scenario-kd-pre-ner-full-xlmr-halfen_data-univner_en66

This model is a fine-tuned version of FacebookAI/xlm-roberta-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 51.6812
• Precision: 0.7610
• Recall: 0.7712
• F1: 0.7661
• Accuracy: 0.9817

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
107.359	1.28	500	73.7771	0.7295	0.7008	0.7149	0.9786
64.0888	2.55	1000	62.1069	0.7445	0.7298	0.7371	0.9808
56.4531	3.83	1500	57.7062	0.7495	0.7340	0.7416	0.9805
52.7856	5.1	2000	54.8858	0.7482	0.7598	0.7540	0.9814
50.5437	6.38	2500	53.2901	0.7482	0.7505	0.7494	0.9811
49.0825	7.65	3000	52.2441	0.7503	0.7557	0.7530	0.9815
48.2494	8.93	3500	51.6812	0.7610	0.7712	0.7661	0.9817

Framework versions

• Transformers 4.33.3
• Pytorch 2.1.1+cu121
• Datasets 2.14.5
• Tokenizers 0.13.3