favs_token_classification_v2_uncased

This model is a fine-tuned version of bert-base-uncased on the token_classification_v2 dataset. It achieves the following results on the evaluation set:

• Loss: 0.5006
• Precision: 0.6599
• Recall: 0.7823
• F1: 0.7159
• Accuracy: 0.8547

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: 1.5e-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	Precision	Recall	F1	Accuracy
2.2447	1.0	13	1.9089	0.2	0.0968	0.1304	0.3576
1.9589	2.0	26	1.5848	0.2734	0.2823	0.2778	0.4477
1.729	3.0	39	1.3636	0.3128	0.4516	0.3696	0.6134
1.4278	4.0	52	1.1854	0.4302	0.5968	0.5	0.7122
1.3046	5.0	65	1.0341	0.5183	0.6855	0.5903	0.7413
1.1599	6.0	78	0.9163	0.5188	0.6694	0.5845	0.75
0.9263	7.0	91	0.8235	0.5399	0.7097	0.6132	0.7645
0.8721	8.0	104	0.7627	0.5176	0.7097	0.5986	0.7733
0.7879	9.0	117	0.7070	0.5366	0.7097	0.6111	0.7849
0.6881	10.0	130	0.6575	0.5427	0.7177	0.6181	0.7936
0.6414	11.0	143	0.6076	0.5660	0.7258	0.6360	0.8110
0.6096	12.0	156	0.5804	0.6090	0.7661	0.6786	0.8285
0.5812	13.0	169	0.5661	0.6282	0.7903	0.7000	0.8343
0.5006	14.0	182	0.5503	0.6144	0.7581	0.6787	0.8285
0.5289	15.0	195	0.5366	0.6267	0.7581	0.6861	0.8372
0.4447	16.0	208	0.5222	0.6419	0.7661	0.6985	0.8459
0.435	17.0	221	0.5120	0.6599	0.7823	0.7159	0.8517
0.4454	18.0	234	0.5058	0.6667	0.7903	0.7232	0.8547
0.422	19.0	247	0.5013	0.6599	0.7823	0.7159	0.8547
0.4285	20.0	260	0.5006	0.6599	0.7823	0.7159	0.8547

Framework versions

• Transformers 4.21.1
• Pytorch 1.12.1
• Datasets 2.4.0
• Tokenizers 0.12.1