Table 1 Search space and final results for hyperparameter optimization.
From: Diagnostic forecasting of battery degradation through contrastive learning
Hyperparameter | Symbol | Search Space |
|---|---|---|
Embedding dimension | \(d_{\text {embed}}\) | \(\{64, 128, 256\}\) |
Queue size | q | \(\{512, 1024, 2048, 4096\}\) |
Conv hidden size | \(h_{\text {conv}}\) | \(\{128, 256, 512\}\) |
Operational hidden size | \(h_{\text {op}}\) | \(\{64, 128, 256\}\) |
Attention head size | a | \(\{1, 2, 4\}\) |
Dropout rate | \(\delta\) | \(\{0.05, 0.2\}\) |
