DeepGlint-AI

/

UniME-LLaVA-1.6-7B

like 5

Follow

DeepGlint 26

This PR refines the model card's metadata by:

• Updating the pipeline_tag to zero-shot-image-classification to better reflect the model's core functionality of learning discriminative multimodal embeddings for tasks like retrieval and classification.
• Adding llava and multimodal to the tags list for improved discoverability and accuracy, as the model is based on LLAVA and is inherently multimodal.