datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
Mobile-O-Post-Train
Mobile-O Post-Training Data
Unified Multimodal Post-Training · ~105K Quadruplet Samples
📌 Overview
This dataset is used for Stage 3: Unified Multimodal Post-Training of Mobile-O, a unified multimodal model for on-device understanding and generation.
The goal of this stage is to jointly improve both image generation and visual understanding through a multi-task objective using quadruplet samples.
📊 Dataset Format
Each sample is a quadruplet consisting of:… See the full description on the dataset page: https://huggingface.co/datasets/Amshaker/Mobile-O-Post-Train.MobileManiBenchMobile-R1
Dataset Card for Mobile-R1
Dataset Structure
images/: All the screenshots
data.jsonl: The trajectory data
Data Fields
All screenshots are stored in the images/ directory.
We describe the structure of a single trajectory entry from the file data.jsonl, which contains the full interaction trajectories and action history.
app_name: String. The name of the mobile application (e.g., "闲鱼" / Xianyu) where the task is performed.
trajectory_length: Integer. Number of… See the full description on the dataset page: https://huggingface.co/datasets/Mobile-R1/Mobile-R1.Mobile-O-Pre-Train
Mobile-O Pre-Training Data
Cross-Modal Alignment · 9M Text-Image Pairs
📌 Overview
This dataset is used for Stage 1: Cross-Modal Alignment pre-training of Mobile-O, a unified multimodal model for on-device understanding and generation.
The goal of this stage is to align the DiT diffusion decoder and Mobile Conditioning Projector (MCP) with the frozen VLM backbone using large-scale text-image pairs.
📊 Dataset Composition
Source
Samples
Description… See the full description on the dataset page: https://huggingface.co/datasets/Amshaker/Mobile-O-Pre-Train.MobileMem
MobileMem
MobileMem is the dataset release for MobileMem: On-Device Memory for Continually Evolving Agents, a benchmark for evaluating long-term memory systems in realistic mobile-assistant scenarios.
MobileMem models a continually evolving personal assistant that receives heterogeneous, temporally ordered interactions from users, assistants, and mobile applications. The benchmark evaluates whether a memory system can preserve, update, retrieve, and reason over personal… See the full description on the dataset page: https://huggingface.co/datasets/zjunlp/MobileMem.aloha_mobile_cabinetThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "aloha",
"total_episodes": 85,
"total_frames": 127500,
"total_tasks": 1,
"chunks_size": 1000,
"fps": 50,
"splits": {
"train": "0:85"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path": "videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4",
"features": {… See the full description on the dataset page: https://huggingface.co/datasets/lerobot/aloha_mobile_cabinet.MobileViews
🚀 MobileViews: A Large-Scale Mobile GUI Dataset
MobileViews is a large-scale dataset designed to support research on mobile agents and mobile user interface (UI) analysis. The first release, MobileViews-600K, includes over 600,000 mobile UI screenshot-view hierarchy (VH) pairs collected from over 20,000 apps on the Google Play Store. This dataset is based on the DroidBot, which we have optimized for large-scale data collection, capturing more comprehensive interaction details while… See the full description on the dataset page: https://huggingface.co/datasets/mllmTeam/MobileViews.mobile-aloha
Mobile Aloha Dataset Mirror
link to paper
Uni-GUI-Mobile
Uni-GUI-Mobile
A mobile GUI agent trajectory dataset, used as part of the training data for UI-MOPD (Multi-platform On-Policy Distillation for Continual GUI Agent Learning).
Dataset Statistics
Metric
Value
Trajectories
871
Total Steps
~14K
Platform
Mobile (1080x2400)
Applications
10 categories
Coordinate System
Normalized to [0, 999]
Applications
App
Description
calendar
Calendar event management
chrome
Mobile… See the full description on the dataset page: https://huggingface.co/datasets/UI-MOPD/Uni-GUI-Mobile.MobileSAMMobileBench-v2
Mobile-Bench v2
Dataset Description:
anbu-instruction.zip is a subset of the ambiguous category split in Mobile-Bench-v2.
simple_tasks_sample.json is a sample of the common-simple dataset, and simple_normal_tasks.json is the complete dataset for common-simple.
complex_tasks_sample.json is a sample of the common-complex dataset, and complex_normal_tasks.json is the complete dataset for common-complex.
These two datasets should be used with the Mobile3M dataset.… See the full description on the dataset page: https://huggingface.co/datasets/xwk123/MobileBench-v2.aloha_mobile_shrimpThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "aloha",
"total_episodes": 18,
"total_frames": 67500,
"total_tasks": 1,
"chunks_size": 1000,
"fps": 50,
"splits": {
"train": "0:18"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path": "videos/{video_key}/chunk-{chunk_index:03d}/file-{file_index:03d}.mp4",
"features": {… See the full description on the dataset page: https://huggingface.co/datasets/lerobot/aloha_mobile_shrimp.Mobile3M
MobileVLM
Android Cuttlefish Emulators
Pre-training Tasks
Copyright (C) 2024 Xiaomi Corporation.
The source code included in this project is licensed under the Apache 2.0 license.
The dataset included in this project is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) license.
1. License
The dataset of this project is licensed under the Creative Commons… See the full description on the dataset page: https://huggingface.co/datasets/xwk123/Mobile3M.Mobile-R1
Dataset Card for Mobile-R1
Dataset Structure
images/: All the screenshots
data.jsonl: The trajectory data
Data Instances
Data Fields
All screenshots are stored in the images/ directory.
We describe the structure of a single trajectory entry from the file data.jsonl, which contains the full interaction trajectories and action history.
app_name: String. The name of the mobile application (e.g., "闲鱼" / Xianyu) where the task is performed.… See the full description on the dataset page: https://huggingface.co/datasets/PG23/Mobile-R1.MobileWorld-Eval-Results
MobileWorld-Eval-Results
Evaluation results of the UI-MOPD trained model (Qwen3-VL-8B-Thinking) on a mobile agent benchmark. Contains full execution trajectories including screenshots, marked action visualizations, and task outcomes across 117 diverse mobile tasks.
Evaluation Summary
Metric
Value
Model
Qwen3-VL-8B-Thinking
Total Tasks
117
Successful
12
Success Rate
10.3%
Action Space
mobile_use
Max Steps
50
Avg Steps
32.5
Total Steps
~3.8K… See the full description on the dataset page: https://huggingface.co/datasets/UI-MOPD/MobileWorld-Eval-Results.MobilePoG
MobilePoG: Benchmark for Mobile Point-of-Gaze Calibration.
Dataset Summary
Figure 1: Collection procedure of the MobilePoG dataset.
MobilePoG was introduced in the paper "Pose-Robust Calibration Strategy for Point-of-Gaze Estimation on Mobile Phones" (BMVC 2025). It was collected and curated to support research on point-of-gaze estimation and calibration, with the primary goal of enabling valid evaluation of personalized… See the full description on the dataset page: https://huggingface.co/datasets/HomieZ/MobilePoG.MobileSpectralCCDatasetmobile-actions
Mobile Actions: A Dataset for On-Device Function Calling
The dataset contains conversational traces designed to train lightweight models (such as FunctionGemma 270M) to translate natural language instructions into executable function calls for Android OS system tools.
Dataset Format
The dataset is provided in JSONL format. Each line represents a data sample. The
dataset is pre-split into training and evaluation sets. This distinction is
denoted by the metadata field… See the full description on the dataset page: https://huggingface.co/datasets/google/mobile-actions.mobileworldbenchAgiBot-g1_mobile_accessory_storage_box_a
AgiBot-g1_mobile_accessory_storage_box_a
📋 Overview
This dataset uses an extended format based on LeRobot and is fully compatible with LeRobot.
Robot Type: ruantong_a2d
| Codebase Version: v2.1
End-Effector Type: two_finger_gripper
🏠 Scene Types
This dataset covers the following scene types:
home
🤖 Atomic Actions
This dataset includes the following atomic actions:
grasp
place
pick
📊 Dataset Statistics
Metric
Value
Total… See the full description on the dataset page: https://huggingface.co/datasets/RoboCOIN/AgiBot-g1_mobile_accessory_storage_box_a.aloha_mobile_wipe_wineThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.0",
"robot_type": "aloha",
"total_episodes": 50,
"total_frames": 65000,
"total_tasks": 1,
"total_videos": 150,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 50,
"splits": {
"train": "0:50"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/lerobot/aloha_mobile_wipe_wine.AgiBot-g1_mobile_accessory_storage_box_d
AgiBot-g1_mobile_accessory_storage_box_d
📋 Overview
This dataset uses an extended format based on LeRobot and is fully compatible with LeRobot.
Robot Type: ruantong_a2d
| Codebase Version: v2.1
End-Effector Type: two_finger_gripper
🏠 Scene Types
This dataset covers the following scene types:
home
🤖 Atomic Actions
This dataset includes the following atomic actions:
grasp
place
pick
📊 Dataset Statistics
Metric
Value
Total… See the full description on the dataset page: https://huggingface.co/datasets/RoboCOIN/AgiBot-g1_mobile_accessory_storage_box_d.AIRBOT_MMK2_mobile_calculator_box
AIRBOT_MMK2_mobile_calculator_box
📋 Overview
This dataset uses an extended format based on LeRobot and is fully compatible with LeRobot.
Robot Type: discover_robotics_aitbot_mmk2
| Codebase Version: v2.1
End-Effector Type: five_finger_hand
🏠 Scene Types
This dataset covers the following scene types:
home
🤖 Atomic Actions
This dataset includes the following atomic actions:
grasp
place
pick
📊 Dataset Statistics
Metric
Value… See the full description on the dataset page: https://huggingface.co/datasets/RoboCOIN/AIRBOT_MMK2_mobile_calculator_box.Cobot_Magic_mobile_cube_blackboard
Cobot_Magic_mobile_cube_blackboard
📋 Overview
This dataset uses an extended format based on LeRobot and is fully compatible with LeRobot.
Robot Type: agilex_cobot_decoupled_magic
| Codebase Version: v2.1
End-Effector Type: two_finger_gripper
🏠 Scene Types
This dataset covers the following scene types:
home
🤖 Atomic Actions
This dataset includes the following atomic actions:
grasp
pick
place
📊 Dataset Statistics
Metric
Value… See the full description on the dataset page: https://huggingface.co/datasets/RoboCOIN/Cobot_Magic_mobile_cube_blackboard.AgiBot-g1_mobile_accessory_storage_box_e
AgiBot-g1_mobile_accessory_storage_box_e
📋 Overview
This dataset uses an extended format based on LeRobot and is fully compatible with LeRobot.
Robot Type: ruantong_a2d
| Codebase Version: v2.1
End-Effector Type: two_finger_gripper
🏠 Scene Types
This dataset covers the following scene types:
home
🤖 Atomic Actions
This dataset includes the following atomic actions:
grasp
place
pick
📊 Dataset Statistics
Metric
Value
Total… See the full description on the dataset page: https://huggingface.co/datasets/RoboCOIN/AgiBot-g1_mobile_accessory_storage_box_e.mobile-gs2The dataset is link to the paper https://huggingface.co/papers/2606.30017
This dataset contains the weights of trained Gaussian Splatting parameters.
We save them as a JSON file to better load for WebGL renderer.
You can try our render on your mobile here https://xiaobiaodu.github.io/flux-gs-project/
All code is released, and you can find them at our project page.
aloha_mobile_chairThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v2.0",
"robot_type": "aloha",
"total_episodes": 55,
"total_frames": 110000,
"total_tasks": 1,
"total_videos": 165,
"total_chunks": 1,
"chunks_size": 1000,
"fps": 50,
"splits": {
"train": "0:55"
},
"data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/lerobot/aloha_mobile_chair.AgiBot-g1_mobile_accessory_storage_box_c
AgiBot-g1_mobile_accessory_storage_box_c
📋 Overview
This dataset uses an extended format based on LeRobot and is fully compatible with LeRobot.
Robot Type: ruantong_a2d
| Codebase Version: v2.1
End-Effector Type: two_finger_gripper
🏠 Scene Types
This dataset covers the following scene types:
home
🤖 Atomic Actions
This dataset includes the following atomic actions:
grasp
place
pick
📊 Dataset Statistics
Metric
Value
Total… See the full description on the dataset page: https://huggingface.co/datasets/RoboCOIN/AgiBot-g1_mobile_accessory_storage_box_c.Mobile-R1
Dataset Card for Mobile-R1
Dataset Structure
images/: All the screenshots
data.jsonl: The trajectory data
Data Fields
All screenshots are stored in the images/ directory.
We describe the structure of a single trajectory entry from the file data.jsonl, which contains the full interaction trajectories and action history.
app_name: String. The name of the mobile application (e.g., "闲鱼" / Xianyu) where the task is performed.
trajectory_length:… See the full description on the dataset page: https://huggingface.co/datasets/huCamus/Mobile-R1.AgiBot-g1_mobile_accessory_storage_box_f
AgiBot-g1_mobile_accessory_storage_box_f
📋 Overview
This dataset uses an extended format based on LeRobot and is fully compatible with LeRobot.
Robot Type: ruantong_a2d
| Codebase Version: v2.1
End-Effector Type: two_finger_gripper
🏠 Scene Types
This dataset covers the following scene types:
home
🤖 Atomic Actions
This dataset includes the following atomic actions:
grasp
place
pick
📊 Dataset Statistics
Metric
Value
Total… See the full description on the dataset page: https://huggingface.co/datasets/RoboCOIN/AgiBot-g1_mobile_accessory_storage_box_f.