BERT-legal-de-cased_German_legal_SQuAD_part_augmented_17

This model was trained from scratch on the None dataset. It achieves the following results on the evaluation set:

• Loss: 3.6973

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: 160
• eval_batch_size: 40
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 17

Training results

Training Loss	Epoch	Step	Validation Loss
No log	1.0	3	6.1820
No log	2.0	6	6.1072
No log	3.0	9	6.0006
No log	4.0	12	5.5083
No log	5.0	15	5.2355
No log	6.0	18	5.0300
No log	7.0	21	4.7332
No log	8.0	24	4.4631
No log	9.0	27	4.3110
No log	10.0	30	4.1210
No log	11.0	33	3.9867
No log	12.0	36	3.9081
No log	13.0	39	3.8287
No log	14.0	42	3.7669
No log	15.0	45	3.7289
No log	16.0	48	3.7066
No log	17.0	51	3.6973

Framework versions

• Transformers 4.36.2
• Pytorch 2.1.2+cu121
• Datasets 2.14.7
• Tokenizers 0.15.0