Hugging Face's logo

Datasets:
saaduddinM
/
OXE_fmb_dataset_embeddings

Tasks:
Robotics
Modalities:
Tabular
Time-series
Formats:
parquet
Size:
1M - 10M
Tags:
LeRobot
language_table
openx
franka
Libraries:
Datasets
Dask
License:
Dataset card Data Studio Files
xet
Community
1
saaduddinM commited on
Commit
7414c59
·
verified ·
1 Parent(s): a886a38

Add files using upload-large-folder tool

Browse files
Files changed (50) hide show
  1. README.md +300 -0
  2. data/shard-00000-of-00064.parquet +3 -0
  3. data/shard-00001-of-00064.parquet +3 -0
  4. data/shard-00002-of-00064.parquet +3 -0
  5. data/shard-00004-of-00064.parquet +3 -0
  6. data/shard-00005-of-00064.parquet +3 -0
  7. data/shard-00006-of-00064.parquet +3 -0
  8. data/shard-00007-of-00064.parquet +3 -0
  9. data/shard-00008-of-00064.parquet +3 -0
  10. data/shard-00011-of-00064.parquet +3 -0
  11. data/shard-00012-of-00064.parquet +3 -0
  12. data/shard-00013-of-00064.parquet +3 -0
  13. data/shard-00019-of-00064.parquet +3 -0
  14. data/shard-00020-of-00064.parquet +3 -0
  15. data/shard-00021-of-00064.parquet +3 -0
  16. data/shard-00024-of-00064.parquet +3 -0
  17. data/shard-00026-of-00064.parquet +3 -0
  18. data/shard-00028-of-00064.parquet +3 -0
  19. data/shard-00029-of-00064.parquet +3 -0
  20. data/shard-00031-of-00064.parquet +3 -0
  21. data/shard-00032-of-00064.parquet +3 -0
  22. data/shard-00033-of-00064.parquet +3 -0
  23. data/shard-00034-of-00064.parquet +3 -0
  24. data/shard-00035-of-00064.parquet +3 -0
  25. data/shard-00036-of-00064.parquet +3 -0
  26. data/shard-00038-of-00064.parquet +3 -0
  27. data/shard-00039-of-00064.parquet +3 -0
  28. data/shard-00040-of-00064.parquet +3 -0
  29. data/shard-00042-of-00064.parquet +3 -0
  30. data/shard-00043-of-00064.parquet +3 -0
  31. data/shard-00044-of-00064.parquet +3 -0
  32. data/shard-00045-of-00064.parquet +3 -0
  33. data/shard-00047-of-00064.parquet +3 -0
  34. data/shard-00048-of-00064.parquet +3 -0
  35. data/shard-00050-of-00064.parquet +3 -0
  36. data/shard-00051-of-00064.parquet +3 -0
  37. data/shard-00052-of-00064.parquet +3 -0
  38. data/shard-00054-of-00064.parquet +3 -0
  39. data/shard-00055-of-00064.parquet +3 -0
  40. data/shard-00056-of-00064.parquet +3 -0
  41. data/shard-00057-of-00064.parquet +3 -0
  42. data/shard-00058-of-00064.parquet +3 -0
  43. data/shard-00061-of-00064.parquet +3 -0
  44. data/shard-00063-of-00064.parquet +3 -0
  45. data/task_text_embeddings.parquet +3 -0
  46. meta/episodes.jsonl +0 -0
  47. meta/info.json +151 -0
  48. meta/sharded_index.json +0 -0
  49. meta/task_text_embeddings_info.json +7 -0
  50. meta/tasks.jsonl +24 -0
README.md ADDED
@@ -0,0 +1,300 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ task_categories:
4
+ - robotics
5
+ tags:
6
+ - LeRobot
7
+ - language_table
8
+ - openx
9
+ - franka
10
+ configs:
11
+ - config_name: default
12
+ data_files: data/*.parquet
13
+ ---
14
+ Language Table (LeRobot) — Embedding-Only Release
15
+ (DINOv3 + SigLIP2 image features; EmbeddingGemma task-text features)
16
+
17
+ This repository packages a re-encoded variant of [IPEC-COMMUNITY/fmb_dataset_lerobot](https://huggingface.co/datasets/IPEC-COMMUNITY/fmb_dataset_lerobot) where raw videos are replaced by fixed-length image embeddings, and task strings are augmented with text embeddings. All indices, splits, and semantics remain consistent with the source dataset while storage and I/O are substantially lighter. To make the dataset practical to upload/download and stream from the Hub, we also consolidated tiny per-episode Parquet files into N large Parquet shards under a single data/ folder. The file meta/sharded_index.json preserves a precise mapping from each original episode (referenced by a normalized identifier of the form data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet) to its shard path and row range, so you keep original addressing without paying the small-file tax.
18
+
19
+ - Robot: Franka
20
+ - Modalities kept: states, actions, timestamps, frame/episode indices, image embeddings, task-text embeddings
21
+ - Removed:
22
+ - observation.images.image_side_1
23
+ - observation.images.image_side_2
24
+ - observation.images.image_wrist_1
25
+ - observation.images.image_wrist_2
26
+ - License: apache-2.0 (inherits from source)
27
+
28
+ ----------------------------------------------------------------
29
+
30
+ Quick Stats
31
+
32
+ From meta/info.json and meta/task_text_embeddings_info.json:
33
+
34
+ - Episodes: 8,612
35
+ - Frames: 1,137,459
36
+ - Tasks (unique): 24
37
+ - Chunks (original layout): 9 (chunks_size=1000)
38
+ - Shards (this release): 64 Parquet files under data/ (see meta/sharded_index.json)
39
+ - FPS: 10
40
+ - Image embeddings (per frame):
41
+ - observation.images.image_side_1_dinov3 → float32 [1024] (DINOv3 ViT-L/16 CLS)
42
+ - observation.images.image_side_1_siglip2 → float32 [768] (SigLIP2-base)
43
+ - observation.images.image_side_2_dinov3 → float32 [1024] (DINOv3 ViT-L/16 CLS)
44
+ - observation.images.image_side_2_siglip2 → float32 [768] (SigLIP2-base)
45
+ - observation.images.image_wrist_1_dinov3 → float32 [1024] (DINOv3 ViT-L/16 CLS)
46
+ - observation.images.image_wrist_1_siglip2 → float32 [768] (SigLIP2-base)
47
+ - observation.images.image_wrist_2_dinov3 → float32 [1024] (DINOv3 ViT-L/16 CLS)
48
+ - observation.images.image_wrist_2_siglip2 → float32 [768] (SigLIP2-base)
49
+ - Task-text embeddings (per unique task):
50
+ - embedding → float32 [768] from google/embeddinggemma-300m
51
+ - Count: 24 rows (one per task)
52
+
53
+ Note: This is an embedding-only package. The original pixel arrays listed under “Removed” are dropped.
54
+
55
+ ----------------------------------------------------------------
56
+
57
+ <details>
58
+ <summary><b>Contents</b></summary>
59
+
60
+ <pre>
61
+ .
62
+ |-- meta/
63
+ | |-- info.json
64
+ | |-- sharded_index.json
65
+ | |-- tasks.jsonl
66
+ | |-- episodes.jsonl
67
+ | `-- task_text_embeddings_info.json
68
+ |-- data/
69
+ | |-- shard-00000-of-000NN.parquet
70
+ | |-- shard-00001-of-000NN.parquet
71
+ | |-- ...
72
+ | `-- task_text_embeddings.parquet
73
+ `-- README.md
74
+ </pre>
75
+ </details>
76
+
77
+ ----------------------------------------------------------------
78
+
79
+ How This Was Generated (Reproducible Pipeline)
80
+
81
+ 1) Episode → Image Embeddings (drop pixels)
82
+ convert_lerobot_to_embeddings_mono.py (GPU-accelerated preprocessing).
83
+ Adds:
84
+ - observation.images.image_side_1_dinov3 (float32[1024])
85
+ - observation.images.image_side_1_siglip2 (float32[768])
86
+ - observation.images.image_side_2_dinov3 (float32[1024])
87
+ - observation.images.image_side_2_siglip2 (float32[768])
88
+ - observation.images.image_wrist_1_dinov3 (float32[1024])
89
+ - observation.images.image_wrist_1_siglip2 (float32[768])
90
+ - observation.images.image_wrist_2_dinov3 (float32[1024])
91
+ - observation.images.image_wrist_2_siglip2 (float32[768])
92
+ Removes:
93
+ - observation.images.image_side_1
94
+ - observation.images.image_side_2
95
+ - observation.images.image_wrist_1
96
+ - observation.images.image_wrist_2
97
+
98
+ 2) Task-Text Embeddings (one row per unique task)
99
+ build_task_text_embeddings.py with SentenceTransformer("google/embeddinggemma-300m") → data/task_text_embeddings.parquet + meta/task_text_embeddings_info.json.
100
+
101
+ 3) Data Consolidation (this release)
102
+ All per-episode Parquets were consolidated into N large Parquet shards in one data/ folder.
103
+ - The index meta/sharded_index.json records, for each episode, its normalized source identifier data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet, the destination shard path, and the (row_offset, num_rows) range inside that shard.
104
+ - This preserves original addressing while making Hub sync/clone/stream far faster and more reliable.
105
+
106
+ ----------------------------------------------------------------
107
+
108
+ Metadata (Excerpts)
109
+
110
+ meta/task_text_embeddings_info.json
111
+
112
+ ~~~json
113
+ {
114
+ "model": "google/embeddinggemma-300m",
115
+ "dimension": 768,
116
+ "normalized": true,
117
+ "count": 24,
118
+ "file": "task_text_embeddings.parquet"
119
+ }
120
+ ~~~
121
+
122
+
123
+ meta/info.json (embedding-only + shards)
124
+
125
+ ~~~json
126
+ {
127
+ "codebase_version": "v2.1-embeddings-sharded",
128
+ "robot_type": "franka",
129
+ "total_episodes": 8612,
130
+ "total_frames": 1137459,
131
+ "total_tasks": 24,
132
+ "total_videos": 34448,
133
+ "total_chunks": 9,
134
+ "chunks_size": 1000,
135
+ "fps": 10,
136
+ "splits": {
137
+ "train": "0:8612"
138
+ },
139
+ "data_path": "data/shard-{shard_id:05d}-of-{num_shards:05d}.parquet",
140
+ "features": {
141
+ "observation.state": {
142
+ "dtype": "float32",
143
+ "shape": [
144
+ 8
145
+ ],
146
+ "names": {
147
+ "motors": [
148
+ "x",
149
+ "y",
150
+ "z",
151
+ "roll",
152
+ "pitch",
153
+ "yaw",
154
+ "pad",
155
+ "gripper"
156
+ ]
157
+ }
158
+ },
159
+ "action": {
160
+ "dtype": "float32",
161
+ "shape": [
162
+ 7
163
+ ],
164
+ "names": {
165
+ "motors": [
166
+ "x",
167
+ "y",
168
+ "z",
169
+ "roll",
170
+ "pitch",
171
+ "yaw",
172
+ "gripper"
173
+ ]
174
+ }
175
+ },
176
+ "timestamp": {
177
+ "dtype": "float32",
178
+ "shape": [
179
+ 1
180
+ ],
181
+ "names": null
182
+ },
183
+ "frame_index": {
184
+ "dtype": "int64",
185
+ "shape": [
186
+ 1
187
+ ],
188
+ "names": null
189
+ },
190
+ "episode_index": {
191
+ "dtype": "int64",
192
+ "shape": [
193
+ 1
194
+ ],
195
+ "names": null
196
+ },
197
+ "index": {
198
+ "dtype": "int64",
199
+ "shape": [
200
+ 1
201
+ ],
202
+ "names": null
203
+ },
204
+ "task_index": {
205
+ "dtype": "int64",
206
+ "shape": [
207
+ 1
208
+ ],
209
+ "names": null
210
+ },
211
+ "observation.images.image_side_1_dinov3": {
212
+ "dtype": "float32",
213
+ "shape": [
214
+ 1024
215
+ ],
216
+ "names": null
217
+ },
218
+ "observation.images.image_side_1_siglip2": {
219
+ "dtype": "float32",
220
+ "shape": [
221
+ 768
222
+ ],
223
+ "names": null
224
+ },
225
+ "observation.images.image_side_2_dinov3": {
226
+ "dtype": "float32",
227
+ "shape": [
228
+ 1024
229
+ ],
230
+ "names": null
231
+ },
232
+ "observation.images.image_side_2_siglip2": {
233
+ "dtype": "float32",
234
+ "shape": [
235
+ 768
236
+ ],
237
+ "names": null
238
+ },
239
+ "observation.images.image_wrist_1_dinov3": {
240
+ "dtype": "float32",
241
+ "shape": [
242
+ 1024
243
+ ],
244
+ "names": null
245
+ },
246
+ "observation.images.image_wrist_1_siglip2": {
247
+ "dtype": "float32",
248
+ "shape": [
249
+ 768
250
+ ],
251
+ "names": null
252
+ },
253
+ "observation.images.image_wrist_2_dinov3": {
254
+ "dtype": "float32",
255
+ "shape": [
256
+ 1024
257
+ ],
258
+ "names": null
259
+ },
260
+ "observation.images.image_wrist_2_siglip2": {
261
+ "dtype": "float32",
262
+ "shape": [
263
+ 768
264
+ ],
265
+ "names": null
266
+ }
267
+ },
268
+ "video_keys": [
269
+ "observation.images.image_side_1",
270
+ "observation.images.image_side_2",
271
+ "observation.images.image_wrist_1",
272
+ "observation.images.image_wrist_2"
273
+ ],
274
+ "num_shards": 64,
275
+ "index_path": "meta/sharded_index.json"
276
+ }
277
+ ~~~
278
+
279
+ ----------------------------------------------------------------
280
+
281
+ Environment & Dependencies
282
+
283
+ Python ≥ 3.9 • PyTorch ≥ 2.1 • transformers • sentence-transformers • pyarrow • tqdm • decord (and optionally av)
284
+
285
+ ----------------------------------------------------------------
286
+
287
+ Provenance, License, and Citation
288
+
289
+ - Source dataset: [IPEC-COMMUNITY/fmb_dataset_lerobot](https://huggingface.co/datasets/IPEC-COMMUNITY/fmb_dataset_lerobot)
290
+ - License: apache-2.0 (inherits from the source)
291
+ - Encoders to cite:
292
+ - facebook/dinov3-vitl16-pretrain-lvd1689m
293
+ - google/siglip2-base-patch16-384
294
+ - google/embeddinggemma-300m
295
+
296
+ ----------------------------------------------------------------
297
+
298
+ Changelog
299
+
300
+ - v2.0-embeddings-sharded — Replaced video tensors with DINOv3 + SigLIP2 features; added EmbeddingGemma task-text embeddings; consolidated per-episode Parquets into N shards with a repo-local index; preserved original indexing/splits via normalized episode identifiers.
data/shard-00000-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8de1bcc593fce6c4ce1615b004603ed60623acd99f09ef1bc96070d7523c18cc
3
+ size 365975247
data/shard-00001-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:100183ae59ca90687f2aa4b33b8458e915f8f6db855562627e4a92747cc026ea
3
+ size 377511656
data/shard-00002-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:41c881f3d89d78589b81ec97323e94798c2eb69770154dd644e4cbc93ac5d5d7
3
+ size 380133059
data/shard-00004-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a68a98725d774017b8b6dba57635507ac7aaed231a283f0f317bf43428b47582
3
+ size 347252317
data/shard-00005-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:77a4e71031ec56fab3d3a3f3d62acab7a9e4c9b065e20d82963557dec25414f1
3
+ size 400689256
data/shard-00006-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4031328258a49b165171dc0de41960569d2c672cd64190f04e69c5d88784fb65
3
+ size 400490091
data/shard-00007-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:37c378a747f7ad1440a2fd7d0f5be3382b178f34fd936dce10a6aab337194f05
3
+ size 335469200
data/shard-00008-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:669b71e5889d55deda086a3ed4786c5ccc341a1551ecf5ec63da046cfb91ea25
3
+ size 334590836
data/shard-00011-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:990365c020a09b06195227d21ce5b2fc102a9b8fc0f81a717a6529a02a057a75
3
+ size 338356260
data/shard-00012-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8e014473d6660895f055460368bc7e36b00eb3e27ab05d40550d292d1977c575
3
+ size 381166310
data/shard-00013-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:71ddde11422316aba931dd26f5aa6e1e6a60c1e7dc007d5d5ed9bcc34f4faf5e
3
+ size 340256794
data/shard-00019-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7b6a55de43e140da19746666f928ac672ba12eeb579f39b31e261b1a124bd2ee
3
+ size 352337381
data/shard-00020-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c5f6b20830e73d6ff8cd086fc0f120f518a9a7a884d61d7836572f31ecfa9ea0
3
+ size 383076565
data/shard-00021-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:03ff5cbba00b6af1e1da908ff154c142a400c7627ef16a01a63aad3213df7d87
3
+ size 352166158
data/shard-00024-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7146b302c89db7e3ad23e393ba6157bc7cf71c366c166d23b0bdb970c1c344e4
3
+ size 375860430
data/shard-00026-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fcba91d978d6e03b600ca7764ffaa6f6ed8a7b2a7209cdc66e2c69c6d1675e56
3
+ size 338778423
data/shard-00028-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f8d4e0541a13e3901f916b578a9acd0bd062e2535e83f83b2f3afacebaeab734
3
+ size 400447326
data/shard-00029-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f4d616d073c5d34b15d8c53b1d5711f760dd44b1b0a76316d9a0451a0a2e6454
3
+ size 403884755
data/shard-00031-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6afc86d1ecbd22ce2192af141cc4a4111a346517ed6507da1cd920beb5c27845
3
+ size 353707470
data/shard-00032-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:080c9d19ddf16f2c58d9bfdbe3c9d06e56c9d2d07540fb50aa90e37968cbf89d
3
+ size 361644190
data/shard-00033-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:03806c24ab885623fbf2e35ae19c07293925af54be1fe3a73e053a09d1935b3a
3
+ size 301567007
data/shard-00034-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c14e363648bc60ef7b79659f8d009162989e2e4ef642008cd9ba06bc7c2ef46
3
+ size 398503005
data/shard-00035-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eb54cb117471b362af53438a80d7a08afad0c073ef51e1214c2cff82a0dc8d23
3
+ size 372243841
data/shard-00036-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:284e134f6602cc012db05e632dc759d9674b1910569553f1cf332550bb26b7d7
3
+ size 352701684
data/shard-00038-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5cfdc4a28cea8cfe084d568eecdfe820233a2ffc79277be10b6fae3d86f79c7d
3
+ size 378587387
data/shard-00039-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ac5a9b34de763b719120e48b7c75b08e917a1fd3e454dea10757535213c3b1bb
3
+ size 324563551
data/shard-00040-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d6298e2ed019bec6b3fa0a8e4ee43fa3f2a79fd12f16d642fb097278fd86d094
3
+ size 354133101
data/shard-00042-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:62a5c2fb28e1cf4bc62f560e044eb581f2bfb4861a00e81a9bef81d0116f611a
3
+ size 349500596
data/shard-00043-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3067b763b8d4038aaf4d8da4aafae88e0c78ab3813f99a3a15a0bc186b9722fc
3
+ size 330702851
data/shard-00044-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f263cc5fe47b5b1b755681f229996ccd21298550b4b0f5614e3fbe00955cba30
3
+ size 343631621
data/shard-00045-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5e2331f6304997d24475384513f1101225c46b6662246a69a9b0f65b3e2aec09
3
+ size 385463983
data/shard-00047-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a2dc0c75d7dd5e677346044f9cc13f4a12e0de0a8bc76bcd257771a1c8cefa87
3
+ size 395843554
data/shard-00048-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1fca9e496fe23b0e38b7bf6ebd417dee25d7fe9184af54310540ce44f3bc3540
3
+ size 368780092
data/shard-00050-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a3b2d64effd93cf1c06c70dcfbfbf317f473cc371a601538c5a9229280ca20ce
3
+ size 337891071
data/shard-00051-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:849c57951addb73052f437de0242617a83fcddec22cf61fdb34058bc6451ee55
3
+ size 385344449
data/shard-00052-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bf81739958b3ade7bcd54b18e2d7c3b40a2e80880d78b551512d3975fd69318b
3
+ size 379511245
data/shard-00054-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7b3b43458977856acc12803f6ae41c1e58ea38a1b5514f4aaa378cb6be316737
3
+ size 369234970
data/shard-00055-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7a5bfef214738ec3b797a493c216d152a7f8ed1ad379ffad27c92f20f2df28c7
3
+ size 385522144
data/shard-00056-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:489897d990577c72c1fc68407b717426d7f33511a3a14e15ad36e64ee0a4acf0
3
+ size 403035980
data/shard-00057-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c41bc1d1a42a2509e77a11b4cae16e857feafb8809a577257b2cdf693f0eb738
3
+ size 369997099
data/shard-00058-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:24fd49d706fae1f6ceacf80a5e54b18be5dfad4a0ff3f2a4a5aef770b18f6da2
3
+ size 332852171
data/shard-00061-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:848d5499e3bae8eddab72d6826752bb62b1b3b35b17bc948e4365239d629fb60
3
+ size 399470827
data/shard-00063-of-00064.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9b56044b4a183dd268902cba3dc63ec91e8ae9f3faffcf353afe23d87c08b857
3
+ size 377512967
data/task_text_embeddings.parquet ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e35e2e12d166a360814bbc8e6cadf8d6f545ed348c9ead80972fa734de8f7593
3
+ size 69857
meta/episodes.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
meta/info.json ADDED
@@ -0,0 +1,151 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "codebase_version": "v2.1-embeddings-sharded",
3
+ "robot_type": "franka",
4
+ "total_episodes": 8612,
5
+ "total_frames": 1137459,
6
+ "total_tasks": 24,
7
+ "total_videos": 34448,
8
+ "total_chunks": 9,
9
+ "chunks_size": 1000,
10
+ "fps": 10,
11
+ "splits": {
12
+ "train": "0:8612"
13
+ },
14
+ "data_path": "data/shard-{shard_id:05d}-of-{num_shards:05d}.parquet",
15
+ "features": {
16
+ "observation.state": {
17
+ "dtype": "float32",
18
+ "shape": [
19
+ 8
20
+ ],
21
+ "names": {
22
+ "motors": [
23
+ "x",
24
+ "y",
25
+ "z",
26
+ "roll",
27
+ "pitch",
28
+ "yaw",
29
+ "pad",
30
+ "gripper"
31
+ ]
32
+ }
33
+ },
34
+ "action": {
35
+ "dtype": "float32",
36
+ "shape": [
37
+ 7
38
+ ],
39
+ "names": {
40
+ "motors": [
41
+ "x",
42
+ "y",
43
+ "z",
44
+ "roll",
45
+ "pitch",
46
+ "yaw",
47
+ "gripper"
48
+ ]
49
+ }
50
+ },
51
+ "timestamp": {
52
+ "dtype": "float32",
53
+ "shape": [
54
+ 1
55
+ ],
56
+ "names": null
57
+ },
58
+ "frame_index": {
59
+ "dtype": "int64",
60
+ "shape": [
61
+ 1
62
+ ],
63
+ "names": null
64
+ },
65
+ "episode_index": {
66
+ "dtype": "int64",
67
+ "shape": [
68
+ 1
69
+ ],
70
+ "names": null
71
+ },
72
+ "index": {
73
+ "dtype": "int64",
74
+ "shape": [
75
+ 1
76
+ ],
77
+ "names": null
78
+ },
79
+ "task_index": {
80
+ "dtype": "int64",
81
+ "shape": [
82
+ 1
83
+ ],
84
+ "names": null
85
+ },
86
+ "observation.images.image_side_1_dinov3": {
87
+ "dtype": "float32",
88
+ "shape": [
89
+ 1024
90
+ ],
91
+ "names": null
92
+ },
93
+ "observation.images.image_side_1_siglip2": {
94
+ "dtype": "float32",
95
+ "shape": [
96
+ 768
97
+ ],
98
+ "names": null
99
+ },
100
+ "observation.images.image_side_2_dinov3": {
101
+ "dtype": "float32",
102
+ "shape": [
103
+ 1024
104
+ ],
105
+ "names": null
106
+ },
107
+ "observation.images.image_side_2_siglip2": {
108
+ "dtype": "float32",
109
+ "shape": [
110
+ 768
111
+ ],
112
+ "names": null
113
+ },
114
+ "observation.images.image_wrist_1_dinov3": {
115
+ "dtype": "float32",
116
+ "shape": [
117
+ 1024
118
+ ],
119
+ "names": null
120
+ },
121
+ "observation.images.image_wrist_1_siglip2": {
122
+ "dtype": "float32",
123
+ "shape": [
124
+ 768
125
+ ],
126
+ "names": null
127
+ },
128
+ "observation.images.image_wrist_2_dinov3": {
129
+ "dtype": "float32",
130
+ "shape": [
131
+ 1024
132
+ ],
133
+ "names": null
134
+ },
135
+ "observation.images.image_wrist_2_siglip2": {
136
+ "dtype": "float32",
137
+ "shape": [
138
+ 768
139
+ ],
140
+ "names": null
141
+ }
142
+ },
143
+ "video_keys": [
144
+ "observation.images.image_side_1",
145
+ "observation.images.image_side_2",
146
+ "observation.images.image_wrist_1",
147
+ "observation.images.image_wrist_2"
148
+ ],
149
+ "num_shards": 64,
150
+ "index_path": "meta/sharded_index.json"
151
+ }
meta/sharded_index.json ADDED
The diff for this file is too large to render. See raw diff
 
meta/task_text_embeddings_info.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "model": "google/embeddinggemma-300m",
3
+ "dimension": 768,
4
+ "normalized": true,
5
+ "count": 24,
6
+ "file": "task_text_embeddings.parquet"
7
+ }
meta/tasks.jsonl ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"task_index": 0, "task": "Insert the 3 prong object."}
2
+ {"task_index": 1, "task": "Insert the hexagon object."}
3
+ {"task_index": 2, "task": "Insert the arch object."}
4
+ {"task_index": 3, "task": "Pick up the arch object."}
5
+ {"task_index": 4, "task": "Insert the square-circle object."}
6
+ {"task_index": 5, "task": "Insert the oval object."}
7
+ {"task_index": 6, "task": "Pick up the red object."}
8
+ {"task_index": 7, "task": "Pick up the blue object."}
9
+ {"task_index": 8, "task": "Insert the round object."}
10
+ {"task_index": 9, "task": "Pick up the star object."}
11
+ {"task_index": 10, "task": "Pick up the yellow object."}
12
+ {"task_index": 11, "task": "Insert the rectangle object."}
13
+ {"task_index": 12, "task": "Pick up the green object."}
14
+ {"task_index": 13, "task": "Pick up the oval object."}
15
+ {"task_index": 14, "task": "Pick up the purple object."}
16
+ {"task_index": 15, "task": "Insert the double-square object."}
17
+ {"task_index": 16, "task": "Pick up the brown object."}
18
+ {"task_index": 17, "task": "Pick up the 3 prong object."}
19
+ {"task_index": 18, "task": "Insert the star object."}
20
+ {"task_index": 19, "task": "Pick up the round object."}
21
+ {"task_index": 20, "task": "Pick up the rectangle object."}
22
+ {"task_index": 21, "task": "Pick up the square-circle object."}
23
+ {"task_index": 22, "task": "Pick up the double-square object."}
24
+ {"task_index": 23, "task": "Pick up the hexagon object."}