bottleneckBERTsmall

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

• Loss: 1.7726

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: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss
2.469	0.2565	500	2.3210
2.308	0.5131	1000	2.1745
2.2322	0.7696	1500	2.0982
2.1424	1.0262	2000	2.0600
2.1064	1.2827	2500	2.0712
2.1129	1.5393	3000	2.0012
2.0676	1.7958	3500	1.9624
1.9941	2.0523	4000	1.9267
1.9514	2.3089	4500	1.9082
1.9382	2.5654	5000	1.8936
1.8954	2.8220	5500	1.8489
1.8913	3.0785	6000	1.8282
1.8724	3.3350	6500	1.8323
1.8135	3.5916	7000	1.8146
1.8178	3.8481	7500	1.7984
1.7824	4.1047	8000	1.7734
1.7669	4.3612	8500	1.7677
1.7687	4.6178	9000	1.7556
1.7572	4.8743	9500	1.7726

Framework versions

• Transformers 4.49.0.dev0
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0