boss-sentiment-12000-bert-base-uncased

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

• F1: 0.7380
• Acc: 0.8773
• Loss: 0.9558

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: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 20

Training results

Training Loss	Epoch	Step	F1	Acc	Validation Loss
0.929	1.0	750	0.6983	0.8609	0.4073
0.481	2.0	1500	0.7179	0.8689	0.3605
0.3564	3.0	2250	0.7269	0.8703	0.3834
0.2369	4.0	3000	0.7006	0.8465	0.5631
0.1536	5.0	3750	0.7237	0.8591	0.6596
0.1228	6.0	4500	0.7285	0.8660	0.7316
0.0831	7.0	5250	0.7454	0.8817	0.6420
0.0687	8.0	6000	0.6955	0.8354	1.1172
0.0541	9.0	6750	0.7143	0.8479	1.0556
0.0465	10.0	7500	0.7473	0.8889	0.7691
0.0404	11.0	8250	0.7209	0.8636	1.0274
0.0315	12.0	9000	0.7082	0.8401	1.2706
0.0329	13.0	9750	0.7380	0.8773	0.9558

Framework versions

• Transformers 4.35.0
• Pytorch 2.1.0+cu121
• Datasets 2.14.6
• Tokenizers 0.14.1