Tracking Meets LoRA: Faster Training, Larger Model, Stronger Performance

Benchmark Model Rank Results
visual-object-tracking-on-got-10kLoRAT-g-378#8Average Overlap: 78.9Success Rate 0.5: 87.8Success Rate 0.75: 80.7
visual-object-tracking-on-got-10kLoRAT-L-378#13Average Overlap: 77.5Success Rate 0.5: 86.2Success Rate 0.75: 78.1
visual-object-tracking-on-lasotLoRAT-g-378#4AUC: 76.2Normalized Precision: 85.3Precision: 83.5
visual-object-tracking-on-lasotLoRAT-L-378#6AUC: 75.1Normalized Precision: 84.1Precision: 82.0
visual-object-tracking-on-lasot-extLoRAT-L-378#3AUC: 56.6Normalized Precision: 69.0Precision: 65.1
visual-object-tracking-on-lasot-extLoRAT-g-378#4AUC: 56.5Normalized Precision: 69.0Precision: 64.9
visual-object-tracking-on-needforspeedLoRAT-g-378#5AUC: 0.681
visual-object-tracking-on-needforspeedLoRAT-L-378#8AUC: 0.667
visual-object-tracking-on-tnl2kLoRAT-g-378#5AUC: 62.7precision: 67.8
visual-object-tracking-on-tnl2kLoRAT-L-378#6AUC: 62.3precision: 67.0
visual-object-tracking-on-trackingnetLoRAT-g-378#9Accuracy: 86.0Normalized Precision: 90.2Precision: 86.1
visual-object-tracking-on-trackingnetLoRAT-L-378#11Accuracy: 85.6Normalized Precision: 89.7Precision: 85.4
visual-object-tracking-on-uav123LoRAT-g-378#1AUC: 0.739
visual-object-tracking-on-uav123LoRAT-L-378#3AUC: 0.725