SOMD-train-bert-v1

This model is a fine-tuned version of google-bert/bert-base-cased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0001
• F1: 0.9914

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: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss	F1
No log	1.0	1243	0.0183	0.3245
No log	2.0	2486	0.0118	0.4825
No log	3.0	3729	0.0103	0.5138
No log	4.0	4972	0.0085	0.5191
No log	5.0	6215	0.0064	0.6228
No log	6.0	7458	0.0053	0.6666
No log	7.0	8701	0.0046	0.6844
No log	8.0	9944	0.0043	0.7041
No log	9.0	11187	0.0036	0.7175
No log	10.0	12430	0.0029	0.7732
No log	11.0	13673	0.0022	0.8057
No log	12.0	14916	0.0027	0.7894
No log	13.0	16159	0.0015	0.8652
No log	14.0	17402	0.0016	0.8579
No log	15.0	18645	0.0012	0.8958
No log	16.0	19888	0.0009	0.9167
No log	17.0	21131	0.0008	0.9395
No log	18.0	22374	0.0006	0.9535
No log	19.0	23617	0.0004	0.9607
No log	20.0	24860	0.0003	0.9712
No log	21.0	26103	0.0003	0.9794
No log	22.0	27346	0.0002	0.9858
No log	23.0	28589	0.0002	0.9888
No log	24.0	29832	0.0001	0.9898
No log	25.0	31075	0.0001	0.9914

Framework versions

• Transformers 4.37.0
• Pytorch 2.1.2
• Datasets 2.1.0
• Tokenizers 0.15.1