TAOBAO-MM: Large-Scale E-Commerce User Interaction Sequences with Multimodal Embeddings

TAOBAO-MM is a large-scale recommendation dataset derived from user interaction logs on Taobao, one of the world's largest e-commerce platforms. It features historical behavior sequences of up to 1,000 interactions per user and includes high-quality multimodal embeddings. The dataset was authored by TaoBao-MM and was last updated on the Hugging Face platform on 2026-01-15.

Use Cases

• Training long-sequence recommendation models based on user interaction histories of up to 1,000 items.
• Developing multimodal recommendation systems based on the included high-quality multimodal embeddings.
• Benchmarking next-item prediction algorithms on large-scale, real-world e-commerce data.
• Studying user behavior patterns on a major global e-commerce platform.

Strengths

• Sequences contain up to 1,000 interactions per user, enabling long-sequence modeling.
• Derived from Taobao, one of the world's largest e-commerce platforms, providing real-world scale.
• Includes high-quality multimodal embeddings for items.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count, file formats, and license details are unknown, which may limit suitability assessment.
• Data may reflect geographic or demographic bias inherent to the Taobao platform.

Provenance

Source

User interaction logs from Taobao.

Collection Method

Derived from platform logs, likely processed to create sequences and embeddings.

Freshness

Last updated 2026-01-15 09:10:46.