spellcorrector_20_02_all_proportions_only_del_consonant_v15

This model is a fine-tuned version of google/canine-s on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0117
• Precision: 0.9973
• Recall: 0.9936
• F1: 0.9954
• Accuracy: 0.9969

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: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.3519	1.0	967	0.0926	0.9946	0.9855	0.9900	0.9797
0.0946	2.0	1934	0.0813	0.9935	0.9850	0.9892	0.9813
0.0845	3.0	2901	0.0720	0.9914	0.9855	0.9884	0.9827
0.0763	4.0	3868	0.0652	0.9849	0.9828	0.9839	0.9840
0.0714	5.0	4835	0.0589	0.9898	0.9855	0.9876	0.9855
0.0641	6.0	5802	0.0524	0.9941	0.9877	0.9909	0.9866
0.0592	7.0	6769	0.0465	0.9892	0.9871	0.9882	0.9878
0.0545	8.0	7736	0.0411	0.9930	0.9882	0.9906	0.9892
0.0505	9.0	8703	0.0367	0.9882	0.9866	0.9874	0.9902
0.0473	10.0	9670	0.0318	0.9951	0.9893	0.9922	0.9914
0.0426	11.0	10637	0.0280	0.9941	0.9898	0.9919	0.9924
0.0406	12.0	11604	0.0243	0.9946	0.9893	0.9919	0.9933
0.0369	13.0	12571	0.0218	0.9968	0.9914	0.9941	0.9942
0.0343	14.0	13538	0.0192	0.9973	0.9941	0.9957	0.9947
0.0309	15.0	14505	0.0172	0.9957	0.9930	0.9944	0.9954
0.0289	16.0	15472	0.0152	0.9957	0.9925	0.9941	0.9959
0.028	17.0	16439	0.0138	0.9957	0.9936	0.9946	0.9963
0.0264	18.0	17406	0.0125	0.9973	0.9936	0.9954	0.9967
0.025	19.0	18373	0.0121	0.9973	0.9936	0.9954	0.9968
0.0244	20.0	19340	0.0117	0.9973	0.9936	0.9954	0.9969

Framework versions

• Transformers 4.37.2
• Pytorch 2.1.0+cu121
• Datasets 2.17.1
• Tokenizers 0.15.2