BLIP

The 'BLIP' dataset/benchmark contains curated data used to evaluate the performance of Vision-Language Models (VLMs) across tasks such as retrieval, captioning, and reasoning.

Papers using BLIP (3)

Can Argus Judge Them All? Comparing VLMs Across Domains2025Investigating The Invertibility Of Multimodal Latent Spaces: Limitations Of Optimization-based Methods2025DeepInsert: Early Layer Bypass for Efficient and Performant Multimodal Understanding2025