bert-tiny-finetuned-pile-of-law-tos

This model is a MLM fine-tuned version of google/bert_uncased_L-2_H-128_A-2 on the pile-of-law/tos dataset. It achieves the following results on the evaluation set:

• Loss: 3.3545

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss
No log	1.0	264	3.5896
3.8119	2.0	528	3.5598
3.8119	3.0	792	3.5263
3.7028	4.0	1056	3.4982
3.7028	5.0	1320	3.5170
3.6286	6.0	1584	3.5143
3.6286	7.0	1848	3.4477
3.553	8.0	2112	3.4044
3.553	9.0	2376	3.4670
3.5179	10.0	2640	3.3991
3.5179	11.0	2904	3.4330
3.4784	12.0	3168	3.4671
3.4784	13.0	3432	3.3489
3.4535	14.0	3696	3.4354
3.4535	15.0	3960	3.4023

Framework versions

• Transformers 4.21.0
• Pytorch 1.12.0+cu113
• Datasets 2.4.0
• Tokenizers 0.12.1