Personalized Multimodal LLM Framework For Long-Term Agent Memory

PersonaVLM is a dataset supporting the development of personalized multimodal agents with long-term memory capabilities. The framework was created by ClareNie and the associated paper was accepted for CVPR 2026. The dataset is hosted on Hugging Face and was last updated in March 2026.

Use Cases

• Training models to extract and summarize personal history from multimodal inputs like text and images.
• Benchmarking agent performance on tasks requiring long-term personalization and memory recall.
• Developing frameworks that integrate remembering, understanding, and evolving capabilities for personalized MLLMs.

Strengths

• Framework is associated with a paper accepted to the CVPR 2026 conference.
• Dataset is hosted on a major platform (Hugging Face) with a last update timestamp of 2026-03-15.

Limitations

• Specific dataset details like row count, column names, file formats, and size are unknown.
• The license for the dataset is unspecified, which could restrict usage.

Provenance

Source

Hugging Face dataset repository created by ClareNie.

Freshness

Last updated on 2026-03-15.

Geography

Region tags indicate 'us', but specific spatial coverage is unknown.