Datasets:
images images listlengths 2 43 | messages stringlengths 5.34k 165k | metadata stringlengths 400 426 |
|---|---|---|
[{"role": "user", "content": [{"type": "image", "index": 0}, {"type": "text", "text": "Open the Pikachu picture on the desktop using GIMP, and then select the \"Venetian Blinds\" filter in the animation."}]}, {"role": "assistant", "content": [{"type": "inline_reasoning", "text": "I need to start working towards the goa... | {"platform": "desktop", "task_type": "navigation", "extra_tools": [], "valid_actions": null, "others": {"id": "agentnet_ubuntu_0030dc52-2a4a-4c0e-895b-48284c200efe", "resolution": [1920, 1080], "os": "ubuntu", "source": "xlangai/AgentNet", "source_id": "0030dc52-2a4a-4c0e-895b-48284c200efe", "domain": "infeasible", "ta... | |
[{"role": "user", "content": [{"type": "image", "index": 0}, {"type": "text", "text": "Open \"Notifications\" in the settings and turn on Notification Popups in Archive Manager."}]}, {"role": "assistant", "content": [{"type": "inline_reasoning", "text": "I need to open the Notifications settings to turn on Notification... | {"platform": "desktop", "task_type": "navigation", "extra_tools": [], "valid_actions": null, "others": {"id": "agentnet_ubuntu_0019cfaa-b2a0-424d-a2b0-b068d7956a7e", "resolution": [1920, 1080], "os": "ubuntu", "source": "xlangai/AgentNet", "source_id": "0019cfaa-b2a0-424d-a2b0-b068d7956a7e", "domain": "infeasible", "ta... | |
[{"role": "user", "content": [{"type": "image", "index": 0}, {"type": "text", "text": "Please help me change the shortcut key for closing the window to 'Ctrl+Shift' and delete another shortcut key for closing the window."}]}, {"role": "assistant", "content": [{"type": "inline_reasoning", "text": "The task involves modi... | {"platform": "desktop", "task_type": "navigation", "extra_tools": [], "valid_actions": null, "others": {"id": "agentnet_ubuntu_00406a3b-9879-40b0-94d8-f7d21b7858d7", "resolution": [1920, 1080], "os": "ubuntu", "source": "xlangai/AgentNet", "source_id": "00406a3b-9879-40b0-94d8-f7d21b7858d7", "domain": "VScode", "task_c... | |
[{"role": "user", "content": [{"type": "image", "index": 0}, {"type": "text", "text": "I'm editing\"The Xiaomi Product Ecosystem Chain.pptx,\"and I'm not very satisfied with the content of part 02 and part 04.Could you please add a strikethrough to the main titles of the chapter pages for these two sections to distingu... | {"platform": "desktop", "task_type": "navigation", "extra_tools": [], "valid_actions": null, "others": {"id": "agentnet_ubuntu_0022f50b-7983-4c7a-9005-01a6d924c152", "resolution": [1920, 1080], "os": "ubuntu", "source": "xlangai/AgentNet", "source_id": "0022f50b-7983-4c7a-9005-01a6d924c152", "domain": "libreoffice_impr... | |
[{"role": "user", "content": [{"type": "image", "index": 0}, {"type": "text", "text": "Can you enable the sound field locator and then change the size to 6.0 and dry to 7.0?"}]}, {"role": "assistant", "content": [{"type": "inline_reasoning", "text": "Since this is the first action and there are no previous actions, I n... | {"platform": "desktop", "task_type": "navigation", "extra_tools": [], "valid_actions": null, "others": {"id": "agentnet_ubuntu_00721815-cc6b-41b2-b078-e743c9108357", "resolution": [1920, 1080], "os": "ubuntu", "source": "xlangai/AgentNet", "source_id": "00721815-cc6b-41b2-b078-e743c9108357", "domain": "VLC", "task_comp... | |
[{"src":"https://datasets-server.huggingface.co/assets/cua-lite/OpenCUA/--/{dataset_git_revision}/--(...TRUNCATED) | "[{\"role\": \"user\", \"content\": [{\"type\": \"image\", \"index\": 0}, {\"type\": \"text\", \"tex(...TRUNCATED) | "{\"platform\": \"desktop\", \"task_type\": \"navigation\", \"extra_tools\": [], \"valid_actions\": (...TRUNCATED) |
[{"src":"https://datasets-server.huggingface.co/assets/cua-lite/OpenCUA/--/{dataset_git_revision}/--(...TRUNCATED) | "[{\"role\": \"user\", \"content\": [{\"type\": \"image\", \"index\": 0}, {\"type\": \"text\", \"tex(...TRUNCATED) | "{\"platform\": \"desktop\", \"task_type\": \"navigation\", \"extra_tools\": [], \"valid_actions\": (...TRUNCATED) |
[{"src":"https://datasets-server.huggingface.co/assets/cua-lite/OpenCUA/--/{dataset_git_revision}/--(...TRUNCATED) | "[{\"role\": \"user\", \"content\": [{\"type\": \"image\", \"index\": 0}, {\"type\": \"text\", \"tex(...TRUNCATED) | "{\"platform\": \"desktop\", \"task_type\": \"navigation\", \"extra_tools\": [], \"valid_actions\": (...TRUNCATED) |
[{"src":"https://datasets-server.huggingface.co/assets/cua-lite/OpenCUA/--/{dataset_git_revision}/--(...TRUNCATED) | "[{\"role\": \"user\", \"content\": [{\"type\": \"image\", \"index\": 0}, {\"type\": \"text\", \"tex(...TRUNCATED) | "{\"platform\": \"desktop\", \"task_type\": \"navigation\", \"extra_tools\": [], \"valid_actions\": (...TRUNCATED) |
[{"src":"https://datasets-server.huggingface.co/assets/cua-lite/OpenCUA/--/{dataset_git_revision}/--(...TRUNCATED) | "[{\"role\": \"user\", \"content\": [{\"type\": \"image\", \"index\": 0}, {\"type\": \"text\", \"tex(...TRUNCATED) | "{\"platform\": \"desktop\", \"task_type\": \"navigation\", \"extra_tools\": [], \"valid_actions\": (...TRUNCATED) |
cua-lite/OpenCUA
cua-lite preprocessed version of xlangai/AgentNet (OpenCUA). Desktop navigation trajectories from Ubuntu and Windows/Mac environments with pyautogui-style actions converted to CUA-lite tool calls.
Origin
Load via datasets
from datasets import load_dataset
# entire dataset
ds = load_dataset("cua-lite/OpenCUA")
# just one (platform, task_type) cohort
ds = load_dataset("cua-lite/OpenCUA", "desktop.navigation")
You can also filter by metadata.platform / metadata.task_type /
metadata.others.* after loading; every row carries a rich metadata
struct (see schema below).
Schema
Each row has these columns:
| column | type | notes |
|---|---|---|
images |
list[Image] | embedded PNG/JPEG bytes; HF viewer renders thumbnails |
messages |
list[struct] | OpenAI-style turns with role + structured content |
metadata |
struct | {platform, task_type, extra_tools, valid_actions, others{...}} |
Coordinate values in messages are normalized to [0, 1000] integers.
Layout
<platform>/<task_type>/<split>/shard-NNNNN-of-NNNNN.parquet # single-variant cohort
<platform>/<task_type>/<split>/<variant>/shard-NNNNN-of-NNNNN.parquet # multi-variant cohort
platform∈ {desktop, mobile, web}task_type∈ {understanding, grounding.action, grounding.point, grounding.bbox, navigation} — used verbatim as the dir component- HF config names are
<platform>.<task_type>(e.g.mobile.grounding.action). The agent registry lookup key in code is<agent>@<platform>@<task_type>(e.g.qwen3_vl@mobile@grounding.action); only this user-facing token uses.between platform and task_type, because@triggers a 403 on the dataset-viewer's signed image URLs. - HF split names stay
train/validation(thedatasetslibrary blacklists<>:/\|?*in split names; everything else is fine in config_name) validationis an in-distribution held-out slice (never used in training);testis reserved for out-of-distribution benchmark datasets
Stats
| platform | task_type | variant | train | validation |
|---|---|---|---|---|
| desktop | navigation | ubuntu | 4,890 | 102 |
| desktop | navigation | win_mac | 17,173 | 371 |
Local mirror & SFT export
For local workflows (SFT export, dedup, mixing across datasets), use
lite.data.hf.download to mirror this repo back to the canonical local
layout:
$CUA_LITE_DATASETS_ROOT/cua-lite/OpenCUA/
images/<hash[:2]>/<hash>.<ext> # content-addressed image store
<platform>/<task_type>/<split>[/<variant>].parquet # rows reference images by relative path
Rows in the local parquet have images: list[str]; bytes are extracted to
the image store. lite.train.utils.data.export_sft consumes the local
form directly with --image-root=$CUA_LITE_DATASETS_ROOT.
- Total unique images: 392,526
- Image store size: 186.57 GB
Notes
(none)
License & citation
See original dataset (xlangai/AgentNet).
- Downloads last month
- 5,006