Add dataset card, paper link, and sample usage
#1
by nielsr HF Staff - opened
- .gitattributes +0 -2
- README.md +0 -12
.gitattributes
CHANGED
|
@@ -58,5 +58,3 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 58 |
# Video files - compressed
|
| 59 |
*.mp4 filter=lfs diff=lfs merge=lfs -text
|
| 60 |
*.webm filter=lfs diff=lfs merge=lfs -text
|
| 61 |
-
*.ply filter=lfs diff=lfs merge=lfs -text
|
| 62 |
-
*.splat filter=lfs diff=lfs merge=lfs -text
|
|
|
|
| 58 |
# Video files - compressed
|
| 59 |
*.mp4 filter=lfs diff=lfs merge=lfs -text
|
| 60 |
*.webm filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
README.md
CHANGED
|
@@ -10,18 +10,6 @@ task_categories:
|
|
| 10 |
|
| 11 |
**Deform360** is a massive multi-view visuotactile dataset for deformable-object research, featuring 198 daily-life objects, 1,980 interaction sequences, and over 215 hours of observations from 41 surround-view cameras and bimanual tactile grippers to capture both global motion and contact-induced local deformations.
|
| 12 |
|
| 13 |
-
## Where the data lives
|
| 14 |
-
|
| 15 |
-
| Repository | Contents |
|
| 16 |
-
| --- | --- |
|
| 17 |
-
| [`brownu/deform360`](https://huggingface.co/datasets/brownu/deform360) (this repo) | `raw/` — the untouched capture directories, one per object |
|
| 18 |
-
| [`brownu/deform360_processed`](https://huggingface.co/datasets/brownu/deform360_processed) | `processed/` — undistorted cameras, refined masks, tracking, Gaussian splats and point clouds |
|
| 19 |
-
|
| 20 |
-
The processed stage is published in its own repository, in a packed per-episode
|
| 21 |
-
layout rather than as loose files. Kept unpacked it runs to well over a million
|
| 22 |
-
files, far past the Hub's [per-repository limits](https://huggingface.co/docs/hub/storage-limits)
|
| 23 |
-
(under 100k files per repo), which degrades every commit and download against it.
|
| 24 |
-
|
| 25 |
## Installation
|
| 26 |
|
| 27 |
To load and interact with the dataset, install the helper library:
|
|
|
|
| 10 |
|
| 11 |
**Deform360** is a massive multi-view visuotactile dataset for deformable-object research, featuring 198 daily-life objects, 1,980 interaction sequences, and over 215 hours of observations from 41 surround-view cameras and bimanual tactile grippers to capture both global motion and contact-induced local deformations.
|
| 12 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13 |
## Installation
|
| 14 |
|
| 15 |
To load and interact with the dataset, install the helper library:
|