Hugging Face's logo

hqfang
/
dit

Safetensors
Model card Files
xet
Community
hqfang commited on
Commit
df70aff
·
verified ·
1 Parent(s): ec32946

Upload folder using huggingface_hub

Browse files
Files changed (8) hide show
  1. pretrained_model/config.json +93 -0
  2. pretrained_model/model.safetensors +3 -0
  3. pretrained_model/train_config.json +223 -0
  4. training_state/optimizer_param_groups.json +126 -0
  5. training_state/optimizer_state.safetensors +3 -0
  6. training_state/rng_state.safetensors +3 -0
  7. training_state/scheduler_state.json +18 -0
  8. training_state/training_step.json +3 -0
pretrained_model/config.json ADDED
@@ -0,0 +1,93 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "n_obs_steps": 2,
3
+ "normalization_mapping": {
4
+ "VISUAL": "MEAN_STD",
5
+ "STATE": "MIN_MAX",
6
+ "ACTION": "MIN_MAX"
7
+ },
8
+ "input_features": {
9
+ "observation.images.first_view": {
10
+ "type": "VISUAL",
11
+ "shape": [
12
+ 3,
13
+ 480,
14
+ 640
15
+ ]
16
+ },
17
+ "observation.images.second_view": {
18
+ "type": "VISUAL",
19
+ "shape": [
20
+ 3,
21
+ 480,
22
+ 640
23
+ ]
24
+ },
25
+ "observation.images.wrist_image": {
26
+ "type": "VISUAL",
27
+ "shape": [
28
+ 3,
29
+ 480,
30
+ 640
31
+ ]
32
+ },
33
+ "observation.state": {
34
+ "type": "STATE",
35
+ "shape": [
36
+ 7
37
+ ]
38
+ }
39
+ },
40
+ "output_features": {
41
+ "action": {
42
+ "type": "ACTION",
43
+ "shape": [
44
+ 7
45
+ ]
46
+ }
47
+ },
48
+ "device": "cuda",
49
+ "use_amp": false,
50
+ "horizon": 16,
51
+ "n_action_steps": 8,
52
+ "drop_n_last_frames": 7,
53
+ "vision_backbone": "resnet50",
54
+ "crop_shape": null,
55
+ "pretrained_backbone_weights": null,
56
+ "use_group_norm": true,
57
+ "spatial_softmax_num_keypoints": 32,
58
+ "use_separate_rgb_encoder_per_camera": false,
59
+ "denoising_model_type": "transformer",
60
+ "diffusion_step_embed_dim": 128,
61
+ "down_dims": [
62
+ 512,
63
+ 1024,
64
+ 2048
65
+ ],
66
+ "kernel_size": 5,
67
+ "n_groups": 8,
68
+ "use_film_scale_modulation": true,
69
+ "embed_dim": 768,
70
+ "num_heads": 12,
71
+ "depth": 8,
72
+ "mlp_ratio": 4.0,
73
+ "noise_scheduler_type": "DDIM",
74
+ "num_train_timesteps": 100,
75
+ "beta_schedule": "squaredcos_cap_v2",
76
+ "beta_start": 0.0001,
77
+ "beta_end": 0.02,
78
+ "prediction_type": "epsilon",
79
+ "clip_sample": true,
80
+ "clip_sample_range": 1.0,
81
+ "num_inference_steps": null,
82
+ "do_mask_loss_for_padding": false,
83
+ "optimizer_lr": 0.0001,
84
+ "optimizer_betas": [
85
+ 0.95,
86
+ 0.999
87
+ ],
88
+ "optimizer_eps": 1e-08,
89
+ "optimizer_weight_decay": 1e-06,
90
+ "optimizer_grad_clip_norm": 1.0,
91
+ "scheduler_name": "cosine",
92
+ "scheduler_warmup_steps": 500
93
+ }
pretrained_model/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:78dc4f3fbbb3e4f6a0a4fc125f086145dad3b32b6d1f9bc7aa6d923fbf1efb17
3
+ size 441367612
pretrained_model/train_config.json ADDED
@@ -0,0 +1,223 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "/net/nfs/prior/shuol/molmoact_lerobot/molmoact_dataset_tabletop/train/",
4
+ "root": null,
5
+ "episodes": null,
6
+ "image_transforms": {
7
+ "enable": true,
8
+ "max_num_transforms": 3,
9
+ "random_order": false,
10
+ "tfs": {
11
+ "resize": {
12
+ "weight": 1.0,
13
+ "type": "Resize",
14
+ "kwargs": {
15
+ "size": [
16
+ 256,
17
+ 342
18
+ ]
19
+ }
20
+ },
21
+ "random_crop": {
22
+ "weight": 1.0,
23
+ "type": "RandomCrop",
24
+ "kwargs": {
25
+ "size": [
26
+ 224,
27
+ 224
28
+ ]
29
+ }
30
+ },
31
+ "brightness": {
32
+ "weight": 1.0,
33
+ "type": "ColorJitter",
34
+ "kwargs": {
35
+ "brightness": [
36
+ 0.8,
37
+ 1.2
38
+ ]
39
+ }
40
+ },
41
+ "contrast": {
42
+ "weight": 1.0,
43
+ "type": "ColorJitter",
44
+ "kwargs": {
45
+ "contrast": [
46
+ 0.8,
47
+ 1.2
48
+ ]
49
+ }
50
+ },
51
+ "saturation": {
52
+ "weight": 1.0,
53
+ "type": "ColorJitter",
54
+ "kwargs": {
55
+ "saturation": [
56
+ 0.5,
57
+ 1.5
58
+ ]
59
+ }
60
+ },
61
+ "hue": {
62
+ "weight": 1.0,
63
+ "type": "ColorJitter",
64
+ "kwargs": {
65
+ "hue": [
66
+ -0.05,
67
+ 0.05
68
+ ]
69
+ }
70
+ },
71
+ "sharpness": {
72
+ "weight": 1.0,
73
+ "type": "SharpnessJitter",
74
+ "kwargs": {
75
+ "sharpness": [
76
+ 0.5,
77
+ 1.5
78
+ ]
79
+ }
80
+ }
81
+ }
82
+ },
83
+ "revision": null,
84
+ "use_imagenet_stats": true,
85
+ "video_backend": "pyav"
86
+ },
87
+ "policy": {
88
+ "type": "diffusion",
89
+ "n_obs_steps": 2,
90
+ "normalization_mapping": {
91
+ "VISUAL": "MEAN_STD",
92
+ "STATE": "MIN_MAX",
93
+ "ACTION": "MIN_MAX"
94
+ },
95
+ "input_features": {
96
+ "observation.images.first_view": {
97
+ "type": "VISUAL",
98
+ "shape": [
99
+ 3,
100
+ 480,
101
+ 640
102
+ ]
103
+ },
104
+ "observation.images.second_view": {
105
+ "type": "VISUAL",
106
+ "shape": [
107
+ 3,
108
+ 480,
109
+ 640
110
+ ]
111
+ },
112
+ "observation.images.wrist_image": {
113
+ "type": "VISUAL",
114
+ "shape": [
115
+ 3,
116
+ 480,
117
+ 640
118
+ ]
119
+ },
120
+ "observation.state": {
121
+ "type": "STATE",
122
+ "shape": [
123
+ 7
124
+ ]
125
+ }
126
+ },
127
+ "output_features": {
128
+ "action": {
129
+ "type": "ACTION",
130
+ "shape": [
131
+ 7
132
+ ]
133
+ }
134
+ },
135
+ "device": "cuda",
136
+ "use_amp": false,
137
+ "horizon": 16,
138
+ "n_action_steps": 8,
139
+ "drop_n_last_frames": 7,
140
+ "vision_backbone": "resnet50",
141
+ "crop_shape": null,
142
+ "pretrained_backbone_weights": null,
143
+ "use_group_norm": true,
144
+ "spatial_softmax_num_keypoints": 32,
145
+ "use_separate_rgb_encoder_per_camera": false,
146
+ "denoising_model_type": "transformer",
147
+ "diffusion_step_embed_dim": 128,
148
+ "down_dims": [
149
+ 512,
150
+ 1024,
151
+ 2048
152
+ ],
153
+ "kernel_size": 5,
154
+ "n_groups": 8,
155
+ "use_film_scale_modulation": true,
156
+ "embed_dim": 768,
157
+ "num_heads": 12,
158
+ "depth": 8,
159
+ "mlp_ratio": 4.0,
160
+ "noise_scheduler_type": "DDIM",
161
+ "num_train_timesteps": 100,
162
+ "beta_schedule": "squaredcos_cap_v2",
163
+ "beta_start": 0.0001,
164
+ "beta_end": 0.02,
165
+ "prediction_type": "epsilon",
166
+ "clip_sample": true,
167
+ "clip_sample_range": 1.0,
168
+ "num_inference_steps": null,
169
+ "do_mask_loss_for_padding": false,
170
+ "optimizer_lr": 0.0001,
171
+ "optimizer_betas": [
172
+ 0.95,
173
+ 0.999
174
+ ],
175
+ "optimizer_eps": 1e-08,
176
+ "optimizer_weight_decay": 1e-06,
177
+ "optimizer_grad_clip_norm": 1.0,
178
+ "scheduler_name": "cosine",
179
+ "scheduler_warmup_steps": 500
180
+ },
181
+ "output_dir": "/app/output/2025-10-24/22-46-38_diffusion",
182
+ "job_name": "diffusion",
183
+ "resume": false,
184
+ "seed": 42,
185
+ "num_workers": 16,
186
+ "batch_size": 64,
187
+ "steps": 200000,
188
+ "eval_freq": 20000,
189
+ "log_freq": 200,
190
+ "save_checkpoint": true,
191
+ "save_freq": 20000,
192
+ "use_policy_training_preset": true,
193
+ "optimizer": {
194
+ "type": "adamw",
195
+ "lr": 0.0001,
196
+ "weight_decay": 1e-06,
197
+ "grad_clip_norm": 1.0,
198
+ "betas": [
199
+ 0.95,
200
+ 0.999
201
+ ],
202
+ "eps": 1e-08
203
+ },
204
+ "scheduler": {
205
+ "type": "diffuser",
206
+ "num_warmup_steps": 500,
207
+ "name": "cosine"
208
+ },
209
+ "eval": {
210
+ "n_episodes": 50,
211
+ "batch_size": 50,
212
+ "use_async_envs": false
213
+ },
214
+ "wandb": {
215
+ "enable": true,
216
+ "disable_artifact": false,
217
+ "project": "steerdp",
218
+ "entity": null,
219
+ "notes": null,
220
+ "run_id": null,
221
+ "mode": null
222
+ }
223
+ }
training_state/optimizer_param_groups.json ADDED
@@ -0,0 +1,126 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "weight_decay": 0.001,
4
+ "lr": 0.0,
5
+ "betas": [
6
+ 0.95,
7
+ 0.999
8
+ ],
9
+ "eps": 1e-08,
10
+ "amsgrad": false,
11
+ "foreach": null,
12
+ "maximize": false,
13
+ "capturable": false,
14
+ "differentiable": false,
15
+ "fused": null,
16
+ "initial_lr": 0.0001,
17
+ "params": [
18
+ 0,
19
+ 1,
20
+ 2,
21
+ 3,
22
+ 4,
23
+ 5,
24
+ 6,
25
+ 7,
26
+ 8,
27
+ 9,
28
+ 10,
29
+ 11,
30
+ 12,
31
+ 13,
32
+ 14,
33
+ 15,
34
+ 16,
35
+ 17,
36
+ 18,
37
+ 19,
38
+ 20,
39
+ 21,
40
+ 22,
41
+ 23,
42
+ 24,
43
+ 25,
44
+ 26,
45
+ 27,
46
+ 28,
47
+ 29,
48
+ 30,
49
+ 31,
50
+ 32,
51
+ 33,
52
+ 34,
53
+ 35,
54
+ 36,
55
+ 37,
56
+ 38,
57
+ 39,
58
+ 40,
59
+ 41,
60
+ 42,
61
+ 43
62
+ ]
63
+ },
64
+ {
65
+ "weight_decay": 0.0,
66
+ "lr": 0.0,
67
+ "betas": [
68
+ 0.95,
69
+ 0.999
70
+ ],
71
+ "eps": 1e-08,
72
+ "amsgrad": false,
73
+ "foreach": null,
74
+ "maximize": false,
75
+ "capturable": false,
76
+ "differentiable": false,
77
+ "fused": null,
78
+ "initial_lr": 0.0001,
79
+ "params": [
80
+ 44,
81
+ 45,
82
+ 46,
83
+ 47,
84
+ 48,
85
+ 49,
86
+ 50,
87
+ 51,
88
+ 52,
89
+ 53,
90
+ 54,
91
+ 55,
92
+ 56,
93
+ 57,
94
+ 58,
95
+ 59,
96
+ 60,
97
+ 61,
98
+ 62,
99
+ 63,
100
+ 64,
101
+ 65,
102
+ 66,
103
+ 67,
104
+ 68,
105
+ 69,
106
+ 70,
107
+ 71,
108
+ 72,
109
+ 73,
110
+ 74,
111
+ 75,
112
+ 76,
113
+ 77,
114
+ 78,
115
+ 79,
116
+ 80,
117
+ 81,
118
+ 82,
119
+ 83,
120
+ 84,
121
+ 85,
122
+ 86,
123
+ 87
124
+ ]
125
+ }
126
+ ]
training_state/optimizer_state.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f532c19a05633bfac22f6763d731e5f1929fe54e9110f5ed1cbabf53c860d177
3
+ size 693017288
training_state/rng_state.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a363e3a7a4cf6905878a57ca6d32970dbd89bce78bb0c9518e9d912ef1692c97
3
+ size 15708
training_state/scheduler_state.json ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "base_lrs": [
3
+ 0.0001,
4
+ 0.0001
5
+ ],
6
+ "last_epoch": 200000,
7
+ "verbose": false,
8
+ "_step_count": 200001,
9
+ "_get_lr_called_within_step": false,
10
+ "_last_lr": [
11
+ 0.0,
12
+ 0.0
13
+ ],
14
+ "lr_lambdas": [
15
+ null,
16
+ null
17
+ ]
18
+ }
training_state/training_step.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ {
2
+ "step": 200000
3
+ }