baseline_BERT_10K_steps

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

• Loss: 0.0356
• Accuracy: 0.9906
• Precision: 0.7827
• Recall: 0.0517
• F1: 0.0970
• Hamming: 0.0094

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: 2e-05
• train_batch_size: 6
• eval_batch_size: 6
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• training_steps: 10000

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1	Hamming
No log	0.0	500	0.1602	0.9902	0.0	0.0	0.0	0.0098
No log	0.0	1000	0.0573	0.9902	0.0	0.0	0.0	0.0098
No log	0.0	1500	0.0504	0.9902	0.0	0.0	0.0	0.0098
No log	0.01	2000	0.0492	0.9902	0.0	0.0	0.0	0.0098
No log	0.01	2500	0.0488	0.9902	0.0	0.0	0.0	0.0098
No log	0.01	3000	0.0485	0.9902	0.0	0.0	0.0	0.0098
No log	0.01	3500	0.0477	0.9902	0.0	0.0	0.0	0.0098
No log	0.01	4000	0.0467	0.9902	0.0	0.0	0.0	0.0098
No log	0.01	4500	0.0455	0.9902	0.0	0.0	0.0	0.0098
No log	0.01	5000	0.0442	0.9902	0.0	0.0	0.0	0.0098
No log	0.02	5500	0.0422	0.9902	0.0	0.0	0.0	0.0098
No log	0.02	6000	0.0408	0.9902	0.0	0.0	0.0	0.0098
No log	0.02	6500	0.0394	0.9902	0.0	0.0	0.0	0.0098
No log	0.02	7000	0.0385	0.9902	1.0	0.0011	0.0022	0.0098
No log	0.02	7500	0.0376	0.9903	0.7949	0.0057	0.0113	0.0097
No log	0.02	8000	0.0368	0.9903	0.8071	0.0146	0.0287	0.0097
No log	0.03	8500	0.0363	0.9905	0.7372	0.0465	0.0874	0.0095
No log	0.03	9000	0.0359	0.9905	0.7811	0.0381	0.0727	0.0095
No log	0.03	9500	0.0357	0.9906	0.8029	0.0562	0.1051	0.0094
0.0665	0.03	10000	0.0356	0.9906	0.7827	0.0517	0.0970	0.0094

Framework versions

• Transformers 4.37.2
• Pytorch 1.12.1+cu113
• Datasets 2.16.1
• Tokenizers 0.15.1