legalbert-large-1.7M-2_class_actions

This model is a fine-tuned version of pile-of-law/legalbert-large-1.7M-2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.6428
• Accuracy: 0.61

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: linear
• num_epochs: 14

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	150	0.6380	0.6333
No log	2.0	300	0.7457	0.55
No log	3.0	450	0.7066	0.45
0.6843	4.0	600	0.7218	0.6767
0.6843	5.0	750	0.6360	0.6067
0.6843	6.0	900	0.6502	0.6033
0.6751	7.0	1050	0.6664	0.6033
0.6751	8.0	1200	0.6490	0.6133
0.6751	9.0	1350	0.6506	0.6067
0.6781	10.0	1500	0.6486	0.61
0.6781	11.0	1650	0.6544	0.6167
0.6781	12.0	1800	0.6425	0.61
0.6781	13.0	1950	0.6417	0.61
0.6756	14.0	2100	0.6428	0.61

Framework versions

• Transformers 4.29.2
• Pytorch 2.0.1+cu118
• Datasets 2.12.0
• Tokenizers 0.13.3