Hugging Face's logo

Mimic-Robotics
/
xvla_full_ttt_6e_100k

Safetensors
Model card Files
xet
 Community
ac-pate commited on
Commit
262c6e9
·
verified ·
1 Parent(s): 6db2bf7

Upload folder using huggingface_hub

Browse files
Files changed (7) hide show
  1. config.json +228 -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 +111 -0
  6. policy_preprocessor_step_7_normalizer_processor.safetensors +3 -0
  7. train_config.json +371 -0
config.json ADDED
@@ -0,0 +1,228 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "type": "xvla",
3
+ "n_obs_steps": 1,
4
+ "input_features": {
5
+ "observation.images.image": {
6
+ "type": "VISUAL",
7
+ "shape": [
8
+ 3,
9
+ 480,
10
+ 640
11
+ ]
12
+ },
13
+ "observation.images.image2": {
14
+ "type": "VISUAL",
15
+ "shape": [
16
+ 3,
17
+ 480,
18
+ 640
19
+ ]
20
+ },
21
+ "observation.state": {
22
+ "type": "STATE",
23
+ "shape": [
24
+ 15
25
+ ]
26
+ },
27
+ "observation.images.image3": {
28
+ "type": "VISUAL",
29
+ "shape": [
30
+ 3,
31
+ 224,
32
+ 224
33
+ ]
34
+ },
35
+ "observation.instruction": {
36
+ "type": "LANGUAGE",
37
+ "shape": [
38
+ 1
39
+ ]
40
+ }
41
+ },
42
+ "output_features": {
43
+ "action": {
44
+ "type": "ACTION",
45
+ "shape": [
46
+ 15
47
+ ]
48
+ }
49
+ },
50
+ "device": "cuda",
51
+ "use_amp": false,
52
+ "use_peft": false,
53
+ "push_to_hub": true,
54
+ "repo_id": "Mimic-Robotics/xvla_odin_full_ttt__xvla_vf",
55
+ "private": null,
56
+ "tags": null,
57
+ "license": null,
58
+ "pretrained_path": "lerobot/xvla-base",
59
+ "chunk_size": 50,
60
+ "n_action_steps": 50,
61
+ "dtype": "bfloat16",
62
+ "normalization_mapping": {
63
+ "STATE": "IDENTITY",
64
+ "ACTION": "MEAN_STD",
65
+ "VISUAL": "IDENTITY"
66
+ },
67
+ "florence_config": {
68
+ "model_type": "florence2",
69
+ "bos_token_id": 0,
70
+ "eos_token_id": 2,
71
+ "ignore_index": -100,
72
+ "pad_token_id": 1,
73
+ "projection_dim": 1024,
74
+ "text_config": {
75
+ "vocab_size": 51289,
76
+ "activation_dropout": 0.1,
77
+ "activation_function": "gelu",
78
+ "add_bias_logits": false,
79
+ "add_final_layer_norm": false,
80
+ "attention_dropout": 0.1,
81
+ "bos_token_id": 0,
82
+ "classif_dropout": 0.1,
83
+ "classifier_dropout": 0.0,
84
+ "d_model": 1024,
85
+ "decoder_attention_heads": 16,
86
+ "decoder_ffn_dim": 4096,
87
+ "decoder_layerdrop": 0.0,
88
+ "decoder_layers": 12,
89
+ "decoder_start_token_id": 2,
90
+ "dropout": 0.1,
91
+ "early_stopping": true,
92
+ "encoder_attention_heads": 16,
93
+ "encoder_ffn_dim": 4096,
94
+ "encoder_layerdrop": 0.0,
95
+ "encoder_layers": 12,
96
+ "eos_token_id": 2,
97
+ "forced_eos_token_id": 2,
98
+ "forced_bos_token_id": 0,
99
+ "gradient_checkpointing": false,
100
+ "init_std": 0.02,
101
+ "is_encoder_decoder": true,
102
+ "label2id": {
103
+ "LABEL_0": 0,
104
+ "LABEL_1": 1,
105
+ "LABEL_2": 2
106
+ },
107
+ "max_position_embeddings": 4096,
108
+ "no_repeat_ngram_size": 3,
109
+ "normalize_before": false,
110
+ "num_hidden_layers": 12,
111
+ "pad_token_id": 1,
112
+ "scale_embedding": false,
113
+ "num_beams": 3
114
+ },
115
+ "vision_config": {
116
+ "model_type": "davit",
117
+ "drop_path_rate": 0.1,
118
+ "patch_size": [
119
+ 7,
120
+ 3,
121
+ 3,
122
+ 3
123
+ ],
124
+ "patch_stride": [
125
+ 4,
126
+ 2,
127
+ 2,
128
+ 2
129
+ ],
130
+ "patch_padding": [
131
+ 3,
132
+ 1,
133
+ 1,
134
+ 1
135
+ ],
136
+ "patch_prenorm": [
137
+ false,
138
+ true,
139
+ true,
140
+ true
141
+ ],
142
+ "enable_checkpoint": false,
143
+ "dim_embed": [
144
+ 256,
145
+ 512,
146
+ 1024,
147
+ 2048
148
+ ],
149
+ "num_heads": [
150
+ 8,
151
+ 16,
152
+ 32,
153
+ 64
154
+ ],
155
+ "num_groups": [
156
+ 8,
157
+ 16,
158
+ 32,
159
+ 64
160
+ ],
161
+ "depths": [
162
+ 1,
163
+ 1,
164
+ 9,
165
+ 1
166
+ ],
167
+ "window_size": 12,
168
+ "projection_dim": 1024,
169
+ "visual_temporal_embedding": {
170
+ "type": "COSINE",
171
+ "max_temporal_embeddings": 100
172
+ },
173
+ "image_pos_embed": {
174
+ "type": "learned_abs_2d",
175
+ "max_pos_embeddings": 50
176
+ },
177
+ "image_feature_source": [
178
+ "spatial_avg_pool",
179
+ "temporal_avg_pool"
180
+ ]
181
+ },
182
+ "vocab_size": 51289,
183
+ "torch_dtype": "float32",
184
+ "is_encoder_decoder": true
185
+ },
186
+ "tokenizer_name": "facebook/bart-large",
187
+ "tokenizer_max_length": 1024,
188
+ "tokenizer_padding_side": "right",
189
+ "pad_language_to": "max_length",
190
+ "hidden_size": 1024,
191
+ "depth": 24,
192
+ "num_heads": 16,
193
+ "mlp_ratio": 4.0,
194
+ "num_domains": 30,
195
+ "len_soft_prompts": 32,
196
+ "dim_time": 32,
197
+ "max_len_seq": 512,
198
+ "use_hetero_proj": false,
199
+ "action_mode": "auto",
200
+ "num_denoising_steps": 10,
201
+ "use_proprio": true,
202
+ "max_state_dim": 20,
203
+ "max_action_dim": 20,
204
+ "domain_feature_key": null,
205
+ "resize_imgs_with_padding": [
206
+ 224,
207
+ 224
208
+ ],
209
+ "num_image_views": 3,
210
+ "empty_cameras": 0,
211
+ "freeze_vision_encoder": true,
212
+ "freeze_language_encoder": false,
213
+ "train_policy_transformer": true,
214
+ "train_soft_prompts": true,
215
+ "optimizer_lr": 0.0001,
216
+ "optimizer_betas": [
217
+ 0.9,
218
+ 0.95
219
+ ],
220
+ "optimizer_eps": 1e-08,
221
+ "optimizer_weight_decay": 0.0001,
222
+ "optimizer_grad_clip_norm": 10.0,
223
+ "optimizer_soft_prompt_lr_scale": 1.0,
224
+ "optimizer_soft_prompt_warmup_lr_scale": null,
225
+ "scheduler_warmup_steps": 1000,
226
+ "scheduler_decay_steps": 300000,
227
+ "scheduler_decay_lr": 2.5e-06
228
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0162b0ce77ee3e276848f4641af030e7043b562c10fdcb2055c4f7f2a20d4e21
3
+ size 1759596986
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
+ 15
13
+ ]
14
+ }
15
+ },
16
+ "norm_map": {
17
+ "STATE": "IDENTITY",
18
+ "ACTION": "MEAN_STD",
19
+ "VISUAL": "IDENTITY"
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 sha256:83d3f6e6195a818669a921668352ae55e1bafacfeddbfd3950edb970f9103e51
3
+ size 10376
policy_preprocessor.json ADDED
@@ -0,0 +1,111 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "name": "policy_preprocessor",
3
+ "steps": [
4
+ {
5
+ "registry_name": "rename_observations_processor",
6
+ "config": {
7
+ "rename_map": {
8
+ "observation.images.left_wrist": "observation.images.image",
9
+ "observation.images.right_wrist": "observation.images.image2"
10
+ }
11
+ }
12
+ },
13
+ {
14
+ "registry_name": "to_batch_processor",
15
+ "config": {}
16
+ },
17
+ {
18
+ "registry_name": "tokenizer_processor",
19
+ "config": {
20
+ "max_length": 50,
21
+ "task_key": "task",
22
+ "padding_side": "right",
23
+ "padding": "max_length",
24
+ "truncation": true,
25
+ "tokenizer_name": "facebook/bart-large"
26
+ }
27
+ },
28
+ {
29
+ "registry_name": "xvla_add_domain_id",
30
+ "config": {
31
+ "domain_id": 0
32
+ }
33
+ },
34
+ {
35
+ "registry_name": "xvla_image_to_float",
36
+ "config": {
37
+ "image_keys": null,
38
+ "validate_range": true
39
+ }
40
+ },
41
+ {
42
+ "registry_name": "xvla_imagenet_normalize",
43
+ "config": {
44
+ "image_keys": null
45
+ }
46
+ },
47
+ {
48
+ "registry_name": "device_processor",
49
+ "config": {
50
+ "device": "cuda",
51
+ "float_dtype": null
52
+ }
53
+ },
54
+ {
55
+ "registry_name": "normalizer_processor",
56
+ "config": {
57
+ "eps": 1e-08,
58
+ "features": {
59
+ "observation.images.image": {
60
+ "type": "VISUAL",
61
+ "shape": [
62
+ 3,
63
+ 480,
64
+ 640
65
+ ]
66
+ },
67
+ "observation.images.image2": {
68
+ "type": "VISUAL",
69
+ "shape": [
70
+ 3,
71
+ 480,
72
+ 640
73
+ ]
74
+ },
75
+ "observation.state": {
76
+ "type": "STATE",
77
+ "shape": [
78
+ 15
79
+ ]
80
+ },
81
+ "observation.images.image3": {
82
+ "type": "VISUAL",
83
+ "shape": [
84
+ 3,
85
+ 224,
86
+ 224
87
+ ]
88
+ },
89
+ "observation.instruction": {
90
+ "type": "LANGUAGE",
91
+ "shape": [
92
+ 1
93
+ ]
94
+ },
95
+ "action": {
96
+ "type": "ACTION",
97
+ "shape": [
98
+ 15
99
+ ]
100
+ }
101
+ },
102
+ "norm_map": {
103
+ "STATE": "IDENTITY",
104
+ "ACTION": "MEAN_STD",
105
+ "VISUAL": "IDENTITY"
106
+ }
107
+ },
108
+ "state_file": "policy_preprocessor_step_7_normalizer_processor.safetensors"
109
+ }
110
+ ]
111
+ }
policy_preprocessor_step_7_normalizer_processor.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:83d3f6e6195a818669a921668352ae55e1bafacfeddbfd3950edb970f9103e51
3
+ size 10376
train_config.json ADDED
@@ -0,0 +1,371 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "dataset": {
3
+ "repo_id": [
4
+ "Mimic-Robotics/mimic_ttt_full30hz_rgb"
5
+ ],
6
+ "root": null,
7
+ "episodes": null,
8
+ "image_transforms": {
9
+ "enable": false,
10
+ "max_num_transforms": 3,
11
+ "random_order": false,
12
+ "tfs": {
13
+ "brightness": {
14
+ "weight": 1.0,
15
+ "type": "ColorJitter",
16
+ "kwargs": {
17
+ "brightness": [
18
+ 0.8,
19
+ 1.2
20
+ ]
21
+ }
22
+ },
23
+ "contrast": {
24
+ "weight": 1.0,
25
+ "type": "ColorJitter",
26
+ "kwargs": {
27
+ "contrast": [
28
+ 0.8,
29
+ 1.2
30
+ ]
31
+ }
32
+ },
33
+ "saturation": {
34
+ "weight": 1.0,
35
+ "type": "ColorJitter",
36
+ "kwargs": {
37
+ "saturation": [
38
+ 0.5,
39
+ 1.5
40
+ ]
41
+ }
42
+ },
43
+ "hue": {
44
+ "weight": 1.0,
45
+ "type": "ColorJitter",
46
+ "kwargs": {
47
+ "hue": [
48
+ -0.05,
49
+ 0.05
50
+ ]
51
+ }
52
+ },
53
+ "sharpness": {
54
+ "weight": 1.0,
55
+ "type": "SharpnessJitter",
56
+ "kwargs": {
57
+ "sharpness": [
58
+ 0.5,
59
+ 1.5
60
+ ]
61
+ }
62
+ },
63
+ "affine": {
64
+ "weight": 1.0,
65
+ "type": "RandomAffine",
66
+ "kwargs": {
67
+ "degrees": [
68
+ -5.0,
69
+ 5.0
70
+ ],
71
+ "translate": [
72
+ 0.05,
73
+ 0.05
74
+ ]
75
+ }
76
+ }
77
+ }
78
+ },
79
+ "revision": null,
80
+ "use_imagenet_stats": true,
81
+ "video_backend": "pyav",
82
+ "streaming": false
83
+ },
84
+ "env": null,
85
+ "policy": {
86
+ "type": "xvla",
87
+ "n_obs_steps": 1,
88
+ "input_features": {
89
+ "observation.images.image": {
90
+ "type": "VISUAL",
91
+ "shape": [
92
+ 3,
93
+ 480,
94
+ 640
95
+ ]
96
+ },
97
+ "observation.images.image2": {
98
+ "type": "VISUAL",
99
+ "shape": [
100
+ 3,
101
+ 480,
102
+ 640
103
+ ]
104
+ },
105
+ "observation.state": {
106
+ "type": "STATE",
107
+ "shape": [
108
+ 15
109
+ ]
110
+ },
111
+ "observation.images.image3": {
112
+ "type": "VISUAL",
113
+ "shape": [
114
+ 3,
115
+ 224,
116
+ 224
117
+ ]
118
+ },
119
+ "observation.instruction": {
120
+ "type": "LANGUAGE",
121
+ "shape": [
122
+ 1
123
+ ]
124
+ }
125
+ },
126
+ "output_features": {
127
+ "action": {
128
+ "type": "ACTION",
129
+ "shape": [
130
+ 15
131
+ ]
132
+ }
133
+ },
134
+ "device": "cuda",
135
+ "use_amp": false,
136
+ "use_peft": false,
137
+ "push_to_hub": true,
138
+ "repo_id": "Mimic-Robotics/xvla_odin_full_ttt__xvla_vf",
139
+ "private": null,
140
+ "tags": null,
141
+ "license": null,
142
+ "pretrained_path": "lerobot/xvla-base",
143
+ "chunk_size": 50,
144
+ "n_action_steps": 50,
145
+ "dtype": "bfloat16",
146
+ "normalization_mapping": {
147
+ "STATE": "IDENTITY",
148
+ "ACTION": "MEAN_STD",
149
+ "VISUAL": "IDENTITY"
150
+ },
151
+ "florence_config": {
152
+ "model_type": "florence2",
153
+ "bos_token_id": 0,
154
+ "eos_token_id": 2,
155
+ "ignore_index": -100,
156
+ "pad_token_id": 1,
157
+ "projection_dim": 1024,
158
+ "text_config": {
159
+ "vocab_size": 51289,
160
+ "activation_dropout": 0.1,
161
+ "activation_function": "gelu",
162
+ "add_bias_logits": false,
163
+ "add_final_layer_norm": false,
164
+ "attention_dropout": 0.1,
165
+ "bos_token_id": 0,
166
+ "classif_dropout": 0.1,
167
+ "classifier_dropout": 0.0,
168
+ "d_model": 1024,
169
+ "decoder_attention_heads": 16,
170
+ "decoder_ffn_dim": 4096,
171
+ "decoder_layerdrop": 0.0,
172
+ "decoder_layers": 12,
173
+ "decoder_start_token_id": 2,
174
+ "dropout": 0.1,
175
+ "early_stopping": true,
176
+ "encoder_attention_heads": 16,
177
+ "encoder_ffn_dim": 4096,
178
+ "encoder_layerdrop": 0.0,
179
+ "encoder_layers": 12,
180
+ "eos_token_id": 2,
181
+ "forced_eos_token_id": 2,
182
+ "forced_bos_token_id": 0,
183
+ "gradient_checkpointing": false,
184
+ "init_std": 0.02,
185
+ "is_encoder_decoder": true,
186
+ "label2id": {
187
+ "LABEL_0": 0,
188
+ "LABEL_1": 1,
189
+ "LABEL_2": 2
190
+ },
191
+ "max_position_embeddings": 4096,
192
+ "no_repeat_ngram_size": 3,
193
+ "normalize_before": false,
194
+ "num_hidden_layers": 12,
195
+ "pad_token_id": 1,
196
+ "scale_embedding": false,
197
+ "num_beams": 3
198
+ },
199
+ "vision_config": {
200
+ "model_type": "davit",
201
+ "drop_path_rate": 0.1,
202
+ "patch_size": [
203
+ 7,
204
+ 3,
205
+ 3,
206
+ 3
207
+ ],
208
+ "patch_stride": [
209
+ 4,
210
+ 2,
211
+ 2,
212
+ 2
213
+ ],
214
+ "patch_padding": [
215
+ 3,
216
+ 1,
217
+ 1,
218
+ 1
219
+ ],
220
+ "patch_prenorm": [
221
+ false,
222
+ true,
223
+ true,
224
+ true
225
+ ],
226
+ "enable_checkpoint": false,
227
+ "dim_embed": [
228
+ 256,
229
+ 512,
230
+ 1024,
231
+ 2048
232
+ ],
233
+ "num_heads": [
234
+ 8,
235
+ 16,
236
+ 32,
237
+ 64
238
+ ],
239
+ "num_groups": [
240
+ 8,
241
+ 16,
242
+ 32,
243
+ 64
244
+ ],
245
+ "depths": [
246
+ 1,
247
+ 1,
248
+ 9,
249
+ 1
250
+ ],
251
+ "window_size": 12,
252
+ "projection_dim": 1024,
253
+ "visual_temporal_embedding": {
254
+ "type": "COSINE",
255
+ "max_temporal_embeddings": 100
256
+ },
257
+ "image_pos_embed": {
258
+ "type": "learned_abs_2d",
259
+ "max_pos_embeddings": 50
260
+ },
261
+ "image_feature_source": [
262
+ "spatial_avg_pool",
263
+ "temporal_avg_pool"
264
+ ]
265
+ },
266
+ "vocab_size": 51289,
267
+ "torch_dtype": "float32",
268
+ "is_encoder_decoder": true
269
+ },
270
+ "tokenizer_name": "facebook/bart-large",
271
+ "tokenizer_max_length": 1024,
272
+ "tokenizer_padding_side": "right",
273
+ "pad_language_to": "max_length",
274
+ "hidden_size": 1024,
275
+ "depth": 24,
276
+ "num_heads": 16,
277
+ "mlp_ratio": 4.0,
278
+ "num_domains": 30,
279
+ "len_soft_prompts": 32,
280
+ "dim_time": 32,
281
+ "max_len_seq": 512,
282
+ "use_hetero_proj": false,
283
+ "action_mode": "auto",
284
+ "num_denoising_steps": 10,
285
+ "use_proprio": true,
286
+ "max_state_dim": 20,
287
+ "max_action_dim": 20,
288
+ "domain_feature_key": null,
289
+ "resize_imgs_with_padding": [
290
+ 224,
291
+ 224
292
+ ],
293
+ "num_image_views": 3,
294
+ "empty_cameras": 0,
295
+ "freeze_vision_encoder": true,
296
+ "freeze_language_encoder": false,
297
+ "train_policy_transformer": true,
298
+ "train_soft_prompts": true,
299
+ "optimizer_lr": 0.0001,
300
+ "optimizer_betas": [
301
+ 0.9,
302
+ 0.95
303
+ ],
304
+ "optimizer_eps": 1e-08,
305
+ "optimizer_weight_decay": 0.0001,
306
+ "optimizer_grad_clip_norm": 10.0,
307
+ "optimizer_soft_prompt_lr_scale": 1.0,
308
+ "optimizer_soft_prompt_warmup_lr_scale": null,
309
+ "scheduler_warmup_steps": 1000,
310
+ "scheduler_decay_steps": 300000,
311
+ "scheduler_decay_lr": 2.5e-06
312
+ },
313
+ "output_dir": "/home/odin/mimic-lerobot/outputs/train/xvla_odin_full_ttt__xvla_vf_20260302_1548",
314
+ "job_name": "xvla_odin_full_ttt__xvla_vf_20260302_1548",
315
+ "resume": false,
316
+ "seed": 1000,
317
+ "num_workers": 8,
318
+ "batch_size": 44,
319
+ "steps": 300000,
320
+ "eval_freq": 20000,
321
+ "log_freq": 200,
322
+ "tolerance_s": 0.0001,
323
+ "save_checkpoint": true,
324
+ "save_freq": 25000,
325
+ "use_policy_training_preset": true,
326
+ "optimizer": {
327
+ "type": "xvla-adamw",
328
+ "lr": 0.0001,
329
+ "weight_decay": 0.0001,
330
+ "grad_clip_norm": 10.0,
331
+ "betas": [
332
+ 0.9,
333
+ 0.95
334
+ ],
335
+ "eps": 1e-08,
336
+ "soft_prompt_lr_scale": 1.0,
337
+ "soft_prompt_warmup_lr_scale": null
338
+ },
339
+ "scheduler": {
340
+ "type": "cosine_decay_with_warmup",
341
+ "num_warmup_steps": 1000,
342
+ "num_decay_steps": 300000,
343
+ "peak_lr": 0.0001,
344
+ "decay_lr": 2.5e-06
345
+ },
346
+ "eval": {
347
+ "n_episodes": 50,
348
+ "batch_size": 50,
349
+ "use_async_envs": false
350
+ },
351
+ "wandb": {
352
+ "enable": true,
353
+ "disable_artifact": false,
354
+ "project": "lerobot",
355
+ "entity": null,
356
+ "notes": null,
357
+ "run_id": "u451vq89",
358
+ "mode": null
359
+ },
360
+ "peft": null,
361
+ "use_rabc": false,
362
+ "rabc_progress_path": null,
363
+ "rabc_kappa": 0.01,
364
+ "rabc_epsilon": 1e-06,
365
+ "rabc_head_mode": "sparse",
366
+ "rename_map": {
367
+ "observation.images.left_wrist": "observation.images.image",
368
+ "observation.images.right_wrist": "observation.images.image2"
369
+ },
370
+ "checkpoint_path": null
371
+ }