This model is a fine-tuned version of google/bert_uncased_L-4_H-128_A-2 on the GLUE QQP dataset. It achieves the following results on the evaluation set:

Loss: 0.3084
Accuracy: 0.8694
F1: 0.8238
Combined Score: 0.8466

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
0.448	1.0	1422	0.3785	0.8204	0.7736	0.7970
0.3757	2.0	2844	0.3562	0.8351	0.7980	0.8165
0.3453	3.0	4266	0.3263	0.8521	0.8052	0.8286
0.3226	4.0	5688	0.3198	0.8560	0.8090	0.8325
0.3049	5.0	7110	0.3245	0.8552	0.8159	0.8355
0.2903	6.0	8532	0.3136	0.8602	0.8207	0.8405
0.277	7.0	9954	0.3188	0.8603	0.8200	0.8401
0.2667	8.0	11376	0.3146	0.8648	0.8258	0.8453
0.2551	9.0	12798	0.3239	0.8612	0.8240	0.8426
0.2444	10.0	14220	0.3084	0.8694	0.8238	0.8466
0.2358	11.0	15642	0.3224	0.8669	0.8296	0.8483
0.2276	12.0	17064	0.3147	0.8699	0.8313	0.8506
0.2191	13.0	18486	0.3185	0.8708	0.8332	0.8520
0.2111	14.0	19908	0.3282	0.8721	0.8337	0.8529
0.2048	15.0	21330	0.3194	0.8738	0.8306	0.8522

Framework versions

Transformers 4.46.3
Pytorch 2.2.1+cu118
Datasets 2.17.0
Tokenizers 0.20.3

gokulsrinivasagan/bert_uncased_L-4_H-128_A-2_qqp

bert_uncased_L-4_H-128_A-2_qqp

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

gokulsrinivasagan/
bert_uncased_L-4_H-128_A-2_qqp