DEV Community

Harpreet Sahota profile picture

Harpreet Sahota

Hi. This blog covers physical AI, robotics, computer vision, and autonomous systems. Hope you enjoy and follow on YouTube!

EmbeddingGemma 2 in FiftyOne: Search Video by Text, Sound, or Your Webcam

EmbeddingGemma 2 in FiftyOne: Search Video by Text, Sound, or Your Webcam

Comments
15 min read
One Camera Isn’t Enough: 9 Robotics Datasets You Can Play on One Timeline

One Camera Isn’t Enough: 9 Robotics Datasets You Can Play on One Timeline

Comments
12 min read
Robots Can See. They Still Can't Feel: 8 Tactile LeRobot Datasets in FiftyOne

Robots Can See. They Still Can't Feel: 8 Tactile LeRobot Datasets in FiftyOne

Comments
25 min read
How to Find Edited Copies of Images: Image Copy Detection with pHash, CLIP, and FiftyOne

How to Find Edited Copies of Images: Image Copy Detection with pHash, CLIP, and FiftyOne

Comments
16 min read
You Can’t Watch 10,000 Episodes. Measure These Instead: Curation Metrics for LeRobot Datasets

You Can’t Watch 10,000 Episodes. Measure These Instead: Curation Metrics for LeRobot Datasets

Comments
35 min read
Your Robot Data Needs Custom Tools. Build Them as FiftyOne Plugins

Your Robot Data Needs Custom Tools. Build Them as FiftyOne Plugins

Comments
16 min read
50 Embodiments, 497 Episodes, One Indexed Dataset: Curating the LeRobot Community Dataset with FiftyOne

50 Embodiments, 497 Episodes, One Indexed Dataset: Curating the LeRobot Community Dataset with FiftyOne

Comments
24 min read
Real vs. Synthetic LeRobot Datasets: 5 Sets and the Sim-to-Real Evidence

Real vs. Synthetic LeRobot Datasets: 5 Sets and the Sim-to-Real Evidence

Comments
13 min read
5 Humanoid and Bimanual LeRobot Datasets, Compared by State and Action Space

5 Humanoid and Bimanual LeRobot Datasets, Compared by State and Action Space

Comments
13 min read
How to Import LeRobot Datasets into FiftyOne Using Agent Skills

How to Import LeRobot Datasets into FiftyOne Using Agent Skills

Comments
8 min read
From Franka Arms to Humanoids: 11 LeRobot v3.0 Datasets for VLA Training in 2026

From Franka Arms to Humanoids: 11 LeRobot v3.0 Datasets for VLA Training in 2026

Comments
26 min read
The Smoothest Episode in the Batch Was Also a Failure: Motion Smoothness vs. Task Success in RoboLab-EgoX Policy Rollouts

The Smoothest Episode in the Batch Was Also a Failure: Motion Smoothness vs. Task Success in RoboLab-EgoX Policy Rollouts

Comments
14 min read
Seven Downloads, One MCAP: Authoring a Full Aria Digital Twin Episode for FiftyOne

Seven Downloads, One MCAP: Authoring a Full Aria Digital Twin Episode for FiftyOne

Comments
18 min read
What Actually Breaks When You Author MCAP by Hand

What Actually Breaks When You Author MCAP by Hand

Comments
10 min read
Your Photogrammetry Dataset Is Secretly a Robotics Log: Converting PIVOT to MCAP Episodes in FiftyOne

Your Photogrammetry Dataset Is Secretly a Robotics Log: Converting PIVOT to MCAP Episodes in FiftyOne

Comments
13 min read
10 Multimodal MCAP Datasets in 2026: SLAM Benchmarks, Sensor Fusion, and Ground-Truth Object Detection in FiftyOne

10 Multimodal MCAP Datasets in 2026: SLAM Benchmarks, Sensor Fusion, and Ground-Truth Object Detection in FiftyOne

Comments
21 min read
A Complete Annotation Workflow with FiftyOne

A Complete Annotation Workflow with FiftyOne

Comments
33 min read
Everyone Is Scoring Robot Data, Nobody Is Looking at It: Mapping the Robot Data Quality Stack

Everyone Is Scoring Robot Data, Nobody Is Looking at It: Mapping the Robot Data Quality Stack

Comments
12 min read
Smooth Isn't the Same as Correct: Triaging Robot Episode Quality with FiftyOne

Smooth Isn't the Same as Correct: Triaging Robot Episode Quality with FiftyOne

Comments
21 min read
Vibe-Checking Qwen3.8-Max on Hard Visual Grounding Tasks

Vibe-Checking Qwen3.8-Max on Hard Visual Grounding Tasks

Comments
27 min read
6,500 Robot Demos, One Bad Grasp: Finding Bad Demonstrations in LIBERO with FiftyOne

6,500 Robot Demos, One Bad Grasp: Finding Bad Demonstrations in LIBERO with FiftyOne

Comments
7 min read
Data Curation, the Voxel51 Agent, and In-App Annotation: A CropAndWeed Walkthrough

Data Curation, the Voxel51 Agent, and In-App Annotation: A CropAndWeed Walkthrough

Comments
17 min read
The Robot Grabbed the Wrong Block

The Robot Grabbed the Wrong Block

Comments
15 min read
A Drone Remembers Everything It Saw. Now You Can Watch It Happen Again.

A Drone Remembers Everything It Saw. Now You Can Watch It Happen Again.

Comments
12 min read
MiniMax Just Raised Another ~$2B. Its Frontier Model Now Runs Inside The FiftyOne App

MiniMax Just Raised Another ~$2B. Its Frontier Model Now Runs Inside The FiftyOne App

Comments
10 min read
Real Roads Don’t Have Four Wheels: Exploring MOTOR in FiftyOne

Real Roads Don’t Have Four Wheels: Exploring MOTOR in FiftyOne

Comments
12 min read
Top Agriculture Computer Vision Datasets in 2026: Crop & Weed Segmentation, Plant Phenotyping, 3D Point Clouds, and Precision Farming Data

Top Agriculture Computer Vision Datasets in 2026: Crop & Weed Segmentation, Plant Phenotyping, 3D Point Clouds, and Precision Farming Data

Comments
19 min read
Visualize Rare Driving Scenarios with FiftyOne: Inside the KITScenes-LongTail Dataset

Visualize Rare Driving Scenarios with FiftyOne: Inside the KITScenes-LongTail Dataset

Comments
10 min read
Nine Cameras, Seven LiDARs, Three Radars, One Group ID: Exploring KITScenes Multimodal in FiftyOne

Nine Cameras, Seven LiDARs, Three Radars, One Group ID: Exploring KITScenes Multimodal in FiftyOne

Comments
14 min read
Some Data You Can't Wait For. Some You Can't Survive.

Some Data You Can't Wait For. Some You Can't Survive.

Comments
14 min read
Beyond Default Boxes: Custom Annotation Styles for Computer Vision Datasets in FiftyOne

Beyond Default Boxes: Custom Annotation Styles for Computer Vision Datasets in FiftyOne

Comments
7 min read
Finding a Needle in the Desert

Finding a Needle in the Desert

Comments
7 min read
LightOnOCR-2: A Compact AI Model Revolutionizing Document OCR

LightOnOCR-2: A Compact AI Model Revolutionizing Document OCR

Comments
4 min read
Building Production-Ready AI Document Understanding Pipelines with GLM-OCR

Building Production-Ready AI Document Understanding Pipelines with GLM-OCR

Comments
13 min read
C-RADIOv4: A Distilled Vision Foundation Model for FiftyOne

C-RADIOv4: A Distilled Vision Foundation Model for FiftyOne

Comments
8 min read
Why Vision-Only Imitation Fails at Delicate Contact: Low-Cost Dexterous Teleoperation & Tactile Policies

Why Vision-Only Imitation Fails at Delicate Contact: Low-Cost Dexterous Teleoperation & Tactile Policies

1
Comments
5 min read
loading...