dal-nlp

Assenmacher-mat's group workspace

Group: distilbert_coreset_fnc1_nosub

806

440

1-5

of 5

Timestamps visible

2023-02-02 16:20:04

Epoch [2]: Train Loss: 0.5305, Train Accuracy: 84.0000

2023-02-02 16:20:04

--------------------------------------------------------------------------------

2023-02-02 16:20:04

Epoch [3] [0/5] eta: 0:00:01 lr: 0.00001957 loss: 0.6692 (0.6692) batch_acc: 80.0000 (80.0000) time: 0.3334 data: 0.0073 max mem: 7437

2023-02-02 16:20:04

Epoch [3] Total time: 0:00:01

2023-02-02 16:20:04

Epoch [3]: Train Loss: 0.5071, Train Accuracy: 84.0000

2023-02-02 16:20:04

--------------------------------------------------------------------------------

2023-02-02 16:20:06

Epoch [4] [0/5] eta: 0:00:01 lr: 0.00000870 loss: 0.6335 (0.6335) batch_acc: 80.0000 (80.0000) time: 0.3364 data: 0.0068 max mem: 7437

2023-02-02 16:20:06

Epoch [4] Total time: 0:00:01

2023-02-02 16:20:06

Epoch [4]: Train Loss: 0.5057, Train Accuracy: 84.0000

2023-02-02 16:20:06

--------------------------------------------------------------------------------

2023-02-02 16:20:06

[2023-02-02 17:20:05,469][root][INFO] - Training took 0.17 minutes

2023-02-02 16:20:06

[2023-02-02 17:20:05,470][root][INFO] - Training stats: {'train_lr_epoch': 4.347826086956522e-06, 'train_loss_epoch': 0.50566745698452, 'train_batch_acc_epoch': 84.0}

2023-02-02 16:20:06

> Evaluation.

2023-02-02 16:20:06

[2023-02-02 17:20:05,470][root][INFO] - Evaluation with 4998 sample

2023-02-02 16:20:06

Testing: [ 0/63] eta: 0:00:24 loss: 0.9598 (0.9598) batch_acc: 70.0000 (70.0000) batch_f1: 0.2745 (0.2745) batch_acc_balanced: 0.3333 (0.3333) time: 0.3933 data: 0.0258 max mem: 7437

2023-02-02 16:20:16

Testing: [25/63] eta: 0:00:14 loss: 0.8957 (0.8804) batch_acc: 72.5000 (73.6538) batch_f1: 0.2122 (0.2312) batch_acc_balanced: 0.2500 (0.2724) time: 0.3898 data: 0.0246 max mem: 7437