Table 4 Cluster quality metrics for each layer of the model trained on MNIST.
From: Deep representation learning using layer-wise VICReg losses
Layer number | DB index | CH index |
|---|---|---|
1 | 3.360 | 532.289 |
2 | 3.339 | 565.690 |
3 | 3.287 | 592.728 |
4 | 3.264 | 533.800 |
From: Deep representation learning using layer-wise VICReg losses
Layer number | DB index | CH index |
|---|---|---|
1 | 3.360 | 532.289 |
2 | 3.339 | 565.690 |
3 | 3.287 | 592.728 |
4 | 3.264 | 533.800 |
