bert_uncased_L-2_H-256_A-4_wnli

This model is a fine-tuned version of google/bert_uncased_L-2_H-256_A-4 on the GLUE WNLI dataset. It achieves the following results on the evaluation set:

• Loss: 0.6966
• Accuracy: 0.5211

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
0.7188	1.0	3	0.7157	0.4085
0.6947	2.0	6	0.6966	0.5211
0.693	3.0	9	0.6977	0.5352
0.699	4.0	12	0.7026	0.5493
0.6941	5.0	15	0.7084	0.3944
0.6908	6.0	18	0.7167	0.3380
0.6915	7.0	21	0.7230	0.3239

Framework versions

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