|
55 | 55 | (AdamP, {'lr': 5e-1, 'weight_decay': 1e-3, 'nesterov': True}, 100), |
56 | 56 | (DiffGrad, {'lr': 5e-1, 'weight_decay': 1e-3}, 100), |
57 | 57 | (DiffRGrad, {'lr': 5e-1, 'weight_decay': 1e-3}, 100), |
58 | | - (Lamb, {'lr': 1e-1, 'weight_decay': 1e-3}, 300), |
59 | | - (Lamb, {'lr': 1e-1, 'weight_decay': 1e-3, 'adam': True, 'eps': 1e-8}, 100), |
| 58 | + (Lamb, {'lr': 1e-1, 'weight_decay': 1e-3}, 100), |
| 59 | + (Lamb, {'lr': 1e-1, 'weight_decay': 1e-3, 'adam': True, 'eps': 1e-8}, 300), |
60 | 60 | (Lamb, {'lr': 1e-1, 'weight_decay': 1e-3, 'pre_norm': True, 'eps': 1e-8}, 100), |
61 | 61 | (LARS, {'lr': 1e-1, 'weight_decay': 1e-3}, 300), |
62 | 62 | (RaLamb, {'lr': 1e-1, 'weight_decay': 1e-4}, 100), |
|
0 commit comments