Search papers, labs, and topics across Lattice.
1
0
3
Target-aware data selection and VLM-based caption refinement can boost food image retrieval performance by over 19%, revealing the potential of fine-tuning multimodal models with curated data.
