Jerryxrc commited on
Commit
3584323
·
verified ·
1 Parent(s): 311c1fe

Polish dataset card copy

Browse files
Files changed (1) hide show
  1. README.md +20 -4
README.md CHANGED
@@ -36,11 +36,13 @@ dataset_info:
36
 
37
  ## Dataset Summary
38
 
39
- Web4Style is a website design understanding and retrieval dataset.
40
 
41
- The current public release contains 300 examples from 300 public websites across 25 categories. It includes author-written annotations describing each site's high-level brief, content structure, functional patterns, and visual style.
42
 
43
- This public release does not include screenshots or other third-party website assets.
 
 
44
 
45
  ## What Is Included
46
 
@@ -54,6 +56,8 @@ Each row contains:
54
  - `function`: summary of the site's key interaction patterns and functional modules
55
  - `style`: summary of the site's visual style and aesthetic direction
56
 
 
 
57
  ## Data Files
58
 
59
  - `data/train-00000-of-00001.parquet`: primary Hugging Face upload artifact
@@ -70,9 +74,11 @@ Potential uses include:
70
  - retrieval of stylistically similar websites
71
  - benchmarking style-aware web understanding systems
72
 
 
 
73
  ## Data Collection and Processing
74
 
75
- The source dataset paired website screenshots with structured textual annotations. For this metadata-only release, the screenshot field was removed and the remaining annotation fields were retained.
76
 
77
  The release script performs the following transformations:
78
 
@@ -95,6 +101,8 @@ Not recommended uses:
95
  - identity extraction or contact scraping
96
  - legal assertions about ownership of third-party content
97
 
 
 
98
  ## Licensing and Rights
99
 
100
  This repository is intended to distribute only the dataset packaging and the author-written metadata fields included in this release.
@@ -117,6 +125,8 @@ This dataset is not intended for person-level profiling, identity inference, or
117
  - Websites change over time, so the linked sources may no longer match the original capture context.
118
  - The dataset is better suited to style and structure understanding than to factual verification about the linked websites.
119
 
 
 
120
  ## Release Notes
121
 
122
  - Split count: `train=300`
@@ -132,6 +142,12 @@ dataset = load_dataset("parquet", data_files="data/train-00000-of-00001.parquet"
132
  print(dataset["train"][0])
133
  ```
134
 
 
 
 
 
 
 
135
  ## Publishing
136
 
137
  This directory is structured to be pushed directly to a Hugging Face dataset repository root.
 
36
 
37
  ## Dataset Summary
38
 
39
+ Web4Style is a curated dataset for research on website design understanding, style analysis, and text-based retrieval.
40
 
41
+ The current public release contains 300 examples from 300 public websites across 25 categories. Each example provides author-written annotations that summarize a site's design intent, information architecture, interaction patterns, and visual direction.
42
 
43
+ This public release is metadata-only and does not include screenshots or other third-party website assets.
44
+
45
+ Web4Style is intended for benchmark creation, retrieval experiments, prompt design research, and analysis of how website characteristics can be represented in structured natural language.
46
 
47
  ## What Is Included
48
 
 
56
  - `function`: summary of the site's key interaction patterns and functional modules
57
  - `style`: summary of the site's visual style and aesthetic direction
58
 
59
+ Together, these fields describe each website at a level that is useful for style-aware modeling without redistributing the original page imagery.
60
+
61
  ## Data Files
62
 
63
  - `data/train-00000-of-00001.parquet`: primary Hugging Face upload artifact
 
74
  - retrieval of stylistically similar websites
75
  - benchmarking style-aware web understanding systems
76
 
77
+ The dataset is especially suitable for experiments where the target representation is textual rather than pixel-level reconstruction.
78
+
79
  ## Data Collection and Processing
80
 
81
+ The source dataset paired website screenshots with structured textual annotations. For this public release, the screenshot field was removed and only the annotation metadata was retained.
82
 
83
  The release script performs the following transformations:
84
 
 
101
  - identity extraction or contact scraping
102
  - legal assertions about ownership of third-party content
103
 
104
+ Web4Style should be understood as a compact, design-focused annotation dataset rather than a comprehensive crawl of the public web.
105
+
106
  ## Licensing and Rights
107
 
108
  This repository is intended to distribute only the dataset packaging and the author-written metadata fields included in this release.
 
125
  - Websites change over time, so the linked sources may no longer match the original capture context.
126
  - The dataset is better suited to style and structure understanding than to factual verification about the linked websites.
127
 
128
+ Because the release excludes screenshots, it is not suitable for image-based benchmark tasks unless paired with separately licensed visual data.
129
+
130
  ## Release Notes
131
 
132
  - Split count: `train=300`
 
142
  print(dataset["train"][0])
143
  ```
144
 
145
+ Example applications include:
146
+
147
+ - retrieving examples with similar design characteristics
148
+ - generating structured web design briefs from category-conditioned prompts
149
+ - studying the relationship between site function and visual direction
150
+
151
  ## Publishing
152
 
153
  This directory is structured to be pushed directly to a Hugging Face dataset repository root.