tinybert_train_book_v2_qqp

This model is a fine-tuned version of gokulsrinivasagan/tinybert_train_book_v2 on the GLUE QQP dataset. It achieves the following results on the evaluation set:

• Loss: 0.2906
• Accuracy: 0.8763
• F1: 0.8292
• Combined Score: 0.8528

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	F1	Combined Score
0.397	1.0	1422	0.3445	0.8484	0.7789	0.8137
0.2958	2.0	2844	0.3029	0.8695	0.8221	0.8458
0.2432	3.0	4266	0.2906	0.8763	0.8292	0.8528
0.1995	4.0	5688	0.2984	0.8798	0.8315	0.8557
0.1629	5.0	7110	0.3093	0.8847	0.8454	0.8651
0.1334	6.0	8532	0.3454	0.8849	0.8435	0.8642
0.1097	7.0	9954	0.3595	0.8847	0.8452	0.8649
0.0925	8.0	11376	0.3933	0.8854	0.8441	0.8647

Framework versions

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