HongxinLi commited on
Commit
ccf3da6
·
verified ·
1 Parent(s): 798872e

Add dataset card (README) with schema and field documentation

Browse files
Files changed (1) hide show
  1. README.md +153 -0
README.md ADDED
@@ -0,0 +1,153 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: cc-by-nc-4.0
3
+ task_categories:
4
+ - image-to-text
5
+ - visual-question-answering
6
+ language:
7
+ - en
8
+ tags:
9
+ - gui
10
+ - gui-agent
11
+ - ui-understanding
12
+ - screenshot
13
+ - visual-grounding
14
+ - ocr
15
+ pretty_name: UIPro MultiUI
16
+ size_categories:
17
+ - 1M<n<10M
18
+ ---
19
+
20
+ # UIPro-MultiUI-Data-v1
21
+
22
+ Part of the **UIPro** GUI-agent training suite (ICCV 2025). This repository packages the
23
+ **MultiUI** source into the unified UIPro instruction-tuning format, with coordinates
24
+ normalized to a **[0, 1000]** grid.
25
+
26
+ > Web UI grounding, OCR, web QA and embedded QA derived from MultiUI.
27
+
28
+ ## Dataset at a glance
29
+
30
+ | | |
31
+ | :--- | :--- |
32
+ | **Total samples** | 3,410,901 |
33
+ | **Valid images** | 3,410,901 |
34
+ | **Avg. samples / image** | 1.0 |
35
+ | **Coordinate scale** | 0–1000 |
36
+ | **Source dataset** | [MultiUI](https://huggingface.co/datasets/neulab/MultiUI) |
37
+
38
+ ### Samples by task
39
+
40
+ | Task | Count |
41
+ | :--- | ---: |
42
+ | IntentGnd | 1,328,726 |
43
+ | EmbedQA | 934,059 |
44
+ | ElemGnd | 729,642 |
45
+ | OCR | 207,062 |
46
+ | TitleOCR | 164,420 |
47
+ | WebQA | 46,992 |
48
+
49
+ ## Repository file structure
50
+
51
+ | File | Description |
52
+ | :--- | :--- |
53
+ | `MultiUI_gnd_ref_qa_scale1000_3410k.json` | The dataset: a JSON **list** of 3,410,901 sample objects (schema below). |
54
+ | `MultiUI_gnd_ref_qa_scale1000_3410k_sample.json` | A small preview slice of the same schema, for quick inspection without downloading everything. |
55
+ | `MultiUI_gnd_ref_qa_scale1000_3410k_images.zip` | All screenshots referenced by the `image` field, preserving the relative paths stored there. |
56
+ | `MultiUI_gnd_ref_qa_scale1000_3410k_info.json` | Full generation report — per-task counts, image statistics, invalid-element breakdown, and the exact processing config. |
57
+
58
+ Unzip `MultiUI_gnd_ref_qa_scale1000_3410k_images.zip` and each sample's `image` path resolves relative to the extraction root.
59
+
60
+ ## Sample schema — every field explained
61
+
62
+ Each element of the main JSON list is one training sample. This dataset's samples use the
63
+ following fields:
64
+
65
+ | Field | Meaning |
66
+ | :--- | :--- |
67
+ | `conversations` | The vision-language dialogue: a list of turns, each `{"from": "human"|"gpt", "value": ...}`. The **human** turn holds the instruction/question and contains the `<image>` placeholder marking where the screenshot is inserted; the **gpt** turn is the ground-truth answer. |
68
+ | `id` | Unique sample identifier, formatted `autogui_<dataset>_<task>_<n>`. The `<task>` segment (e.g. `intentgnd`, `textloc`, `ocr`, `elemgnd`, `elemref`) tells you which task the sample belongs to. |
69
+ | `image` | Path to the screenshot **inside `_images.zip`**, relative to the archive root. Load the image by joining this path with your extraction directory. |
70
+ | `task_attr` | The task's target attribute in plain form — for grounding tasks the referred element's text/instruction; for OCR/referring tasks the queried coordinate string. Useful for filtering or building custom prompts without parsing the conversation. |
71
+ | `unnormalized_box` | Ground-truth bounding box in **original image pixels**, as `[x1, y1, x2, y2]` (top-left, bottom-right). Present when a box is available. Note: the answer in the `gpt` turn is **normalized to 0–1000**, while this field is the raw-pixel box — divide by width/height and multiply by 1000 to reconcile them. |
72
+
73
+ > **Note:** Fields tied to a bounding box (e.g. `unnormalized_box`) are only present on samples
74
+ > that have a box; point-only answers (e.g. some intent-grounding samples) may omit them.
75
+
76
+ ### The `conversations` field in detail
77
+
78
+ `conversations` is a list of turns that a vision-language model consumes directly:
79
+
80
+ - Each turn is `{"from": "...", "value": "..."}`.
81
+ - `from` is either **`human`** (the prompt) or **`gpt`** (the ground-truth response).
82
+ - The token **`<image>`** inside a human turn marks where the screenshot is spliced into the
83
+ prompt — replace it with the actual image when tokenizing.
84
+
85
+ ### Coordinate system
86
+
87
+ - Answers are **normalized to the `0–1000` range** relative to image width/height.
88
+ - A **point** answer is formatted `(x,y)`; a **bounding box** answer is `(x1,y1,x2,y2)`.
89
+ - Prompts ending in `(with point)` expect a point; `(with bbox)` expect a box.
90
+ - `unnormalized_box`, when present, is the same box in **raw pixels** — combine it with `wxh`
91
+ (or the image's true size) to convert between pixels and the normalized grid.
92
+
93
+ ## Example
94
+
95
+ ```json
96
+ {
97
+ "id": "autogui_multiui_embedqa_873493",
98
+ "image": "MultiUI/v0.8_exclude_v0.7/caption_data_new/Training9_pc1280_curated_start_idx300000/data/002908721/0_embed.png",
99
+ "conversations": [
100
+ {
101
+ "from": "human",
102
+ "value": "<image>\nHow many eyes does the robot have?\nPlease answer the question with a detailed and comprehensive explanation."
103
+ },
104
+ {
105
+ "from": "gpt",
106
+ "value": "The caption describes the robot's head as having two circular eyes that convey a friendly expression, indicating that the robot has a pair of eyes."
107
+ }
108
+ ]
109
+ }
110
+ ```
111
+
112
+ ## Usage
113
+
114
+ ```python
115
+ import json, os, zipfile
116
+ from PIL import Image
117
+ from huggingface_hub import hf_hub_download
118
+
119
+ repo = "HongxinLi/UIPro-MultiUI-Data-v1"
120
+ samples = json.load(open(hf_hub_download(repo, "MultiUI_gnd_ref_qa_scale1000_3410k.json", repo_type="dataset")))
121
+
122
+ images_zip = hf_hub_download(repo, "MultiUI_gnd_ref_qa_scale1000_3410k_images.zip", repo_type="dataset")
123
+ with zipfile.ZipFile(images_zip) as zf:
124
+ zf.extractall("images/")
125
+
126
+ s = samples[0]
127
+ print(s["conversations"])
128
+ img = Image.open(os.path.join("images", s["image"])) # screenshot for this sample
129
+ print("image size:", img.size)
130
+ ```
131
+
132
+ ## About UIPro
133
+
134
+ UIPro is a generalist GUI agent trained on 20.6M understanding tasks across 13 task types,
135
+ followed by agent continued fine-tuning. See the project repository for the full data
136
+ pipeline, training recipes and evaluation scripts:
137
+ **https://github.com/ZJULiHongxin/UIPro**
138
+
139
+ ## License
140
+
141
+ Released under **CC BY-NC 4.0** (non-commercial research use). The underlying screenshots and
142
+ annotations remain subject to the terms of their original source, [MultiUI](https://huggingface.co/datasets/neulab/MultiUI).
143
+
144
+ ## Citation
145
+
146
+ ```bibtex
147
+ @inproceedings{uipro2025,
148
+ title = {UIPro: A Generalist GUI Agent},
149
+ author = {Li, Hongxin and others},
150
+ booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
151
+ year = {2025}
152
+ }
153
+ ```