N_bert_twitterfin_padding10model

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: 0.9504
• Accuracy: 0.8941

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: 2e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.5941	1.0	597	0.3676	0.8681
0.3264	2.0	1194	0.3223	0.8886
0.2271	3.0	1791	0.4276	0.8886
0.1373	4.0	2388	0.5792	0.8819
0.0979	5.0	2985	0.6505	0.8832
0.0411	6.0	3582	0.7322	0.8878
0.0376	7.0	4179	0.7613	0.8807
0.022	8.0	4776	0.7982	0.8894
0.0217	9.0	5373	0.8054	0.8886
0.0266	10.0	5970	0.8280	0.8932
0.0142	11.0	6567	0.8836	0.8857
0.0062	12.0	7164	0.8788	0.8907
0.0119	13.0	7761	0.8796	0.8941
0.0031	14.0	8358	0.8968	0.8903
0.0096	15.0	8955	0.8962	0.8915
0.0027	16.0	9552	0.9295	0.8945
0.0024	17.0	10149	0.9298	0.8961
0.0027	18.0	10746	0.9663	0.8932
0.0017	19.0	11343	0.9372	0.8932
0.0024	20.0	11940	0.9504	0.8941

Framework versions

• Transformers 4.33.2
• Pytorch 2.0.1+cu117
• Datasets 2.14.5
• Tokenizers 0.13.3