Datasets:
Search is not available for this dataset
image imagewidth (px) 443 640 |
|---|
DM0018 · E-commerce Product Image-Text Pair Samples / 电商商品图文对样例
| Item / 项目 | Value / 值 |
|---|---|
| Dataset ID / 数据集编号 | DM0018 |
| Files / 数据文件 | 58 |
| Size / 数据体积 | 3.45 MB |
| Provided by / 提供方 | Zovda |
| Contact / 联系邮箱 | lucas@zovda.com |
Overview / 简介
A multimodal image-text corpus built from real e-commerce product pages: each product pairs a full image set with its title and structured attributes, supporting image-text alignment, image-based product search, caption generation and attribute extraction.
基于真实电商商品详情页构建的多模态图文语料：每个商品一组完整商品图、商品标题与结构化属性，支撑图文对齐、以图搜商品、图生标题与图生属性。
What's inside / 数据内容
samples.json: 20 product records withitemId,text_caption,structured_tags,task_types,image_roles, plus per-imagepreview_urland original dimensions.samples.json：20 个商品的记录，含itemId、text_caption、structured_tags、task_types、image_roles与逐图preview_url/ 原始尺寸信息。images/: 57 compressed preview images (long edge ≤ 640 px), named by product and image order.images/：57 张压缩预览图（长边 ≤ 640px），按商品与图序命名。- Five image roles: hero, flat-lay, detail, SKU card and model shot.
- 图片角色分五类：主图、平铺图、细节图、SKU 色卡、模特图。
Specifications / 规格
- Typically 4–5 images per product, image sets keyed by product
- 每个商品通常 4–5 张图，图集以商品为单位
- Structured attributes come straight from page snapshots with no model inference, fully traceable
- 结构化属性直接取自详情页快照，不做模型推断，可回溯原文
- Month × batch stratified sampling with a fixed reproducible seed
- 样本按「月份 × 批次」分层随机抽样，随机种子固定可复现
Suitable tasks / 适用任务
- Image-text alignment
- 图文对齐
- Image-based product search / cross-modal retrieval
- 以图搜商品 / 跨模态检索
- Caption and attribute generation from images
- 图生标题、图生属性
- Image-text consistency and compliance review
- 图文一致性与合规审核
Files / 文件清单
| # | File | Size |
|---|---|---|
| 1 | ecommerce-imgtext-pairs/images/1005577602864_1.jpg |
69.86 KB |
| 2 | ecommerce-imgtext-pairs/images/1005577602864_2.jpg |
62.15 KB |
| 3 | ecommerce-imgtext-pairs/images/1005577602864_5.jpg |
35.32 KB |
| 4 | ecommerce-imgtext-pairs/images/1009117320635_1.jpg |
63.48 KB |
| 5 | ecommerce-imgtext-pairs/images/1009117320635_2.jpg |
55.22 KB |
| 6 | ecommerce-imgtext-pairs/images/1009117320635_5.jpg |
71.92 KB |
| 7 | ecommerce-imgtext-pairs/images/1012152342277_1.jpg |
106.87 KB |
| 8 | ecommerce-imgtext-pairs/images/1012152342277_2.jpg |
113.86 KB |
| 9 | ecommerce-imgtext-pairs/images/1012152342277_5.jpg |
87.25 KB |
| 10 | ecommerce-imgtext-pairs/images/1028602194934_1.jpg |
36.33 KB |
| 11 | ecommerce-imgtext-pairs/images/1028602194934_2.jpg |
48.04 KB |
| 12 | ecommerce-imgtext-pairs/images/1028602194934_3.jpg |
35.18 KB |
| 13 | ecommerce-imgtext-pairs/images/1028602194934_5.jpg |
31.86 KB |
| 14 | ecommerce-imgtext-pairs/images/1035806666969_1.jpg |
26.33 KB |
| 15 | ecommerce-imgtext-pairs/images/1035806666969_2.jpg |
44.19 KB |
| 16 | ecommerce-imgtext-pairs/images/1035806666969_3.jpg |
56.94 KB |
| 17 | ecommerce-imgtext-pairs/images/1038077470334_1.jpg |
38.93 KB |
| 18 | ecommerce-imgtext-pairs/images/1038077470334_2.jpg |
52.75 KB |
| 19 | ecommerce-imgtext-pairs/images/1038077470334_5.jpg |
51.30 KB |
| 20 | ecommerce-imgtext-pairs/images/1040351526281_1.jpg |
47.59 KB |
| 21 | ecommerce-imgtext-pairs/images/1040351526281_2.jpg |
53.25 KB |
| 22 | ecommerce-imgtext-pairs/images/1040351526281_5.jpg |
63.55 KB |
| 23 | ecommerce-imgtext-pairs/images/1043257704911_1.jpg |
55.33 KB |
| 24 | ecommerce-imgtext-pairs/images/1043257704911_2.jpg |
62.24 KB |
| 25 | ecommerce-imgtext-pairs/images/616097048676_1.jpg |
62.91 KB |
| 26 | ecommerce-imgtext-pairs/images/616097048676_2.jpg |
61.96 KB |
| 27 | ecommerce-imgtext-pairs/images/616097048676_5.jpg |
99.40 KB |
| 28 | ecommerce-imgtext-pairs/images/827472116016_1.jpg |
173.56 KB |
| 29 | ecommerce-imgtext-pairs/images/827472116016_2.jpg |
140.78 KB |
| 30 | ecommerce-imgtext-pairs/images/837839695509_1.jpg |
58.70 KB |
| 31 | ecommerce-imgtext-pairs/images/837839695509_2.jpg |
61.08 KB |
| 32 | ecommerce-imgtext-pairs/images/842502069005_1.jpg |
67.01 KB |
| 33 | ecommerce-imgtext-pairs/images/842502069005_2.jpg |
69.63 KB |
| 34 | ecommerce-imgtext-pairs/images/842502069005_5.jpg |
22.80 KB |
| 35 | ecommerce-imgtext-pairs/images/906663032059_1.jpg |
57.38 KB |
| 36 | ecommerce-imgtext-pairs/images/906663032059_2.jpg |
49.83 KB |
| 37 | ecommerce-imgtext-pairs/images/906663032059_5.jpg |
46.32 KB |
| 38 | ecommerce-imgtext-pairs/images/917455226150_1.jpg |
76.87 KB |
| 39 | ecommerce-imgtext-pairs/images/917455226150_2.jpg |
53.31 KB |
| 40 | ecommerce-imgtext-pairs/images/917455226150_5.jpg |
78.78 KB |
| 41 | ecommerce-imgtext-pairs/images/923893125169_1.jpg |
62.61 KB |
| 42 | ecommerce-imgtext-pairs/images/923893125169_2.jpg |
64.16 KB |
| 43 | ecommerce-imgtext-pairs/images/923893125169_5.jpg |
63.04 KB |
| 44 | ecommerce-imgtext-pairs/images/925366451364_1.jpg |
82.63 KB |
| 45 | ecommerce-imgtext-pairs/images/925366451364_2.jpg |
52.17 KB |
| 46 | ecommerce-imgtext-pairs/images/925366451364_5.jpg |
64.92 KB |
| 47 | ecommerce-imgtext-pairs/images/940720724536_1.jpg |
53.54 KB |
| 48 | ecommerce-imgtext-pairs/images/940720724536_2.jpg |
58.46 KB |
| 49 | ecommerce-imgtext-pairs/images/940720724536_5.jpg |
57.98 KB |
| 50 | ecommerce-imgtext-pairs/images/951477319682_1.jpg |
54.99 KB |
| 51 | ecommerce-imgtext-pairs/images/951477319682_2.jpg |
37.26 KB |
| 52 | ecommerce-imgtext-pairs/images/951477319682_5.jpg |
71.81 KB |
| 53 | ecommerce-imgtext-pairs/images/962122359297_1.jpg |
81.67 KB |
| 54 | ecommerce-imgtext-pairs/images/962122359297_2.jpg |
40.24 KB |
| 55 | ecommerce-imgtext-pairs/images/982035617960_1.jpg |
42.43 KB |
| 56 | ecommerce-imgtext-pairs/images/982035617960_2.jpg |
33.01 KB |
| 57 | ecommerce-imgtext-pairs/images/982035617960_5.jpg |
38.85 KB |
| 58 | ecommerce-imgtext-pairs/samples.json |
26.69 KB |
Usage / 使用方式
from huggingface_hub import snapshot_download
snapshot_download(
repo_id="Zovda/dm0018-ecommerce-image-text-pair-samples",
repo_type="dataset",
local_dir="./dm0018-ecommerce-image-text-pair-samples",
)
Rights and disclaimer / 授权与声明
Titles and images originate from publicly available product pages and remain the property of their rightsholders. This dataset demonstrates data organisation and cleaning capability only; formal external licensing requires source authorisation and legal clearance first.
商品标题与图片来源于公开商品详情页，著作权归相应权利人所有。本数据集仅用于展示数据组织与清洗加工能力；正式对外授权须先完成来源授权与法务定性。
This dataset is shared to demonstrate data organisation, technical specifications and delivery structure. For formal licensing, large-scale access or custom production, contact lucas@zovda.com.
本数据集仅用于展示数据组织方式、技术规格与交付结构。正式授权、规模化取数与定制生产请联系 lucas@zovda.com。
About Zovda / 关于 Zovda
Zovda builds, cleans, annotates and packages training data across video, audio, images, e-commerce image-text pairs, exam question banks, embodied AI and coding-agent corpora.
Zovda 提供多模态与文本训练语料的数据采集、清洗、标注与结构化交付服务，覆盖视频、音频、图像、电商图文、K12 与职业题库、具身智能、代码智能体等方向。
Contact / 联系方式: lucas@zovda.com
- Downloads last month
- 269