bert-base-uncased-issues-128

This model is a fine-tuned version of bert-base-uncased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0420
• Micro f1: 0.6972
• Macro f1: 0.5740

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: 3e-05
• train_batch_size: 4
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Micro f1	Macro f1
0.1542	1.0	404	0.0850	0.0	0.0
0.0853	2.0	808	0.0826	0.0	0.0
0.0828	3.0	1212	0.0776	0.0	0.0
0.0734	4.0	1616	0.0673	0.1735	0.0294
0.061	5.0	2020	0.0597	0.3500	0.0797
0.049	6.0	2424	0.0556	0.4449	0.1180
0.0398	7.0	2828	0.0497	0.5027	0.1615
0.0321	8.0	3232	0.0472	0.5609	0.2498
0.0261	9.0	3636	0.0481	0.5212	0.2442
0.0218	10.0	4040	0.0441	0.6238	0.3251
0.0182	11.0	4444	0.0423	0.6346	0.3880
0.0154	12.0	4848	0.0412	0.6573	0.4223
0.0126	13.0	5252	0.0419	0.6586	0.4570
0.0109	14.0	5656	0.0416	0.6646	0.4813
0.0092	15.0	6060	0.0401	0.6677	0.5041
0.0079	16.0	6464	0.0412	0.6859	0.5323
0.0068	17.0	6868	0.0435	0.6667	0.5372
0.0059	18.0	7272	0.0452	0.6735	0.5617
0.0053	19.0	7676	0.0420	0.6735	0.5590
0.0042	20.0	8080	0.0420	0.6972	0.5740

Framework versions

• Transformers 4.26.1
• Pytorch 1.13.1
• Datasets 2.10.1
• Tokenizers 0.11.0