scenario-kd-po-ner-full-xlmr_data-univner_en66

This model is a fine-tuned version of haryoaw/scenario-TCR-NER_data-univner_en on the None dataset. It achieves the following results on the evaluation set:

• Loss: 53.9174
• Precision: 0.7598
• Recall: 0.7433
• F1: 0.7514
• Accuracy: 0.9811

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: 3e-05
• train_batch_size: 8
• eval_batch_size: 32
• seed: 66
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• 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
87.1202	1.2755	500	72.9152	0.7420	0.6967	0.7186	0.9789
64.1514	2.5510	1000	64.7342	0.7457	0.7195	0.7323	0.9795
58.039	3.8265	1500	60.8792	0.7492	0.7143	0.7313	0.9794
54.4866	5.1020	2000	57.6543	0.7537	0.7474	0.7505	0.9811
52.1316	6.3776	2500	56.1872	0.7638	0.7329	0.7480	0.9804
50.5995	7.6531	3000	54.8584	0.7544	0.7505	0.7525	0.9810
49.7948	8.9286	3500	53.9174	0.7598	0.7433	0.7514	0.9811

Framework versions

• Transformers 4.44.2
• Pytorch 2.1.1+cu121
• Datasets 2.14.5
• Tokenizers 0.19.1