scenario-KD-SCR-DIV2-data-glue-qnli-model-bert-base-uncased-run-1

This model is a fine-tuned version of bert-base-uncased on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.7514
• Accuracy: 0.8627

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 16
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 6969

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.4263	1.0	3273	1.6907	0.8545
1.7748	2.0	6547	1.8491	0.8499
1.1414	3.0	9820	1.9422	0.8545
0.8965	4.0	13094	1.7533	0.8552
0.7756	5.0	16367	1.7103	0.8570
0.6527	6.0	19641	1.6665	0.8569
0.6056	7.0	22914	1.5879	0.8620
0.5559	8.0	26188	1.6570	0.8618
0.5154	9.0	29461	1.5519	0.8658
0.4752	10.0	32735	1.6905	0.8612
0.4581	11.0	36008	1.6075	0.8644
0.4322	12.0	39282	1.6963	0.8614
0.3969	13.0	42555	1.6467	0.8660
0.393	14.0	45829	1.6735	0.8680
0.3651	15.0	49102	1.7631	0.8614
0.3464	16.0	52376	1.7957	0.8645
0.3455	17.0	55649	1.7008	0.8680
0.3276	18.0	58923	1.7183	0.8669
0.3239	19.0	62196	1.7514	0.8627

Framework versions

• Transformers 4.36.2
• Pytorch 2.1.2
• Datasets 2.16.0
• Tokenizers 0.15.0