Knowledge Continuity Regularized Network
Dataset: ANLI Round: None
Trainer Hyperparameters:
lr
= 5e-05per_device_batch_size
= 8gradient_accumulation_steps
= 2weight_decay
= 1e-09seed
= 42
Regularization Hyperparameters
numerical stability denominator constant
= 1.0lambda
= 1.0alpha
= 1.0beta
= 1.0
Extended Logs:
eval_loss | eval_accuracy | epoch |
---|---|---|
1.176 | 0.364 | 1.0 |
1.142 | 0.393 | 2.0 |
1.140 | 0.402 | 3.0 |
1.143 | 0.396 | 4.0 |
1.125 | 0.412 | 5.0 |
1.152 | 0.392 | 6.0 |
1.134 | 0.407 | 7.0 |
1.140 | 0.407 | 8.0 |
1.128 | 0.420 | 9.0 |
1.145 | 0.393 | 10.0 |
1.117 | 0.431 | 11.0 |
1.122 | 0.426 | 12.0 |
1.111 | 0.434 | 13.0 |
1.130 | 0.418 | 14.0 |
1.122 | 0.428 | 15.0 |
1.115 | 0.431 | 16.0 |
1.110 | 0.437 | 17.0 |
1.104 | 0.440 | 18.0 |
1.094 | 0.450 | 19.0 |
Test Accuracy: 0.456
- Downloads last month
- 0
Unable to determine this model’s pipeline type. Check the
docs
.