spellcorrector_17_02_050_qwerty

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: 0.0163
• Precision: 0.9930
• Recall: 0.9887
• F1: 0.9909
• Accuracy: 0.9952

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
0.4639	1.0	967	0.1649	0.9619	0.9624	0.9622	0.9608
0.1737	2.0	1934	0.1300	0.9620	0.9656	0.9638	0.9664
0.145	3.0	2901	0.1099	0.9678	0.9694	0.9686	0.9708
0.1222	4.0	3868	0.0906	0.9699	0.9699	0.9699	0.9752
0.105	5.0	4835	0.0736	0.9726	0.9699	0.9712	0.9792
0.0933	6.0	5802	0.0633	0.9758	0.9732	0.9745	0.9817
0.0807	7.0	6769	0.0531	0.9822	0.9780	0.9801	0.9844
0.0715	8.0	7736	0.0468	0.9839	0.9828	0.9834	0.9864
0.0643	9.0	8703	0.0404	0.9833	0.9823	0.9828	0.9880
0.0575	10.0	9670	0.0356	0.9903	0.9866	0.9884	0.9894
0.0525	11.0	10637	0.0317	0.9887	0.9866	0.9876	0.9905
0.0481	12.0	11604	0.0281	0.9908	0.9871	0.9890	0.9915
0.0444	13.0	12571	0.0255	0.9919	0.9871	0.9895	0.9923
0.0417	14.0	13538	0.0235	0.9924	0.9877	0.9900	0.9930
0.0382	15.0	14505	0.0211	0.9925	0.9882	0.9903	0.9937
0.0358	16.0	15472	0.0198	0.9930	0.9887	0.9909	0.9941
0.034	17.0	16439	0.0184	0.9930	0.9882	0.9906	0.9946
0.0323	18.0	17406	0.0173	0.9930	0.9882	0.9906	0.9949
0.0306	19.0	18373	0.0167	0.9930	0.9887	0.9909	0.9951
0.0304	20.0	19340	0.0163	0.9930	0.9887	0.9909	0.9952

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.17.0
• Tokenizers 0.15.2