rubert-tiny2-ner-drugname

This model is a fine-tuned version of cointegrated/rubert-tiny2 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0549
• Precision: 0.7232
• Recall: 0.7690
• F1: 0.7454
• Accuracy: 0.9883

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: 0.0002
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	61	0.0493	0.6413	0.7468	0.6901	0.9833
No log	2.0	122	0.0417	0.6406	0.8291	0.7228	0.9855
No log	3.0	183	0.0387	0.7588	0.7468	0.7528	0.9879
No log	4.0	244	0.0396	0.7385	0.7595	0.7488	0.9883
No log	5.0	305	0.0425	0.6897	0.7595	0.7229	0.9874
No log	6.0	366	0.0465	0.6991	0.7722	0.7338	0.9876
No log	7.0	427	0.0487	0.7062	0.7911	0.7463	0.9877
No log	8.0	488	0.0521	0.7076	0.7658	0.7356	0.9882
0.0306	9.0	549	0.0540	0.7262	0.7722	0.7485	0.9883
0.0306	10.0	610	0.0549	0.7232	0.7690	0.7454	0.9883

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1