bert-base-chinese-chn-finetuned-augmentation-LUNAR

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

• Loss: 0.2282
• F1: 0.7890
• Roc Auc: 0.8637
• Accuracy: 0.7323

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: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 100
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Roc Auc	Accuracy
0.2216	1.0	315	0.2200	0.5555	0.7352	0.5949
0.1695	2.0	630	0.1692	0.6542	0.7784	0.6839
0.1031	3.0	945	0.1674	0.6900	0.8028	0.6926
0.0671	4.0	1260	0.1707	0.7356	0.8239	0.7085
0.0415	5.0	1575	0.1884	0.7489	0.8419	0.7014
0.0289	6.0	1890	0.1993	0.7604	0.8532	0.6998
0.0204	7.0	2205	0.2331	0.7568	0.8558	0.6791
0.014	8.0	2520	0.2070	0.7714	0.8467	0.7149
0.0069	9.0	2835	0.2256	0.7823	0.8684	0.7053
0.0055	10.0	3150	0.2207	0.7839	0.8611	0.7260
0.0064	11.0	3465	0.2197	0.7875	0.8597	0.7252
0.0061	12.0	3780	0.2282	0.7890	0.8637	0.7323
0.0046	13.0	4095	0.2316	0.7865	0.8584	0.7284
0.0022	14.0	4410	0.2339	0.7763	0.8519	0.7307
0.0025	15.0	4725	0.2339	0.7800	0.8536	0.7315
0.0028	16.0	5040	0.2328	0.7802	0.8537	0.7299

Framework versions

• Transformers 4.45.1
• Pytorch 2.4.0
• Datasets 3.0.1
• Tokenizers 0.20.0