Action Recognition on UCF101 (Base Accuracy, New Accuracy, Harmonic Mean)
89.83Base AccuracyMMRL + AmPLe
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| MMRL + AmPLe2025.12 | 89.83 | 83.1 | 86.33 | |
| 2SFSBackbone=ViT-L/14, Shots=162025.03 | 89.8 | 82.24 | 85.85 | |
| MMABackbone=ViT-L/14, Shots=162025.03 | 88.6 | 81.37 | 84.83 | |
| ProVP2025.03 | 88.56 | 75.55 | 81.54 | |
| MMRLreproduction=true2025.12 | 88.37 | 79.5 | 83.7 | |
| MMRL2025.03 | 88.1 | 80.07 | 83.89 | |
| MMRLreproduction=false2025.12 | 88.1 | 80.07 | 83.89 | |
| MMA + AmPLe2025.12 | 88.03 | 78.83 | 83.18 | |
| TransAgentVision Encoder=ViT-B/16, Evaluation Protocol=Base-to-novel generalization2024.10 | 87.6 | 80.47 | 83.88 | |
| AttriPrompt2025.12 | 87.5 | 81.63 | 84.46 | |
| TCP2025.03 | 87.13 | 80.77 | 83.83 | |
| PromptSRCShots=162024.04 | 87.1 | 78.8 | 82.74 | |
| PromptSRCVision Encoder=ViT-B/16, Evaluation Protocol=Base-to-novel generalization2024.10 | 87.1 | 78.8 | 82.74 | |
| PromptSRC2025.03 | 87.1 | 78.8 | 82.74 | |
| P-SRCYear=20232026.03 | 87.1 | 78.8 | 82.74 | |
| ProMetaRShots=162024.04 | 86.97 | 79.84 | 83.25 | |
| MaPLe + AmPLe2025.12 | 86.87 | 79 | 82.75 | |
| HPTBackbone=ViT-B/162023.12 | 86.52 | 80.06 | 83.16 | |
| HPTYear=20242026.03 | 86.52 | 80.06 | 83.16 | |
| 2SFSBackbone=ViT-B/16, Shots=82025.03 | 86.37 | 78.58 | 82.29 | |
| 2SFSBackbone=ViT-B/16, Shots=8, Evaluation Protocol=base-to-novel generalization2025.03 | 86.37 | 78.58 | 82.29 | |
| MMA2025.03 | 86.23 | 80.03 | 82.2 | |
| MMAreproduction=false2025.12 | 86.23 | 80.03 | 83.01 | |
| MMAYear=20242026.03 | 86.23 | 80.03 | 82.2 | |
| MMAreproduction=true2025.12 | 86.13 | 78.57 | 82.18 | |
| LwEIBYear=20252026.03 | 85.73 | 82.37 | 84.02 | |
| HeBAYear=Ours2026.03 | 85.73 | 82.69 | 84.18 | |
| LASP-VYear=20232026.03 | 85.53 | 78.2 | 81.7 | |
| MetaPrompt2025.03 | 85.33 | 77.72 | 81.35 | |
| IPO2025.12 | 85.32 | 80.92 | 83.06 | |
| MaPLeVision Encoder=ViT-B/16, Evaluation Protocol=Base-to-novel generalization2024.10 | 85.23 | 71.97 | 78.04 | |
| ProDA2025.03 | 85.23 | 71.97 | 78.04 | |
| ProDAYear=20222026.03 | 85.23 | 71.97 | 78.04 | |
| MMABackbone=ViT-B/16, Shots=82025.03 | 84.8 | 78.1 | 81.31 | |
| MMABackbone=ViT-B/16, Shots=8, Evaluation Protocol=base-to-novel generalization2025.03 | 84.8 | 78.1 | 81.31 | |
| CoOpShots=162024.04 | 84.69 | 56.05 | 67.46 | |
| CoOpVision Encoder=ViT-B/16, Evaluation Protocol=Base-to-novel generalization2024.10 | 84.69 | 56.05 | 67.46 | |
| CoOp2025.03 | 84.69 | 56.05 | 67.46 | |
| CoOp2025.12 | 84.69 | 56.05 | 67.46 | |
| CoOpYear=20222026.03 | 84.69 | 56.05 | 67.46 | |
| IVLPShots=162024.04 | 84.23 | 76.78 | 80.33 | |
| RPOShots=162024.04 | 83.67 | 75.43 | 79.34 | |
| RPOVision Encoder=ViT-B/16, Evaluation Protocol=Base-to-novel generalization2024.10 | 83.67 | 75.43 | 79.34 | |
| RPOYear=20232026.03 | 83.67 | 75.43 | 79.34 | |
| MaPLereproduction=true2025.12 | 83.33 | 77.07 | 80.08 | |
| MaPLeBackbone=ViT-B/162023.12 | 83 | 78.66 | 80.77 | |
| MaPLeShots=162024.04 | 83 | 78.66 | 80.77 | |
| MaPLe2025.03 | 83 | 78.66 | 80.77 | |
| MaPLereproduction=false2025.12 | 83 | 78.66 | 80.77 | |
| KgCoOp2025.03 | 82.89 | 76.67 | 79.65 | |
| KgCoOpYear=20232026.03 | 82.89 | 76.67 | 79.65 | |
| CoCoOpBackbone=ViT-B/162023.12 | 82.33 | 73.45 | 77.64 | |
| CoCoOpShots=162024.04 | 82.33 | 73.45 | 77.64 | |
| CoCoOpVision Encoder=ViT-B/16, Evaluation Protocol=Base-to-novel generalization2024.10 | 82.33 | 73.45 | 77.64 | |
| CoCoOp2025.03 | 82.33 | 73.45 | 77.64 | |
| CoCoOp2025.12 | 82.33 | 73.45 | 77.64 | |
| CoOpOpYear=20222026.03 | 82.33 | 73.45 | 77.64 | |
| 2SFSVisual backbone=ViT-B/32, Shots per base class=4, Generalization setting=base-to-novel2025.03 | 82.04 | 74.44 | 78.05 | |
| 2SFSBackbone=ViT-B/32, Shots=42025.03 | 82.04 | 74.44 | 78.05 | |
| UNIGRAMShots=162024.04 | 82 | 78.06 | 79.98 | |
| CLIPBackbone=ViT-L/14, Shots=Zero-shot2025.03 | 79.94 | 79.66 | 79.8 | |
| MaPLeYear=20232026.03 | 78.66 | 80.77 | 83 | |
| DiMPLe2025.12 | 78.1 | 75.33 | 76.69 | |
| MMAVisual backbone=ViT-B/32, Shots per base class=4, Generalization setting=base-to-novel2025.03 | 77.2 | 74.43 | 75.79 | |
| MMABackbone=ViT-B/32, Shots=42025.03 | 77.2 | 74.43 | 75.79 | |
| CLIPBackbone=ViT-B/162023.12 | 70.53 | 77.5 | 73.85 | |
| CLIPShots=Zero-shot2024.04 | 70.53 | 77.5 | 73.85 | |
| CLIPVision Encoder=ViT-B/16, Evaluation Protocol=Base-to-novel generalization2024.10 | 70.53 | 77.5 | 73.85 | |
| CLIPBackbone=ViT-B/16, Shots=82025.03 | 70.53 | 77.5 | 73.85 | |
| CLIPBackbone=ViT-B/16, Shots=8, Evaluation Protocol=base-to-novel generalization2025.03 | 70.53 | 77.5 | 73.85 | |
| CLIPzero-shot=true2025.03 | 70.53 | 77.5 | 73.85 | |
| CLIP2025.12 | 70.53 | 77.5 | 73.85 | |
| CLIPYear=20212026.03 | 70.53 | 77.5 | 73.85 | |
| CLIPVisual backbone=ViT-B/32, Shots per base class=4, Generalization setting=base-to-novel2025.03 | 69.08 | 72.96 | 70.97 | |
| CLIPBackbone=ViT-B/32, Shots=42025.03 | 69.08 | 72.96 | 70.97 |