-
Notifications
You must be signed in to change notification settings - Fork 1
Expand file tree
/
Copy pathindex.html
More file actions
847 lines (780 loc) · 42.8 KB
/
Copy pathindex.html
File metadata and controls
847 lines (780 loc) · 42.8 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<meta name="description"
content="CLAMP: Contrastive Learning for 3D Multi-View Action-Conditioned Robotic Manipulation Pretraining">
<meta name="keywords" content="CLAMP, Pretraining, RoboticManipulation, MultiView, 3D">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>CLAMP: Contrastive Learning for 3D Multi-View Action-Conditioned Robotic Manipulation Pretraining</title>
<!-- Global site tag (gtag.js) - Google Analytics -->
<!-- <script async src="https://www.googletagmanager.com/gtag/js?id=G-PYVRSFMDRL"></script> -->
<!-- <script>
window.dataLayer = window.dataLayer || [];
function gtag() {
dataLayer.push(arguments);
}
gtag('js', new Date());
gtag('config', 'G-PYVRSFMDRL');
</script> -->
<link href="https://fonts.googleapis.com/css?family=Google+Sans|Noto+Sans|Castoro" rel="stylesheet">
<link rel="stylesheet" href="./static/css/bulma.min.css">
<link rel="stylesheet" href="./static/css/bulma-carousel.min.css">
<link rel="stylesheet" href="./static/css/bulma-slider.min.css">
<link rel="stylesheet" href="https://cdn.jsdelivr.net/gh/jpswalsh/academicons@1/css/academicons.min.css">
<link rel="stylesheet" href="./static/css/index.css">
<link rel="icon" href="./static/images/favicon.svg">
<script src="https://ajax.googleapis.com/ajax/libs/jquery/3.5.1/jquery.min.js"></script>
<script src="./static/js/bulma-carousel.min.js"></script>
<script src="./static/js/bulma-slider.min.js"></script>
<script src="./static/js/index.js"></script>
<script id="MathJax-script" async src="https://cdn.jsdelivr.net/npm/mathjax@3/es5/tex-mml-chtml.js"></script>
</head>
<body>
<section class="hero">
<div class="hero-body">
<div class="container is-max-desktop">
<div class="columns is-centered">
<div class="column has-text-centered">
<h1 class="title is-1 publication-title">CLAMP: Contrastive Learning for 3D Multi-View Action-Conditioned
Robotic Manipulation Pretraining</h1>
<div class="is-size-5 publication-authors">
<span class="author-block">
<a href="https://arthurliu.com/">I-Chun Arthur Liu</a><sup>1,2</sup>,</span>
<span class="author-block">
<a href="https://ieor.columbia.edu/content/krzysztof-choromanski">Krzysztof
Choromanski</a><sup>1</sup>,</span>
<span class="author-block">
<a href="https://shhuang.github.io/">Sandy Huang</a><sup>1</sup>,</span>
<span class="author-block">
<a href="https://storage.googleapis.com/www.connorschenck.com/index.html">Connor
Schenck</a><sup>1</sup></span>
</div>
<div class="is-size-5 publication-authors">
<span class="author-block"><sup>1</sup>Google DeepMind</span>
<span class="author-block"><sup>2</sup>University of Southern California</span>
</div>
<span class="author-block"><sup></sup>Robotics: Science and Systems (RSS) 2026</span>
<div class="column has-text-centered">
<div class="publication-links">
<!-- PDF Link. -->
<!-- <span class="link-block">
<a href="https://arxiv.org/pdf" class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="fas fa-file-pdf"></i>
</span>
<span>Paper</span>
</a>
</span> -->
<span class="link-block">
<a href="https://arxiv.org/abs/2602.00937" class="external-link button is-normal is-rounded is-dark">
<span class="icon">
<i class="ai ai-arxiv"></i>
</span>
<span>arXiv</span>
</a>
</span>
</div>
</div>
</div>
</div>
</div>
</div>
</section>
<section class="hero teaser">
<div class="container is-max-desktop">
<div class="hero-body">
<img src="./static/images/hero_figure_v2.webp" class="interpolation-image" alt="Hero Figure" />
<h2 class="subtitle has-text-centered">
<span class="dnerf">CLAMP</span> is a 3D pre-training framework for robotic manipulation
</h2>
</div>
</div>
</section>
<section class="hero is-light is-small">
<div class="hero-body">
<div class="container">
<div id="results-carousel" class="carousel results-carousel">
<div class="item">
<video poster autoplay muted loop playsinline height="100%">
<source src=".\static\videos\hero_teaser\open_drawers_s1.mp4" type="video/mp4">
</video>
</div>
<div class="item">
<video poster autoplay muted loop playsinline height="100%">
<source src=".\static\videos\hero_teaser\mug_on_plate_s1.mp4" type="video/mp4">
</video>
</div>
<div class="item">
<video poster autoplay muted loop playsinline height="100%">
<source src=".\static\videos\hero_teaser\plate_on_rack_s1.mp4" type="video/mp4">
</video>
</div>
<div class="item">
<video poster autoplay muted loop playsinline height="100%">
<source src=".\static\videos\hero_teaser\close_drawers_s1.mp4" type="video/mp4">
</video>
</div>
<div class="item">
<video poster autoplay muted loop playsinline height="100%">
<source src=".\static\videos\hero_teaser\recycle_cans_s1.mp4" type="video/mp4">
</video>
</div>
<div class="item" style="
box-sizing: border-box;
padding: 5px; /* moved spacing INSIDE */
margin: 0; /* remove outer spacing */
overflow: hidden;
border: 0px solid #bbb;
border-radius: 10px;
font-size: 0;
width: 100%;
">
<div style="
width: 100%;
aspect-ratio: 16 / 9;
overflow: hidden;
border-radius: 8px; /* border-radius - padding */
">
<video autoplay muted loop playsinline style="
width: 100%;
height: 100%;
object-fit: cover;
display: block;
">
<source src="./static/videos/hero_teaser/can_opener_s1.mp4" type="video/mp4">
</video>
</div>
</div>
<div class="item" style="
box-sizing: border-box;
padding: 5px; /* moved spacing INSIDE */
margin: 0; /* remove outer spacing */
overflow: hidden;
border: 0px solid #bbb;
border-radius: 10px;
font-size: 0;
width: 100%;
">
<div style="
width: 100%;
aspect-ratio: 16 / 9;
overflow: hidden;
border-radius: 8px; /* border-radius - padding */
">
<video autoplay muted loop playsinline style="
width: 100%;
height: 100%;
object-fit: cover;
display: block;
">
<source src="./static/videos/hero_teaser/pen_in_container_s1.mp4" type="video/mp4">
</video>
</div>
</div>
<div class="item" style="
box-sizing: border-box;
padding: 5px; /* moved spacing INSIDE */
margin: 0; /* remove outer spacing */
overflow: hidden;
border: 0px solid #bbb;
border-radius: 10px;
font-size: 0;
width: 100%;
">
<div style="
width: 100%;
aspect-ratio: 16 / 9;
overflow: hidden;
border-radius: 8px; /* border-radius - padding */
">
<video autoplay muted loop playsinline style="
width: 100%;
height: 100%;
object-fit: cover;
display: block;
">
<source src="./static/videos/hero_teaser/plate_on_rack_sim_s1.mp4" type="video/mp4">
</video>
</div>
</div>
<div class="item" style="
box-sizing: border-box;
padding: 5px; /* moved spacing INSIDE */
margin: 0; /* remove outer spacing */
overflow: hidden;
border: 0px solid #bbb;
border-radius: 10px;
font-size: 0;
width: 100%;
">
<div style="
width: 100%;
aspect-ratio: 16 / 9;
overflow: hidden;
border-radius: 8px; /* border-radius - padding */
">
<video autoplay muted loop playsinline style="
width: 100%;
height: 100%;
object-fit: cover;
display: block;
">
<source src="./static/videos/hero_teaser/screwdriver_s1.mp4" type="video/mp4">
</video>
</div>
</div>
<div class="item" style="
box-sizing: border-box;
padding: 5px; /* moved spacing INSIDE */
margin: 0; /* remove outer spacing */
overflow: hidden;
border: 0px solid #bbb;
border-radius: 10px;
font-size: 0;
width: 100%;
">
<div style="
width: 100%;
aspect-ratio: 16 / 9;
overflow: hidden;
border-radius: 8px; /* border-radius - padding */
">
<video autoplay muted loop playsinline style="
width: 100%;
height: 100%;
object-fit: cover;
display: block;
">
<source src="./static/videos/hero_teaser/open_drawer_sim_s1.mp4" type="video/mp4">
</video>
</div>
</div>
<div class="item" style="
box-sizing: border-box;
padding: 5px; /* moved spacing INSIDE */
margin: 0; /* remove outer spacing */
overflow: hidden;
border: 0px solid #bbb;
border-radius: 10px;
font-size: 0;
width: 100%;
">
<div style="
width: 100%;
aspect-ratio: 16 / 9;
overflow: hidden;
border-radius: 8px; /* border-radius - padding */
">
<video autoplay muted loop playsinline style="
width: 100%;
height: 100%;
object-fit: cover;
display: block;
">
<source src="./static/videos/hero_teaser/mug_on_plate_sim_s1.mp4" type="video/mp4">
</video>
</div>
</div>
</div>
</div>
</div>
</section>
<section class="section">
<div class="container is-max-desktop">
<!-- Abstract. -->
<div class="columns is-centered has-text-centered">
<div class="column is-four-fifths">
<h2 class="title is-3 section-heading is-centered">Abstract</h2>
<div class="content has-text-justified abstract-panel">
<p>
Leveraging pre-trained 2D image representations in behavior cloning policies has achieved great success
and has become a standard approach for robotic manipulation. However, such representations fail to capture
the 3D spatial information about objects and scenes that is essential for precise manipulation. In this
work, we introduce <b>C</b>ontrastive <b>L</b>earning for 3D Multi-View
<b>A</b>ction-Conditioned Robotic <b>M</b>anipulation <b>P</b>retraining (<b>CLAMP</b>), a novel 3D
pre-training framework that utilizes point clouds and robot actions. From
the merged point cloud computed from RGB-D images and camera extrinsics, we re-render multi-view
four-channel image observations with depth and 3D coordinates, including dynamic wrist views, to provide
clearer views of target objects for high-precision manipulation tasks. The pre-trained encoders learn to
associate the 3D geometric and positional information of objects with robot action patterns via
contrastive learning on large-scale simulated robot trajectories. During encoder pre-training, we
pre-train a Diffusion Policy to initialize the policy weights for fine-tuning, which is essential for
improving fine-tuning sample efficiency and performance. After pre-training, we fine-tune the policy on a
limited amount of task demonstrations using the learned image and action representations. We demonstrate
that this pre-training and fine-tuning design substantially improves learning efficiency and policy
performance on unseen tasks. Furthermore, we show that <b>CLAMP</b> outperforms state-of-the-art baselines
across six simulated tasks and five real-world tasks.
</p>
</div>
</div>
</div>
<!--/ Abstract. -->
</div>
</section>
<section class="section">
<div class="container is-max-desktop">
<!-- Method Overview -->
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-3 section-heading">Method Overview</h2>
<img src="./static/images/figure_method.webp" class="interpolation-image feature-figure"
alt="Interpolate start reference image." />
<div class="content has-text-justified">
<p>
<b>(i):</b> <b>CLAMP</b> consists of three encoders: image, action and text. The image encoder is a
Vision Transformer (ViT) that takes five multi-view observations (overhead, back-right,
front-left, wrist-left, and wrist-right) as input. These views are rendered from a merged point
cloud and include depth and 3D coordinates. The action encoder is a Transformer encoder that takes a
history of previous actions \( (a_{t−1}, \dots , a_{t−H}) \) as input. The text input
includes a general task description, each object's name and normalized position in the global coordinate
frame, and an integer indicating task progress, and is processed by a CLIP text encoder. The intuition is
to learn image and action representations that capture correlations between action patterns and robot
states as observed in images,
grounded by text that describes the spatial and temporal information about the objects and task. We also
pre-train a Diffusion Policy in parallel.
</p>
<p>
<b>(ii):</b> The Diffusion Policy is first initialized with the pre-trained weights from stage (i) and
use the frozen <b>CLAMP</b> encoders to extract image and action embeddings. These embeddings are fed into
a Transformer encoder together with RGB feature maps from ResNet-50 backbones and proprioceptive features
from the robot to predict noise of shape 50×14 for the next 50 actions \( (\epsilon_{t+1} , \dots ,
\epsilon_{t+50} ) \).
</p>
</div>
</div>
</div>
<!-- Method Overview -->
<!-- Image Representation -->
<div class="columns is-centered has-text-centered">
<div class="column is-full-width">
<h2 class="title is-3 section-heading">Image Representation</h2>
<img src="./static/images/input_image_v2.png" class="interpolation-image feature-figure"
alt="Input image for CLAMP." />
<div class="content has-text-justified">
<p>
We re-render multi-view four-channel observations from the merged point cloud that include depth and 3D
coordinates, using both fixed viewpoints and dynamic wrist views, which provide clearer observations for
high-precision manipulation tasks. Each virtual view has dimensions \( 224 \times 224 \times 4 \), and we
tile the views horizontally to form a \( 224 \times 1120 \times 4 \) input (example shown above).
</p>
</div>
<img src="./static/images/string.webp" class="interpolation-image feature-figure"
alt="STRING relative positional encoding." />
<div class="content has-text-justified">
<p>
We apply <a href="https://sites.google.com/view/string-robotics">STRING</a> relative positional encoding
to the image encoder, enabling tokens from different views
that correspond to nearby regions in 3D space to be correlated (see the image above for an illustration).
Specifically, tokens from different 2D views are concatenated, but also equipped with STRING positional
encoding mechanism. Consequently, the attention score between two highlighted tokens from two different
views will be modulated by a function that depends only on the vector \( \mathbf{r}_{i,j} \in
\mathbb{R}^{3} \) between the centers of mass of two sub-point clouds (highlighted in red and blue
correspondingly) of the visible sets of points related to those tokens.
</p>
</div>
</div>
</div>
<!-- Image Representation -->
<!-- Quantitative Results -->
<div class="columns is-centered has-text-centered">
<div class="column is-full-width">
<h2 class="title is-3 section-heading">Quantitative Results</h2>
<img src="./static/images/quantitative_results.webp" class="interpolation-image feature-figure"
alt="Quantitative results." />
<div class="content has-text-justified">
<p>
Evaluation curves comparing our method (<span style="color: green;">green</span>) to baselines across
three seeds in simulation.
All methods are trained on 4500 demonstrations for 1M environment steps and evaluated for 50
trials per checkpoint every 40K steps. <b><a href="https://aloha-unleashed.github.io/">ALOHA Unleashed</a>
with CLAMP</b> (<span style="color: green;">green</span>) shows strong performance at very
early training stages, indicating
significant gains in sample efficiency and overall performance due to pre-training.
</p>
</div>
</div>
</div>
<!-- Quantitative Results -->
<!-- Successful Rollouts -->
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-3 section-heading">Successful Rollouts of CLAMP</h2>
<p class="section-intro">
All rollout videos are shown in <b>real time</b> (1x speed).
</p>
<div class="rollouts-grid">
<article class="rollout-card">
<header class="rollout-card-header">
<h3 class="rollout-task-name">Close Drawers</h3>
<span class="rollout-badge real">Real</span>
</header>
<div class="rollout-videos">
<video muted loop playsinline webkit-playsinline preload="none" data-lazy-autoplay poster="./static/videos/successful_rollouts/close_drawers_s1.png">
<source src="./static/videos/successful_rollouts/close_drawers_s1.mp4" type="video/mp4">
</video>
<video muted loop playsinline webkit-playsinline preload="none" data-lazy-autoplay poster="./static/videos/successful_rollouts/close_drawers_s2.png">
<source src="./static/videos/successful_rollouts/close_drawers_s2.mp4" type="video/mp4">
</video>
</div>
</article>
<article class="rollout-card">
<header class="rollout-card-header">
<h3 class="rollout-task-name">Open Drawers</h3>
<span class="rollout-badge real">Real</span>
</header>
<div class="rollout-videos">
<video muted loop playsinline webkit-playsinline preload="none" data-lazy-autoplay poster="./static/videos/successful_rollouts/open_drawers_s1.png">
<source src="./static/videos/successful_rollouts/open_drawers_s1.mp4" type="video/mp4">
</video>
<video muted loop playsinline webkit-playsinline preload="none" data-lazy-autoplay poster="./static/videos/successful_rollouts/open_drawers_s2.png">
<source src="./static/videos/successful_rollouts/open_drawers_s2.mp4" type="video/mp4">
</video>
</div>
</article>
<article class="rollout-card">
<header class="rollout-card-header">
<h3 class="rollout-task-name">Mug on Plate</h3>
<span class="rollout-badge real">Real</span>
</header>
<div class="rollout-videos">
<video muted loop playsinline webkit-playsinline preload="none" data-lazy-autoplay poster="./static/videos/successful_rollouts/mug_on_plate_real_s1.png">
<source src="./static/videos/successful_rollouts/mug_on_plate_real_s1.mp4" type="video/mp4">
</video>
<video muted loop playsinline webkit-playsinline preload="none" data-lazy-autoplay poster="./static/videos/successful_rollouts/mug_on_plate_real_s2.png">
<source src="./static/videos/successful_rollouts/mug_on_plate_real_s2.mp4" type="video/mp4">
</video>
</div>
</article>
<article class="rollout-card">
<header class="rollout-card-header">
<h3 class="rollout-task-name">Plate on Rack</h3>
<span class="rollout-badge real">Real</span>
</header>
<div class="rollout-videos">
<video muted loop playsinline webkit-playsinline preload="none" data-lazy-autoplay poster="./static/videos/successful_rollouts/plate_on_rack_real_s1.png">
<source src="./static/videos/successful_rollouts/plate_on_rack_real_s1.mp4" type="video/mp4">
</video>
<video muted loop playsinline webkit-playsinline preload="none" data-lazy-autoplay poster="./static/videos/successful_rollouts/plate_on_rack_real_s2.png">
<source src="./static/videos/successful_rollouts/plate_on_rack_real_s2.mp4" type="video/mp4">
</video>
</div>
</article>
<article class="rollout-card">
<header class="rollout-card-header">
<h3 class="rollout-task-name">Recycle Cans</h3>
<span class="rollout-badge real">Real</span>
</header>
<div class="rollout-videos">
<video muted loop playsinline webkit-playsinline preload="none" data-lazy-autoplay poster="./static/videos/successful_rollouts/recycle_cans_s1.png">
<source src="./static/videos/successful_rollouts/recycle_cans_s1.mp4" type="video/mp4">
</video>
<video muted loop playsinline webkit-playsinline preload="none" data-lazy-autoplay poster="./static/videos/successful_rollouts/recycle_cans_s2.png">
<source src="./static/videos/successful_rollouts/recycle_cans_s2.mp4" type="video/mp4">
</video>
</div>
</article>
<article class="rollout-card">
<header class="rollout-card-header">
<h3 class="rollout-task-name">Drawer Open</h3>
<span class="rollout-badge sim">Sim</span>
</header>
<div class="rollout-videos">
<video muted loop playsinline webkit-playsinline preload="none" data-lazy-autoplay poster="./static/videos/successful_rollouts/drawer_open_s1.png">
<source src="./static/videos/successful_rollouts/drawer_open_s1.mp4" type="video/mp4">
</video>
<video muted loop playsinline webkit-playsinline preload="none" data-lazy-autoplay poster="./static/videos/successful_rollouts/drawer_open_s2.png">
<source src="./static/videos/successful_rollouts/drawer_open_s2.mp4" type="video/mp4">
</video>
</div>
</article>
<article class="rollout-card">
<header class="rollout-card-header">
<h3 class="rollout-task-name">Mug on Plate</h3>
<span class="rollout-badge sim">Sim</span>
</header>
<div class="rollout-videos">
<video muted loop playsinline webkit-playsinline preload="none" data-lazy-autoplay poster="./static/videos/successful_rollouts/mug_on_plate_sim_s1.png">
<source src="./static/videos/successful_rollouts/mug_on_plate_sim_s1.mp4" type="video/mp4">
</video>
<video muted loop playsinline webkit-playsinline preload="none" data-lazy-autoplay poster="./static/videos/successful_rollouts/mug_on_plate_sim_s2.png">
<source src="./static/videos/successful_rollouts/mug_on_plate_sim_s2.mp4" type="video/mp4">
</video>
</div>
</article>
<article class="rollout-card">
<header class="rollout-card-header">
<h3 class="rollout-task-name">Pen in Container</h3>
<span class="rollout-badge sim">Sim</span>
</header>
<div class="rollout-videos">
<video muted loop playsinline webkit-playsinline preload="none" data-lazy-autoplay poster="./static/videos/successful_rollouts/pen_in_container_s1.png">
<source src="./static/videos/successful_rollouts/pen_in_container_s1.mp4" type="video/mp4">
</video>
<video muted loop playsinline webkit-playsinline preload="none" data-lazy-autoplay poster="./static/videos/successful_rollouts/pen_in_container_s2.png">
<source src="./static/videos/successful_rollouts/pen_in_container_s2.mp4" type="video/mp4">
</video>
</div>
</article>
<article class="rollout-card">
<header class="rollout-card-header">
<h3 class="rollout-task-name">Can Opener in Caddy</h3>
<span class="rollout-badge sim">Sim</span>
</header>
<div class="rollout-videos">
<video muted loop playsinline webkit-playsinline preload="none" data-lazy-autoplay poster="./static/videos/successful_rollouts/can_opener_s1.png">
<source src="./static/videos/successful_rollouts/can_opener_s1.mp4" type="video/mp4">
</video>
<video muted loop playsinline webkit-playsinline preload="none" data-lazy-autoplay poster="./static/videos/successful_rollouts/can_opener_s2.png">
<source src="./static/videos/successful_rollouts/can_opener_s2.mp4" type="video/mp4">
</video>
</div>
</article>
<article class="rollout-card">
<header class="rollout-card-header">
<h3 class="rollout-task-name">Screwdriver in Caddy</h3>
<span class="rollout-badge sim">Sim</span>
</header>
<div class="rollout-videos">
<video muted loop playsinline webkit-playsinline preload="none" data-lazy-autoplay poster="./static/videos/successful_rollouts/screwdriver_s1.png">
<source src="./static/videos/successful_rollouts/screwdriver_s1.mp4" type="video/mp4">
</video>
<video muted loop playsinline webkit-playsinline preload="none" data-lazy-autoplay poster="./static/videos/successful_rollouts/screwdriver_s2.png">
<source src="./static/videos/successful_rollouts/screwdriver_s2.mp4" type="video/mp4">
</video>
</div>
</article>
<article class="rollout-card">
<header class="rollout-card-header">
<h3 class="rollout-task-name">Plate on Rack</h3>
<span class="rollout-badge sim">Sim</span>
</header>
<div class="rollout-videos">
<video muted loop playsinline webkit-playsinline preload="none" data-lazy-autoplay poster="./static/videos/successful_rollouts/plate_on_rack_sim_s1.png">
<source src="./static/videos/successful_rollouts/plate_on_rack_sim_s1.mp4" type="video/mp4">
</video>
<video muted loop playsinline webkit-playsinline preload="none" data-lazy-autoplay poster="./static/videos/successful_rollouts/plate_on_rack_sim_s2.png">
<source src="./static/videos/successful_rollouts/plate_on_rack_sim_s2.mp4" type="video/mp4">
</video>
</div>
</article>
</div>
</div>
</div>
<!-- Successful Rollouts -->
<!-- Failures -->
<div class="columns is-centered">
<div class="column is-full-width">
<h2 class="title is-3 section-heading">Failures</h2>
<p class="section-intro">
Failure cases of <b>CLAMP</b> across different tasks.
</p>
<div class="failures-panel">
<div class="failure-selector-row">
<label class="failure-selector-label" for="failure-selector">
Choose a task to view failure cases of <span class="dnerf">CLAMP</span>:
</label>
<div class="failure-selector-control">
<select id="failure-selector">
<option value="close_drawers_failures">Close Drawers (Real)</option>
<option value="open_drawers_failures">Open Drawers (Real)</option>
<option value="mug_on_plate_real_failures">Mug on Plate (Real)</option>
<option value="plate_on_rack_real_failures">Plate on Rack (Real)</option>
<option value="recycle_cans_failures">Recycle Cans (Real)</option>
<option value="drawer_open_failures">Drawer Open (Sim)</option>
<option value="mug_on_plate_sim_failures">Mug on Plate (Sim)</option>
<option value="pen_in_container_failures">Pen in Container (Sim)</option>
<option value="can_opener_failures">Can Opener in Caddy (Sim)</option>
<option value="screwdriver_failures">Screwdriver in Caddy (Sim)</option>
<option value="plate_on_rack_sim_failures">Plate on Rack (Sim)</option>
</select>
</div>
</div>
<div class="columns is-centered">
<div class="column">
<div id="baseline-container" class="has-text-centered">
<!-- Close Drawers -->
<div id="close_drawers_failures" class="failure-class failure-card">
<header class="failure-card-header">
<h3 class="rollout-task-name">Close Drawers</h3>
<span class="rollout-badge real">Real</span>
</header>
<div class="failure-videos">
<video muted loop playsinline webkit-playsinline preload="none" poster="./static/videos/failures/close_drawers_f1.png">
<source src="./static/videos/failures/close_drawers_f1.mp4" type="video/mp4">
</video>
<video muted loop playsinline webkit-playsinline preload="none" poster="./static/videos/failures/close_drawers_f2.png">
<source src="./static/videos/failures/close_drawers_f2.mp4" type="video/mp4">
</video>
</div>
</div>
<!-- Open Drawers -->
<div id="open_drawers_failures" class="failure-class failure-card" style="display: none;">
<header class="failure-card-header">
<h3 class="rollout-task-name">Open Drawers</h3>
<span class="rollout-badge real">Real</span>
</header>
<div class="failure-videos">
<video muted loop playsinline webkit-playsinline preload="none" poster="./static/videos/failures/open_drawers_f1.png">
<source src="./static/videos/failures/open_drawers_f1.mp4" type="video/mp4">
</video>
<video muted loop playsinline webkit-playsinline preload="none" poster="./static/videos/failures/open_drawers_f2.png">
<source src="./static/videos/failures/open_drawers_f2.mp4" type="video/mp4">
</video>
</div>
</div>
<!-- Mug on Plate (Real) -->
<div id="mug_on_plate_real_failures" class="failure-class failure-card" style="display: none;">
<header class="failure-card-header">
<h3 class="rollout-task-name">Mug on Plate</h3>
<span class="rollout-badge real">Real</span>
</header>
<div class="failure-videos">
<video muted loop playsinline webkit-playsinline preload="none" poster="./static/videos/failures/mug_on_plate_real_f1.png">
<source src="./static/videos/failures/mug_on_plate_real_f1.mp4" type="video/mp4">
</video>
<video muted loop playsinline webkit-playsinline preload="none" poster="./static/videos/failures/mug_on_plate_real_f2.png">
<source src="./static/videos/failures/mug_on_plate_real_f2.mp4" type="video/mp4">
</video>
</div>
</div>
<!-- Plate on Rack (Real) -->
<div id="plate_on_rack_real_failures" class="failure-class failure-card" style="display: none;">
<header class="failure-card-header">
<h3 class="rollout-task-name">Plate on Rack</h3>
<span class="rollout-badge real">Real</span>
</header>
<div class="failure-videos">
<video muted loop playsinline webkit-playsinline preload="none" poster="./static/videos/failures/plate_on_rack_real_f1.png">
<source src="./static/videos/failures/plate_on_rack_real_f1.mp4" type="video/mp4">
</video>
<video muted loop playsinline webkit-playsinline preload="none" poster="./static/videos/failures/plate_on_rack_real_f2.png">
<source src="./static/videos/failures/plate_on_rack_real_f2.mp4" type="video/mp4">
</video>
</div>
</div>
<!-- Recycle Cans -->
<div id="recycle_cans_failures" class="failure-class failure-card" style="display: none;">
<header class="failure-card-header">
<h3 class="rollout-task-name">Recycle Cans</h3>
<span class="rollout-badge real">Real</span>
</header>
<div class="failure-videos">
<video muted loop playsinline webkit-playsinline preload="none" poster="./static/videos/failures/recycle_cans_f1.png">
<source src="./static/videos/failures/recycle_cans_f1.mp4" type="video/mp4">
</video>
<video muted loop playsinline webkit-playsinline preload="none" poster="./static/videos/failures/recycle_cans_f2.png">
<source src="./static/videos/failures/recycle_cans_f2.mp4" type="video/mp4">
</video>
</div>
</div>
<!-- Drawer Open -->
<div id="drawer_open_failures" class="failure-class failure-card" style="display: none;">
<header class="failure-card-header">
<h3 class="rollout-task-name">Drawer Open</h3>
<span class="rollout-badge sim">Sim</span>
</header>
<div class="failure-videos single">
<video muted loop playsinline webkit-playsinline preload="none" poster="./static/videos/failures/drawer_open_f1.png">
<source src="./static/videos/failures/drawer_open_f1.mp4" type="video/mp4">
</video>
</div>
</div>
<!-- Mug on Plate (Sim) -->
<div id="mug_on_plate_sim_failures" class="failure-class failure-card" style="display: none;">
<header class="failure-card-header">
<h3 class="rollout-task-name">Mug on Plate</h3>
<span class="rollout-badge sim">Sim</span>
</header>
<div class="failure-videos">
<video muted loop playsinline webkit-playsinline preload="none" poster="./static/videos/failures/mug_on_plate_sim_f1.png">
<source src="./static/videos/failures/mug_on_plate_sim_f1.mp4" type="video/mp4">
</video>
<video muted loop playsinline webkit-playsinline preload="none" poster="./static/videos/failures/mug_on_plate_sim_f2.png">
<source src="./static/videos/failures/mug_on_plate_sim_f2.mp4" type="video/mp4">
</video>
</div>
</div>
<!-- Pen in Container -->
<div id="pen_in_container_failures" class="failure-class failure-card" style="display: none;">
<header class="failure-card-header">
<h3 class="rollout-task-name">Pen in Container</h3>
<span class="rollout-badge sim">Sim</span>
</header>
<div class="failure-videos">
<video muted loop playsinline webkit-playsinline preload="none" poster="./static/videos/failures/pen_in_container_f1.png">
<source src="./static/videos/failures/pen_in_container_f1.mp4" type="video/mp4">
</video>
<video muted loop playsinline webkit-playsinline preload="none" poster="./static/videos/failures/pen_in_container_f2.png">
<source src="./static/videos/failures/pen_in_container_f2.mp4" type="video/mp4">
</video>
</div>
</div>
<!-- Can Opener in Caddy -->
<div id="can_opener_failures" class="failure-class failure-card" style="display: none;">
<header class="failure-card-header">
<h3 class="rollout-task-name">Can Opener in Caddy</h3>
<span class="rollout-badge sim">Sim</span>
</header>
<div class="failure-videos">
<video muted loop playsinline webkit-playsinline preload="none" poster="./static/videos/failures/can_opener_f1.png">
<source src="./static/videos/failures/can_opener_f1.mp4" type="video/mp4">
</video>
<video muted loop playsinline webkit-playsinline preload="none" poster="./static/videos/failures/can_opener_f2.png">
<source src="./static/videos/failures/can_opener_f2.mp4" type="video/mp4">
</video>
</div>
</div>
<!-- Screwdriver in Caddy -->
<div id="screwdriver_failures" class="failure-class failure-card" style="display: none;">
<header class="failure-card-header">
<h3 class="rollout-task-name">Screwdriver in Caddy</h3>
<span class="rollout-badge sim">Sim</span>
</header>
<div class="failure-videos single">
<video muted loop playsinline webkit-playsinline preload="none" poster="./static/videos/failures/screwdriver_f1.png">
<source src="./static/videos/failures/screwdriver_f1.mp4" type="video/mp4">
</video>
</div>
</div>
<!-- Plate on Rack (Sim) -->
<div id="plate_on_rack_sim_failures" class="failure-class failure-card" style="display: none;">
<header class="failure-card-header">
<h3 class="rollout-task-name">Plate on Rack</h3>
<span class="rollout-badge sim">Sim</span>
</header>
<div class="failure-videos">
<video muted loop playsinline webkit-playsinline preload="none" poster="./static/videos/failures/plate_on_rack_sim_f1.png">
<source src="./static/videos/failures/plate_on_rack_sim_f1.mp4" type="video/mp4">
</video>
<video muted loop playsinline webkit-playsinline preload="none" poster="./static/videos/failures/plate_on_rack_sim_f2.png">
<source src="./static/videos/failures/plate_on_rack_sim_f2.mp4" type="video/mp4">
</video>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
<!-- Failures -->
</div>
</section>
<footer class="footer">
<div class="container">
<div class="columns is-centered">
<div class="column is-8">
<div class="content">
<p> Website template modified from <a href="https://github.com/nerfies/nerfies.github.io">NeRFies</a>.</p>
This website is licensed under a <a rel="license"
href="http://creativecommons.org/licenses/by-sa/4.0/">Creative
Commons Attribution-ShareAlike 4.0 International License</a>.
</p>
</div>
</div>
</div>
</div>
</footer>
</body>
</html>