aspect_complaint_spanbert

This model is a fine-tuned version of SpanBERT/spanbert-large-cased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.2032
• F1: 0.8368
• Roc Auc: 0.8935
• Accuracy: 0.5292

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: 48
• eval_batch_size: 48
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Roc Auc	Accuracy
No log	1.0	97	0.2810	0.6988	0.7718	0.0
No log	2.0	194	0.2209	0.7688	0.8260	0.2577
No log	3.0	291	0.2017	0.7958	0.8477	0.3565
No log	4.0	388	0.1882	0.8120	0.8678	0.4184
No log	5.0	485	0.1773	0.8272	0.8693	0.4416
0.2417	6.0	582	0.1732	0.8262	0.8792	0.4811
0.2417	7.0	679	0.1719	0.8331	0.8795	0.4759
0.2417	8.0	776	0.1722	0.8341	0.8814	0.5017
0.2417	9.0	873	0.1797	0.8347	0.8818	0.4923
0.2417	10.0	970	0.1856	0.8328	0.8859	0.5112
0.1287	11.0	1067	0.1866	0.8343	0.8868	0.5275
0.1287	12.0	1164	0.1868	0.8378	0.8898	0.5258
0.1287	13.0	1261	0.1923	0.8322	0.8864	0.5120
0.1287	14.0	1358	0.1916	0.8400	0.8929	0.5387
0.1287	15.0	1455	0.1975	0.8368	0.8919	0.5421
0.0781	16.0	1552	0.1974	0.8403	0.8951	0.5369
0.0781	17.0	1649	0.2022	0.8329	0.8911	0.5241
0.0781	18.0	1746	0.2012	0.8360	0.8921	0.5284
0.0781	19.0	1843	0.2027	0.8367	0.8936	0.5335
0.0781	20.0	1940	0.2032	0.8368	0.8935	0.5292

Framework versions

• Transformers 4.41.1
• Pytorch 1.13.1+cu117
• Datasets 2.19.1
• Tokenizers 0.19.1