vision benchmarks

5 papers

• Computational Budget Should Be Considered in Data Selection
• CrossSpectra: Exploiting Cross-Layer Smoothness for Parameter-Efficient Fine-Tuning
• Ravan: Multi-Head Low-Rank Adaptation for Federated Fine-Tuning
• The Unseen Threat: Residual Knowledge in Machine Unlearning under Perturbed Samples
• Towards Minimizing Feature Drift in Model Merging: Layer-wise Task Vector Fusion for Adaptive Knowledge Integration