bert_uncased_L-2_H-128_A-2_wnli

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

• Loss: 0.6972
• Accuracy: 0.4648

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: 5e-05
• train_batch_size: 256
• eval_batch_size: 256
• seed: 10
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.6962	1.0	3	0.6984	0.4507
0.6909	2.0	6	0.6972	0.4648
0.6962	3.0	9	0.6980	0.4507
0.6922	4.0	12	0.6989	0.4507
0.6926	5.0	15	0.6998	0.4366
0.6907	6.0	18	0.7011	0.4085
0.6946	7.0	21	0.7024	0.4085

Framework versions

• Transformers 4.46.3
• Pytorch 2.2.1+cu118
• Datasets 2.17.0
• Tokenizers 0.20.3