Hugging Face's logo

continuallearning
/
dit_flow_mt_franka_pretrain

Safetensors
Model card Files
xet
 Community
continuallearning commited on
Commit
ed9018b
·
verified ·
1 Parent(s): a0b75e0

Upload folder using huggingface_hub

Browse files
Files changed (3) hide show
  1. config.json +102 -0
  2. model.safetensors +3 -0
  3. train_config.json +236 -0
config.json ADDED
@@ -0,0 +1,102 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "ditflow_mt",
3
+ "n_obs_steps": 2,
4
+ "input_features": {
5
+ "observation.images.primary": {
6
+ "type": "VISUAL",
7
+ "shape": [
8
+ 3,
9
+ 256,
10
+ 256
11
+ ]
12
+ },
13
+ "observation.images.wrist": {
14
+ "type": "VISUAL",
15
+ "shape": [
16
+ 3,
17
+ 256,
18
+ 256
19
+ ]
20
+ },
21
+ "observation.state.cartesian": {
22
+ "type": "STATE",
23
+ "shape": [
24
+ 6
25
+ ]
26
+ },
27
+ "observation.state.gripper": {
28
+ "type": "STATE",
29
+ "shape": [
30
+ 1
31
+ ]
32
+ },
33
+ "observation.state.joints": {
34
+ "type": "STATE",
35
+ "shape": [
36
+ 7
37
+ ]
38
+ },
39
+ "observation.state.target": {
40
+ "type": "STATE",
41
+ "shape": [
42
+ 6
43
+ ]
44
+ },
45
+ "observation.state": {
46
+ "type": "STATE",
47
+ "shape": [
48
+ 20
49
+ ]
50
+ }
51
+ },
52
+ "output_features": {
53
+ "action": {
54
+ "type": "ACTION",
55
+ "shape": [
56
+ 7
57
+ ]
58
+ }
59
+ },
60
+ "device": "cuda",
61
+ "use_amp": false,
62
+ "use_peft": false,
63
+ "push_to_hub": false,
64
+ "repo_id": null,
65
+ "private": null,
66
+ "tags": null,
67
+ "license": null,
68
+ "pretrained_path": null,
69
+ "horizon": 16,
70
+ "n_action_steps": 8,
71
+ "normalization_mapping": {
72
+ "VISUAL": "MEAN_STD",
73
+ "STATE": "MIN_MAX",
74
+ "ACTION": "MIN_MAX"
75
+ },
76
+ "drop_n_last_frames": 7,
77
+ "tokenizer_max_length": 48,
78
+ "language_model_name": "openai/clip-vit-base-patch32",
79
+ "freeze_language_pretrained": true,
80
+ "vit_name": "facebook/dinov2-base",
81
+ "frequency_embedding_dim": 256,
82
+ "hidden_dim": 512,
83
+ "num_blocks": 6,
84
+ "num_heads": 16,
85
+ "dropout": 0.1,
86
+ "dim_feedforward": 4096,
87
+ "activation": "gelu",
88
+ "training_noise_sampling": "uniform",
89
+ "clip_sample": true,
90
+ "clip_sample_range": 1.0,
91
+ "num_inference_steps": 100,
92
+ "do_mask_loss_for_padding": false,
93
+ "optimizer_lr": 0.0001,
94
+ "optimizer_betas": [
95
+ 0.95,
96
+ 0.999
97
+ ],
98
+ "optimizer_eps": 1e-08,
99
+ "optimizer_weight_decay": 1e-06,
100
+ "scheduler_name": "cosine",
101
+ "scheduler_warmup_steps": 500
102
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1dad717b4391227edc41f128040d48aca8771ac498f108c0c03a7785d5b51bbb
3
+ size 774628052
train_config.json ADDED
@@ -0,0 +1,236 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": "continuallearning/pretraining_v2_lerobot30",
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
+ "affine": {
62
+ "weight": 1.0,
63
+ "type": "RandomAffine",
64
+ "kwargs": {
65
+ "degrees": [
66
+ -5.0,
67
+ 5.0
68
+ ],
69
+ "translate": [
70
+ 0.05,
71
+ 0.05
72
+ ]
73
+ }
74
+ }
75
+ }
76
+ },
77
+ "revision": null,
78
+ "use_imagenet_stats": true,
79
+ "video_backend": "pyav",
80
+ "streaming": false
81
+ },
82
+ "env": null,
83
+ "policy": {
84
+ "type": "ditflow_mt",
85
+ "n_obs_steps": 2,
86
+ "input_features": {
87
+ "observation.images.primary": {
88
+ "type": "VISUAL",
89
+ "shape": [
90
+ 3,
91
+ 256,
92
+ 256
93
+ ]
94
+ },
95
+ "observation.images.wrist": {
96
+ "type": "VISUAL",
97
+ "shape": [
98
+ 3,
99
+ 256,
100
+ 256
101
+ ]
102
+ },
103
+ "observation.state.cartesian": {
104
+ "type": "STATE",
105
+ "shape": [
106
+ 6
107
+ ]
108
+ },
109
+ "observation.state.gripper": {
110
+ "type": "STATE",
111
+ "shape": [
112
+ 1
113
+ ]
114
+ },
115
+ "observation.state.joints": {
116
+ "type": "STATE",
117
+ "shape": [
118
+ 7
119
+ ]
120
+ },
121
+ "observation.state.target": {
122
+ "type": "STATE",
123
+ "shape": [
124
+ 6
125
+ ]
126
+ },
127
+ "observation.state": {
128
+ "type": "STATE",
129
+ "shape": [
130
+ 20
131
+ ]
132
+ }
133
+ },
134
+ "output_features": {
135
+ "action": {
136
+ "type": "ACTION",
137
+ "shape": [
138
+ 7
139
+ ]
140
+ }
141
+ },
142
+ "device": "cuda",
143
+ "use_amp": false,
144
+ "use_peft": false,
145
+ "push_to_hub": false,
146
+ "repo_id": null,
147
+ "private": null,
148
+ "tags": null,
149
+ "license": null,
150
+ "pretrained_path": null,
151
+ "horizon": 16,
152
+ "n_action_steps": 8,
153
+ "normalization_mapping": {
154
+ "VISUAL": "MEAN_STD",
155
+ "STATE": "MIN_MAX",
156
+ "ACTION": "MIN_MAX"
157
+ },
158
+ "drop_n_last_frames": 7,
159
+ "tokenizer_max_length": 48,
160
+ "language_model_name": "openai/clip-vit-base-patch32",
161
+ "freeze_language_pretrained": true,
162
+ "vit_name": "facebook/dinov2-base",
163
+ "frequency_embedding_dim": 256,
164
+ "hidden_dim": 512,
165
+ "num_blocks": 6,
166
+ "num_heads": 16,
167
+ "dropout": 0.1,
168
+ "dim_feedforward": 4096,
169
+ "activation": "gelu",
170
+ "training_noise_sampling": "uniform",
171
+ "clip_sample": true,
172
+ "clip_sample_range": 1.0,
173
+ "num_inference_steps": 100,
174
+ "do_mask_loss_for_padding": false,
175
+ "optimizer_lr": 0.0001,
176
+ "optimizer_betas": [
177
+ 0.95,
178
+ 0.999
179
+ ],
180
+ "optimizer_eps": 1e-08,
181
+ "optimizer_weight_decay": 1e-06,
182
+ "scheduler_name": "cosine",
183
+ "scheduler_warmup_steps": 500
184
+ },
185
+ "output_dir": "outputs/real/dit_flow_mt_franka_pretrain",
186
+ "job_name": "dit_flow_mt_franka_pretrain",
187
+ "resume": false,
188
+ "seed": 42,
189
+ "num_workers": 16,
190
+ "batch_size": 32,
191
+ "steps": 200000,
192
+ "eval_freq": 0,
193
+ "log_freq": 100,
194
+ "tolerance_s": 0.0001,
195
+ "save_checkpoint": true,
196
+ "save_freq": 50000,
197
+ "use_policy_training_preset": true,
198
+ "optimizer": {
199
+ "type": "adam",
200
+ "lr": 0.0001,
201
+ "weight_decay": 1e-06,
202
+ "grad_clip_norm": 10.0,
203
+ "betas": [
204
+ 0.95,
205
+ 0.999
206
+ ],
207
+ "eps": 1e-08
208
+ },
209
+ "scheduler": {
210
+ "type": "diffuser",
211
+ "num_warmup_steps": 500,
212
+ "name": "cosine"
213
+ },
214
+ "eval": {
215
+ "n_episodes": 50,
216
+ "batch_size": 50,
217
+ "use_async_envs": false
218
+ },
219
+ "wandb": {
220
+ "enable": true,
221
+ "disable_artifact": true,
222
+ "project": "clare_experiments",
223
+ "entity": "ralf-roemer",
224
+ "notes": null,
225
+ "run_id": "6qhxj60k",
226
+ "mode": null
227
+ },
228
+ "peft": null,
229
+ "use_rabc": false,
230
+ "rabc_progress_path": null,
231
+ "rabc_kappa": 0.01,
232
+ "rabc_epsilon": 1e-06,
233
+ "rabc_head_mode": "sparse",
234
+ "rename_map": {},
235
+ "checkpoint_path": null
236
+ }