bert-finetuned-ner-synonym-replacement-model

This model is a fine-tuned version of bert-base-cased on the combined training dataset(tweetner7(train_2021)+augmented dataset(train_2021) using synonym replacment technique (lsoni/combined_tweetner7_synonym_replacement_augmented_dataset) and dataset used for evaluation is combined evaluation dataset(tweetner7(validation_2021)+augmented dataset(validation_2021) using synonym replacment technique (lsoni/combined_tweetner7_synonym_replacement_augmented_dataset_eval). It achieves the following results on the evaluation set:

• Loss: 0.4484
• Precision: 0.6804
• Recall: 0.6727
• F1: 0.6765
• Accuracy: 0.8780

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	Precision	Recall	F1	Accuracy
0.5525	1.0	624	0.4142	0.7040	0.6120	0.6548	0.8774
0.3293	2.0	1248	0.4101	0.7067	0.6628	0.6841	0.8833
0.2536	3.0	1872	0.4484	0.6804	0.6727	0.6765	0.8780

Framework versions

• Transformers 4.20.1
• Pytorch 1.12.1
• Datasets 2.10.1
• Tokenizers 0.12.1