spellcorrector_0511_v2

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: 0.1552
• Precision: 0.9703
• Recall: 0.9736
• F1: 0.9720
• Accuracy: 0.9734

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: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.2251	1.0	1945	0.1881	0.9152	0.9603	0.9372	0.9531
0.1741	2.0	3890	0.1464	0.9391	0.9651	0.9520	0.9619
0.1467	3.0	5835	0.1302	0.9536	0.9585	0.9560	0.9645
0.1278	4.0	7780	0.1230	0.9576	0.9637	0.9606	0.9665
0.1158	5.0	9725	0.1126	0.9627	0.9651	0.9639	0.9695
0.1047	6.0	11670	0.1099	0.9638	0.9668	0.9653	0.9703
0.0964	7.0	13615	0.1090	0.9641	0.9684	0.9663	0.9712
0.0856	8.0	15560	0.1087	0.9664	0.9688	0.9676	0.9714
0.0778	9.0	17505	0.1120	0.9675	0.9679	0.9677	0.9712
0.0712	10.0	19450	0.1126	0.9664	0.9722	0.9693	0.9724
0.0656	11.0	21395	0.1144	0.9678	0.9701	0.9690	0.9718
0.0582	12.0	23340	0.1184	0.9682	0.9696	0.9689	0.9723
0.0532	13.0	25285	0.1215	0.9686	0.9712	0.9699	0.9727
0.0485	14.0	27230	0.1269	0.9697	0.9718	0.9707	0.9721
0.0447	15.0	29175	0.1293	0.9693	0.9717	0.9705	0.9727
0.039	16.0	31120	0.1317	0.9690	0.9719	0.9705	0.9723
0.0363	17.0	33065	0.1376	0.9689	0.9721	0.9705	0.9724
0.0333	18.0	35010	0.1396	0.9695	0.9721	0.9708	0.9721
0.0303	19.0	36955	0.1424	0.9700	0.9740	0.9720	0.9731
0.0274	20.0	38900	0.1456	0.9700	0.9734	0.9717	0.9736
0.0262	21.0	40845	0.1499	0.9692	0.9732	0.9712	0.9726
0.0232	22.0	42790	0.1522	0.9702	0.9732	0.9717	0.9733
0.0229	23.0	44735	0.1543	0.9706	0.9732	0.9719	0.9736
0.0214	24.0	46680	0.1543	0.9703	0.9738	0.9721	0.9733
0.0204	25.0	48625	0.1552	0.9703	0.9736	0.9720	0.9734

Framework versions

• Transformers 4.35.0
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1