File size: 2,289 Bytes
3b351d3
 
 
 
 
 
 
 
 
 
 
 
5e3a34d
 
3b351d3
 
 
 
 
 
 
 
19a51be
3b351d3
 
 
 
 
 
67323bc
3b351d3
67323bc
 
 
3b351d3
67323bc
 
3b351d3
67323bc
 
609c065
67323bc
609c065
3b351d3
 
 
 
e0ac04d
 
 
38ddb80
 
e0ac04d
 
 
 
 
 
ac779ca
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
---
license: cc-by-nc-4.0
task_categories:
- robotics
- video-classification
- object-detection
tags:
- ego4d
- pov
- manipulation
- fhd
- 4k
- egocentric
- humanoid
size_categories:
- 1G<10G
---

# High-Fidelity Robotics Manipulation Samples (POV)

**Producer:** [Trainthem AI](https://trainthemai.com)  
**Location:** Argentina  
**Hardware:** Smartphone (Headstrap) / 60fps
**Resolution:** FHD / 4K Samples

## Dataset Summary
This dataset contains high-resolution sample footage of fine motor skills, fluid manipulation, and daily living tasks, designed for training VLA (Vision-Language-Action) models.

**Samples Included in this Repo:**

1. **Fluid Manipulation:**
   - `making_orange_juice.mp4` (Squeezing, pouring, dynamic fluids)
   - `preparing_mate_argentinian_drink.mp4` (Precision pouring of hot water, cultural task)

2. **Complex Cooking:**
   - `preparing_salad.mp4` (Mixing ingredients, opening cans, fine motor skills)

3. **Deformable Objects:**
   - `storing_kitchen_towels.mp4` (Folding and storing cloth/laundry)

4. **Household Maintenance & Sorting:**
   - `organizing_cutlery_drawer.mp4` (Sorting cutlery, organizing chaotic drawer)
   - `watering_plants.mp4` (General household maintenance, tool use)

## Why this data is different
Unlike crowdsourced data, this footage is hardware-stabilized and captured by trained operators, minimizing motion blur and rolling shutter artifacts.

## Ready to scale your data pipeline?
This repository is a sample of our production quality. We don't just sell static libraries; we act as a **Continuous Data Partner** for robotics teams.

Note: This dataset demonstrates our visual capabilities. We are fully equipped to integrate teleoperation hardware (VR gloves, exoskeletons) to capture synchronized action-state data.

**Partner with Trainthem AI to build your custom dataset:**
* **Continuous Production:** We can supply **150+ hours/week** of fresh data (scalable).
* **Custom Scenarios:** You define the tasks (e.g., specific tools, kitchen layouts, lighting), and we film them.
* **Managed Lab:** All data is hardware-stabilized and QA-verified.

**Start a Pilot Batch:** [diego.pousa@trainthemai.com](mailto:diego.pousa@trainthemai.com)
**Website:** [https://trainthemai.com/contact](https://trainthemai.com/contact)