Improving Zero-shot Generalization of Learned Prompts via Unsupervised Knowledge Distillation [ECCV24]

ECCV 2024 main CLIP prompt-learning knowledge-distillation computer-vision

We intruce Knowledge Distillation Prompt Learning (KDPL), a novel approach for improving zero-shot generalization of vision-language models (VLMs) using unsupervised knowledge distillation.

03 July 2024

RE-tune - Incremental Fine-Tuning of Biomedical VLMs [NeurIPS23 - W]

NeurIPS 2023 workshop medical-imaging multi-label classification chest-xray

In this work, we introduce RE-tune, a novel approach for incrementally fine-tuning pre-trained Multimodal Biomedical Vision-Language models (VLMs) for multi-label chest disease diagnosis.

09 December 2023