bert-ner-essays-find_span

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

• Loss: 0.1978
• B-span: {'precision': 0.8451327433628318, 'recall': 0.8856259659969088, 'f1-score': 0.8649056603773585, 'support': 647.0}
• I-span: {'precision': 0.9613473219215903, 'recall': 0.9557182067703568, 'f1-score': 0.9585244999082401, 'support': 10930.0}
• O: {'precision': 0.89764120320277, 'recall': 0.9040976460331299, 'f1-score': 0.9008578564447822, 'support': 4588.0}
• Accuracy: 0.9383
• Macro avg: {'precision': 0.9013737561623975, 'recall': 0.9151472729334652, 'f1-score': 0.9080960055767937, 'support': 16165.0}
• Weighted avg: {'precision': 0.9386145965884964, 'recall': 0.9382616764614908, 'f1-score': 0.9384103056993428, 'support': 16165.0}

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss	B-span	I-span	O	Accuracy	Macro avg	Weighted avg
No log	1.0	196	0.1948	{'precision': 0.8323076923076923, 'recall': 0.8361669242658424, 'f1-score': 0.8342328450269854, 'support': 647.0}	{'precision': 0.9544583371360774, 'recall': 0.9568161024702653, 'f1-score': 0.9556357655229132, 'support': 10930.0}	{'precision': 0.8977621763931549, 'recall': 0.8918918918918919, 'f1-score': 0.89481740651651, 'support': 4588.0}	0.9336	{'precision': 0.8948427352789748, 'recall': 0.894958306209333, 'f1-score': 0.8948953390221361, 'support': 16165.0}	{'precision': 0.9334776100904544, 'recall': 0.9335601608413239, 'f1-score': 0.9335149909678719, 'support': 16165.0}
No log	2.0	392	0.1840	{'precision': 0.8016528925619835, 'recall': 0.8995363214837713, 'f1-score': 0.8477785870356882, 'support': 647.0}	{'precision': 0.9520368530394725, 'recall': 0.9643183897529735, 'f1-score': 0.9581382664424344, 'support': 10930.0}	{'precision': 0.9198717948717948, 'recall': 0.8757628596338274, 'f1-score': 0.8972755694506476, 'support': 4588.0}	0.9366	{'precision': 0.8911871801577503, 'recall': 0.9132058569568574, 'f1-score': 0.9010641409762568, 'support': 16165.0}	{'precision': 0.936888587694453, 'recall': 0.9365914011753789, 'f1-score': 0.936446910650632, 'support': 16165.0}
0.2568	3.0	588	0.1978	{'precision': 0.8451327433628318, 'recall': 0.8856259659969088, 'f1-score': 0.8649056603773585, 'support': 647.0}	{'precision': 0.9613473219215903, 'recall': 0.9557182067703568, 'f1-score': 0.9585244999082401, 'support': 10930.0}	{'precision': 0.89764120320277, 'recall': 0.9040976460331299, 'f1-score': 0.9008578564447822, 'support': 4588.0}	0.9383	{'precision': 0.9013737561623975, 'recall': 0.9151472729334652, 'f1-score': 0.9080960055767937, 'support': 16165.0}	{'precision': 0.9386145965884964, 'recall': 0.9382616764614908, 'f1-score': 0.9384103056993428, 'support': 16165.0}

Framework versions

• Transformers 4.37.1
• Pytorch 2.1.2+cu121
• Datasets 2.16.1
• Tokenizers 0.15.1