hBERTv2_new_pretrain_w_init_48_wnli

This model is a fine-tuned version of gokuls/bert_12_layer_model_v2_complete_training_new_wt_init_48 on the GLUE WNLI dataset. It achieves the following results on the evaluation set:

• Loss: 0.6864
• Accuracy: 0.5634

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: 4e-05
• train_batch_size: 128
• eval_batch_size: 128
• seed: 10
• distributed_type: multi-GPU
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.9651	1.0	5	0.6864	0.5634
0.7199	2.0	10	0.6992	0.5634
0.7059	3.0	15	0.7100	0.3380
0.7029	4.0	20	0.7014	0.4648
0.7045	5.0	25	0.7246	0.4366
0.7259	6.0	30	0.7128	0.5634

Framework versions

• Transformers 4.29.2
• Pytorch 1.14.0a0+410ce96
• Datasets 2.12.0
• Tokenizers 0.13.3