distilbert-base-uncased-multil-cls-legal

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

• Loss: 0.5448
• Accuracy: 0.9022
• F1: 0.9015

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: 32
• eval_batch_size: 32
• 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	Accuracy	F1
2.67	1.0	396	1.9327	0.5209	0.4806
1.5362	2.0	792	1.0998	0.7061	0.6869
0.8991	3.0	1188	0.7546	0.8013	0.7975
0.5899	4.0	1584	0.6136	0.8403	0.8392
0.4082	5.0	1980	0.5527	0.8601	0.8589
0.2874	6.0	2376	0.5200	0.8736	0.8731
0.2136	7.0	2772	0.4991	0.8831	0.8815
0.1564	8.0	3168	0.4946	0.8853	0.8843
0.1123	9.0	3564	0.4814	0.8928	0.8920
0.0866	10.0	3960	0.4959	0.8912	0.8908
0.0685	11.0	4356	0.5060	0.8928	0.8923
0.0508	12.0	4752	0.5114	0.8997	0.8989
0.037	13.0	5148	0.5199	0.8978	0.8971
0.0316	14.0	5544	0.5236	0.9003	0.8993
0.0243	15.0	5940	0.5253	0.9022	0.9015
0.021	16.0	6336	0.5385	0.9025	0.9019
0.0177	17.0	6732	0.5396	0.9038	0.9032
0.014	18.0	7128	0.5449	0.9025	0.9018
0.014	19.0	7524	0.5467	0.9010	0.9002
0.0103	20.0	7920	0.5448	0.9022	0.9015

Framework versions

• Transformers 4.31.0
• Pytorch 2.0.1+cu118
• Datasets 2.14.4
• Tokenizers 0.13.3