Image Classification on ImageNet Preserved Classes
94Church AccuracySD-v1.5
Evaluation results
| Method | Links | ||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| SD-v1.5Diffusion Model Backbone=SD-v1.5, Classifier Backbone=ResNet-502026.02 | 94 | — | 93 | 94 | 93 | 95 | 92 | 93 | 93 | 93 | 93 | 93 | 93 | 93 | |
| Neuron-Max Scaling (NMS)Diffusion Model Backbone=SD-v1.5, Classifier Backbone=ResNet-50, Sign Retention Threshold=Top-0.62026.02 | 91 | — | 89 | 90 | 91 | 95 | 92 | 89 | 90 | 91 | 88 | 92 | 92 | 91 | |
| Neuron AverageDiffusion Model Backbone=SD-v1.5, Classifier Backbone=ResNet-50, Sign Retention Threshold=Top-0.62026.02 | 81 | — | 87 | 88 | 89 | 83 | 84 | 77 | 89 | 83 | 77 | 86 | 84 | 84 | |
| Neuron SampleDiffusion Model Backbone=SD-v1.5, Classifier Backbone=ResNet-50, Sign Retention Threshold=Top-0.62026.02 | 80 | — | 86 | 86 | 87 | 82 | 86 | 76 | 88 | 83 | 74 | 86 | 83 | 83 | |
| Quant RecoverDiffusion Model Backbone=SD-v1.5, Classifier Backbone=ResNet-502026.02 | 79 | — | 82 | 81 | 75 | 83 | 81 | 76 | 79 | 83 | 73 | 84 | 77 | 79 | |
| Concept PruneDiffusion Model Backbone=SD-v1.5, Classifier Backbone=ResNet-502026.02 | 78 | — | 88 | 81 | 74 | 82 | 81 | 75 | 79 | 83 | 76 | 83 | 77 | 80 |