bert-base-uncased-nsp-5000-1e-06-8

This model is a fine-tuned version of bert-base-uncased on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4119

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: 1e-06
• train_batch_size: 32
• eval_batch_size: 1024
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss
No log	1.0	157	0.6923
0.6979	2.0	314	0.6900
0.6928	3.0	471	0.6867
0.6891	4.0	628	0.6805
0.6891	5.0	785	0.6697
0.6805	6.0	942	0.6527
0.6637	7.0	1099	0.6301
0.6388	8.0	1256	0.6087
0.6103	9.0	1413	0.5909
0.6103	10.0	1570	0.5760
0.5846	11.0	1727	0.5624
0.5665	12.0	1884	0.5438
0.5322	13.0	2041	0.5249
0.5322	14.0	2198	0.5029
0.5096	15.0	2355	0.4886
0.4759	16.0	2512	0.4672
0.4457	17.0	2669	0.4581
0.4306	18.0	2826	0.4493
0.4306	19.0	2983	0.4413
0.4073	20.0	3140	0.4331
0.3974	21.0	3297	0.4267
0.3789	22.0	3454	0.4252
0.3615	23.0	3611	0.4200
0.3615	24.0	3768	0.4183
0.356	25.0	3925	0.4164
0.3588	26.0	4082	0.4152
0.3465	27.0	4239	0.4146
0.3465	28.0	4396	0.4133
0.3431	29.0	4553	0.4123
0.3451	30.0	4710	0.4119

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.0+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1