distilbert-base-uncased-TASTESet-ner

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

• Loss: 0.3816
• Precision: 0.8929
• Recall: 0.9229
• F1: 0.9076
• Accuracy: 0.9130

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	Precision	Recall	F1	Accuracy
No log	1.0	31	1.0797	0.6027	0.6903	0.6435	0.7063
No log	2.0	62	0.6402	0.7681	0.8295	0.7976	0.8304
No log	3.0	93	0.4899	0.8379	0.8789	0.8579	0.8728
No log	4.0	124	0.4232	0.8716	0.8994	0.8853	0.8912
No log	5.0	155	0.3883	0.8798	0.9043	0.8919	0.8992
No log	6.0	186	0.3848	0.8769	0.9103	0.8933	0.9004
No log	7.0	217	0.3684	0.8864	0.9123	0.8991	0.9046
No log	8.0	248	0.3650	0.8930	0.9182	0.9054	0.9087
No log	9.0	279	0.3628	0.8908	0.9197	0.9050	0.9096
No log	10.0	310	0.3674	0.8933	0.9165	0.9047	0.9093
No log	11.0	341	0.3668	0.8958	0.9177	0.9066	0.9120
No log	12.0	372	0.3717	0.8904	0.9234	0.9066	0.9120
No log	13.0	403	0.3693	0.8940	0.9197	0.9067	0.9126
No log	14.0	434	0.3805	0.8913	0.9239	0.9073	0.9135
No log	15.0	465	0.3788	0.8954	0.9202	0.9076	0.9123
No log	16.0	496	0.3803	0.8935	0.9231	0.9081	0.9122
0.3275	17.0	527	0.3814	0.8918	0.9229	0.9071	0.9126
0.3275	18.0	558	0.3823	0.8921	0.9241	0.9079	0.9123
0.3275	19.0	589	0.3827	0.8928	0.9224	0.9074	0.9124
0.3275	20.0	620	0.3816	0.8929	0.9229	0.9076	0.9130

Framework versions

• Transformers 4.26.0
• Pytorch 1.13.1+cu117
• Datasets 2.9.0
• Tokenizers 0.13.2