19
+ }
20
+ }
21
+ }
22
+ },
23
+ "revision": null,
24
+ "use_imagenet_stats": true,
25
+ "video_backend": "torchcodec",
26
+ "streaming": false
27
+ },
28
+ "env": null,
29
+ "policy": {
30
+ "type": "diffusion",
31
+ "n_obs_steps": 2,
32
+ "input_features": {
33
+ "observation.state": {
34
+ "type": "STATE",
35
+ "shape": [
36
+ 14
37
+ ]
38
+ },
39
+ "observation.images.top": {
40
+ "type": "VISUAL",
41
+ "shape": [
42
+ 3,
43
+ 96,
44
+ 128
45
+ ]
46
+ },
47
+ "observation.images.left_wrist": {
48
+ "type": "VISUAL",
49
+ "shape": [
50
+ 3,
51
+ 96,
52
+ 128
53
+ ]
54
+ },
55
+ "observation.images.right_wrist": {
56
+ "type": "VISUAL",
57
+ "shape": [
58
+ 3,
59
+ 96,
60
+ 128
61
+ ]
62
+ }
63
+ },
64
+ "output_features": {
65
+ "action": {
66
+ "type": "ACTION",
67
+ "shape": [
68
+ 28
69
+ ]
70
+ }
71
+ },
72
+ "device": "cuda",
73
+ "use_amp": false,
74
+ "use_peft": false,
75
+ "push_to_hub": true,
76
+ "repo_id": "KrishKrosh/diffusion-tshirt-folding",
77
+ "private": null,
78
+ "tags": null,
79
+ "license": null,
80
+ "pretrained_path": null,
81
+ "horizon": 16,
82
+ "n_action_steps": 8,
83
+ "normalization_mapping": {
84
+ "VISUAL": "MEAN_STD",
85
+ "STATE": "MIN_MAX",
86
+ "ACTION": "MIN_MAX"
87
+ },
88
+ "drop_n_last_frames": 7,
89
+ "vision_backbone": "resnet18",
90
+ "crop_shape": [
91
+ 84,
92
+ 84
93
+ ],
94
+ "crop_is_random": true,
95
+ "pretrained_backbone_weights": null,
96
+ "use_group_norm": true,
97
+ "spatial_softmax_num_keypoints": 32,
98
+ "use_separate_rgb_encoder_per_camera": false,
99
+ "down_dims": [
100
+ 512,
101
+ 1024,
102
+ 2048
103
+ ],
104
+ "kernel_size": 5,
105
+ "n_groups": 8,
106
+ "diffusion_step_embed_dim": 128,
107
+ "use_film_scale_modulation": true,
108
+ "noise_scheduler_type": "DDPM",
109
+ "num_train_timesteps": 100,
110
+ "beta_schedule": "squaredcos_cap_v2",
111
+ "beta_start": 0.0001,
112
+ "beta_end": 0.02,
113
+ "prediction_type": "epsilon",
114
+ "clip_sample": true,
115
+ "clip_sample_range": 1.0,
116
+ "num_inference_steps": null,
117
+ "compile_model": false,
118
+ "compile_mode": "reduce-overhead",
119
+ "do_mask_loss_for_padding": false,
120
+ "optimizer_lr": 0.0001,
121
+ "optimizer_betas": [
122
+ 0.95,
123
+ 0.999
124
+ ],
125
+ "optimizer_eps": 1e-08,
126
+ "optimizer_weight_decay": 1e-06,
127
+ "scheduler_name": "cosine",
128
+ "scheduler_warmup_steps": 500
129
+ },
130
+ "output_dir": "/outputs/diffusion_tshirt_folding",
131
+ "job_name": "diffusion_tshirt_folding",
132
+ "resume": false,
133
+ "seed": 1000,
134
+ "num_workers": 16,
135
+ "batch_size": 128,
136
+ "steps": 100000,
137
+ "eval_freq": -1,
138
+ "log_freq": 100,
139
+ "tolerance_s": 0.0001,
140
+ "save_checkpoint": true,
141
+ "save_freq": 10000,
142
+ "use_policy_training_preset": true,
143
+ "optimizer": {
144
+ "type": "adam",
145
+ "lr": 0.0001,
146
+ "weight_decay": 1e-06,
147
+ "grad_clip_norm": 10.0,
148
+ "betas": [
149
+ 0.95,
150
+ 0.999
151
+ ],
152
+ "eps": 1e-08
153
+ },
154
+ "scheduler": {
155
+ "type": "diffuser",
156
+ "num_warmup_steps": 500,
157
+ "name": "cosine"
158
+ },
159
+ "eval": {
160
+ "n_episodes": 50,
161
+ "batch_size": 50,
162
+ "use_async_envs": false
163
+ },
164
+ "wandb": {
165
+ "enable": true,
166
+ "disable_artifact": false,
167
+ "project": "diffusion-tshirt-folding",
168
+ "entity": null,
169
+ "notes": null,
170
+ "run_id": "hhd8w3ou",
171
+ "mode": null
172
+ },
173
+ "peft": null,
174
+ "use_rabc": false,
175
+ "rabc_progress_path": null,
176
+ "rabc_kappa": 0.01,
177
+ "rabc_epsilon": 1e-06,
178
+ "rabc_head_mode": "sparse",
179
+ "rename_map": {},
180
+ "enable_profiling": false,
181
+ "checkpoint_path": null
182
+ }