--- license: other tags: - cua-lite - gui - sft task_categories: - image-text-to-text configs: - config_name: default data_files: - split: train path: - "*/*/train*parquet" - "*/*/train/*.parquet" - "*/*/train/*/*.parquet" - split: validation path: - "*/*/validation*parquet" - "*/*/validation/*.parquet" - "*/*/validation/*/*.parquet" - config_name: EvoCUA-8B-20260105 data_files: - split: train path: - "*/*/train/EvoCUA-8B-20260105.parquet" - "*/*/train/EvoCUA-8B-20260105/*.parquet" - split: validation path: - "*/*/validation/EvoCUA-8B-20260105.parquet" - "*/*/validation/EvoCUA-8B-20260105/*.parquet" - config_name: Qwen3-VL-2B-Instruct data_files: - split: train path: - "*/*/train/Qwen3-VL-2B-Instruct.parquet" - "*/*/train/Qwen3-VL-2B-Instruct/*.parquet" - split: validation path: - "*/*/validation/Qwen3-VL-2B-Instruct.parquet" - "*/*/validation/Qwen3-VL-2B-Instruct/*.parquet" - config_name: Qwen3-VL-32B-Instruct data_files: - split: train path: - "*/*/train/Qwen3-VL-32B-Instruct.parquet" - "*/*/train/Qwen3-VL-32B-Instruct/*.parquet" - split: validation path: - "*/*/validation/Qwen3-VL-32B-Instruct.parquet" - "*/*/validation/Qwen3-VL-32B-Instruct/*.parquet" - config_name: Qwen3-VL-4B-Instruct data_files: - split: train path: - "*/*/train/Qwen3-VL-4B-Instruct.parquet" - "*/*/train/Qwen3-VL-4B-Instruct/*.parquet" - split: validation path: - "*/*/validation/Qwen3-VL-4B-Instruct.parquet" - "*/*/validation/Qwen3-VL-4B-Instruct/*.parquet" - config_name: Qwen3-VL-8B-Instruct data_files: - split: train path: - "*/*/train/Qwen3-VL-8B-Instruct.parquet" - "*/*/train/Qwen3-VL-8B-Instruct/*.parquet" - split: validation path: - "*/*/validation/Qwen3-VL-8B-Instruct.parquet" - "*/*/validation/Qwen3-VL-8B-Instruct/*.parquet" - config_name: Qwen3.5-27B data_files: - split: train path: - "*/*/train/Qwen3.5-27B.parquet" - "*/*/train/Qwen3.5-27B/*.parquet" - split: validation path: - "*/*/validation/Qwen3.5-27B.parquet" - "*/*/validation/Qwen3.5-27B/*.parquet" - config_name: Qwen3.5-2B data_files: - split: train path: - "*/*/train/Qwen3.5-2B.parquet" - "*/*/train/Qwen3.5-2B/*.parquet" - split: validation path: - "*/*/validation/Qwen3.5-2B.parquet" - "*/*/validation/Qwen3.5-2B/*.parquet" - config_name: Qwen3.5-4B data_files: - split: train path: - "*/*/train/Qwen3.5-4B.parquet" - "*/*/train/Qwen3.5-4B/*.parquet" - split: validation path: - "*/*/validation/Qwen3.5-4B.parquet" - "*/*/validation/Qwen3.5-4B/*.parquet" - config_name: Qwen3.5-9B data_files: - split: train path: - "*/*/train/Qwen3.5-9B.parquet" - "*/*/train/Qwen3.5-9B/*.parquet" - split: validation path: - "*/*/validation/Qwen3.5-9B.parquet" - "*/*/validation/Qwen3.5-9B/*.parquet" - config_name: Qwen3.8-27B data_files: - split: train path: - "*/*/train/Qwen3.8-27B.parquet" - "*/*/train/Qwen3.8-27B/*.parquet" - split: validation path: - "*/*/validation/Qwen3.8-27B.parquet" - "*/*/validation/Qwen3.8-27B/*.parquet" - config_name: ScaleCUA-7B data_files: - split: train path: - "*/*/train/ScaleCUA-7B.parquet" - "*/*/train/ScaleCUA-7B/*.parquet" - config_name: UI-TARS-1.5-7B data_files: - split: train path: - "*/*/train/UI-TARS-1.5-7B.parquet" - "*/*/train/UI-TARS-1.5-7B/*.parquet" - split: validation path: - "*/*/validation/UI-TARS-1.5-7B.parquet" - "*/*/validation/UI-TARS-1.5-7B/*.parquet" - config_name: UI-TARS-7B-DPO data_files: - split: train path: - "*/*/train/UI-TARS-7B-DPO.parquet" - "*/*/train/UI-TARS-7B-DPO/*.parquet" - split: validation path: - "*/*/validation/UI-TARS-7B-DPO.parquet" - "*/*/validation/UI-TARS-7B-DPO/*.parquet" - config_name: gpt-5.4 data_files: - split: train path: - "*/*/train/gpt-5.4.parquet" - "*/*/train/gpt-5.4/*.parquet" - split: validation path: - "*/*/validation/gpt-5.4.parquet" - "*/*/validation/gpt-5.4/*.parquet" - config_name: gpt-5.5 data_files: - split: train path: - "*/*/train/gpt-5.5.parquet" - "*/*/train/gpt-5.5/*.parquet" - split: validation path: - "*/*/validation/gpt-5.5.parquet" - "*/*/validation/gpt-5.5/*.parquet" - config_name: gpt-5.6-sol data_files: - split: validation path: - "*/*/validation/gpt-5.6-sol.parquet" - "*/*/validation/gpt-5.6-sol/*.parquet" - config_name: gpt-6-astra data_files: - split: validation path: - "*/*/validation/gpt-6-astra.parquet" - "*/*/validation/gpt-6-astra/*.parquet" --- # cua-lite/LiteOS LiteOS desktop-use agent rollout dataset. Train split is preserved from HaoranLiu/LiteOS; validation split is staged from .logs/rollout/lite.osworld_eval_g3 with model folder names matching train split configs. ## Origin ## Load via `datasets` ```python from datasets import load_dataset # entire dataset ds = load_dataset("cua-lite/LiteOS") # just one named subset (config) ds = load_dataset("cua-lite/LiteOS", "EvoCUA-8B-20260105") ``` 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_tool_schemas, valid_actions, others{...}}` | Coordinate values in `messages` are normalized to `[0, 1000]` integers. **Image-dedup (`grounding.*` / `understanding` cohorts).** These cohorts are single-image-per-row and many rows share the same screenshot, so to avoid re-embedding identical image bytes once per instruction they are stored *folded*: one row per unique screenshot (image embedded once), carrying an extra **`_folded`** column — a JSON string with the authoritative list of `{messages, metadata}` members for that screenshot. The row's top-level `messages` is the members concatenated for viewer convenience. `use` cohorts are not folded. **Use `lite.data.hf.download` to consume this repo** — it unfolds automatically back to one row per instruction; reading the parquet directly yields the folded form. ## Layout ``` ///shard-NNNNN-of-NNNNN.parquet # single-variant cohort ////shard-NNNNN-of-NNNNN.parquet # multi-variant cohort ``` - `platform` ∈ {desktop, mobile, web} - `task_type` ∈ {understanding, grounding.action, grounding.point, grounding.bbox, use} — used verbatim as the dir component - HF config names are `.` by default (e.g. `mobile.grounding.action`) — UNLESS the dataset was staged with `--config-names`, which sets verbatim, explicitly-chosen config names (see the `configs:` block above for the authoritative list). The agent registry lookup key in code is `@@` (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` (the `datasets` library blacklists `<>:/\|?*` in split names; everything else is fine in config_name) - `validation` is an in-distribution held-out slice (never used in training); `test` is reserved for out-of-distribution benchmark datasets ## Stats | platform | task_type | variant | train | validation | |---|---|---|---:|---:| | desktop | use | EvoCUA-8B-20260105 | 1,703 | 996 | | desktop | use | Qwen3-VL-2B-Instruct | 1,697 | 996 | | desktop | use | Qwen3-VL-32B-Instruct | 1,702 | 996 | | desktop | use | Qwen3-VL-4B-Instruct | 1,703 | 996 | | desktop | use | Qwen3-VL-8B-Instruct | 1,704 | 996 | | desktop | use | Qwen3.5-27B | 5,112 | 996 | | desktop | use | Qwen3.5-2B | 1,701 | 995 | | desktop | use | Qwen3.5-4B | 1,704 | 995 | | desktop | use | Qwen3.5-9B | 1,704 | 996 | | desktop | use | Qwen3.8-27B | 5,046 | 994 | | desktop | use | ScaleCUA-7B | 1,676 | 0 | | desktop | use | UI-TARS-1.5-7B | 1,704 | 996 | | desktop | use | UI-TARS-7B-DPO | 1,704 | 993 | | desktop | use | gpt-5.4 | 1,704 | 996 | | desktop | use | gpt-5.5 | 1,702 | 996 | | desktop | use | gpt-5.6-sol | 0 | 994 | | desktop | use | gpt-6-astra | 0 | 993 | ## 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: ``` $CUA_LITE_DATASETS_ROOT/cua-lite/LiteOS/ images//. # content-addressed image store //[/].parquet # rows reference images by relative path ``` Rows in the local parquet have `images: list[str]`; bytes are extracted to the image store. `lite.train.export.export_sft` consumes the local form directly with `--image-root=$CUA_LITE_DATASETS_ROOT`. - Total unique images: **632,354** - Image store size: **285.91 GB** ## Notes Staged via `lite.data.hf.stage` from rollout log-roots: .logs/rollout/lite.osworld_eval_g3/Qwen3-VL-2B-Instruct, .logs/rollout/lite.osworld_eval_g3/Qwen3-VL-32B-Instruct, .logs/rollout/lite.osworld_eval_g3/Qwen3-VL-4B-Instruct, .logs/rollout/lite.osworld_eval_g3/Qwen3-VL-8B-Instruct, .logs/rollout/lite.osworld_eval_g3/Qwen3.5-27B, .logs/rollout/lite.osworld_eval_g3/Qwen3.5-2B, .logs/rollout/lite.osworld_eval_g3/Qwen3.5-4B, .logs/rollout/lite.osworld_eval_g3/Qwen3.5-9B, .logs/rollout/lite.osworld_eval_g3/Qwen3.8-27B, .logs/rollout/lite.osworld_eval_g3/UI-TARS-1.5-7B, .logs/rollout/lite.osworld_eval_g3/UI-TARS-7B-DPO, .logs/rollout/lite.osworld_eval_g3/evocua_8b, .logs/rollout/lite.osworld_eval_g3/gpt-5.4, .logs/rollout/lite.osworld_eval_g3/gpt-5.5 (row filter: none). ## License & citation other